Apple has introduced a new Mini Apps Partner Programme that halves App Store commissions for qualifying mini apps to 15 percent, marking a significant shift in how Apple wants developers to build and monetise app-within-app experiences.

A New Approach To App Store Revenue

Apple has announced that developers who host mini apps inside a larger iOS or iPadOS app can now qualify for a reduced 15 percent commission on digital purchases made within those mini apps. Mini apps are small, self-contained experiences built using web technologies such as HTML5 and JavaScript. They run inside a host app rather than being downloaded as separate apps from the App Store.

Apple has supported this format since 2017 under its App Review Guideline 4.7, which covers mini apps, mini games, streaming games, chatbots, plug-ins and emulators. This new programme is the first time Apple has offered a financial incentive that directly targets this part of its ecosystem.

The standard App Store commission can be as high as 30 percent for many in-app purchases, so the new 15 percent rate represents a meaningful reduction for developers who operate or contribute to mini app ecosystems. Apple says the aim is to help developers “grow their business” while ensuring that mini apps continue to meet App Store safety, age-rating and payments standards.

What Mini Apps Are And Why They Matter

Mini apps allow users to access small, task-based experiences without installing a separate full app. For example, a user might open a messaging app and launch a mini game, shopping experience, restaurant booking or banking tool directly within it.

The concept has existed for years in China, where WeChat mini programmes have become a major part of digital life. They let users book taxis, play games, access government services or shop online, all from within a single app. Tencent’s ecosystem has grown to such an extent that analysts estimate it has well over a billion active users.

Similar ideas have now appeared in other services. For example, LINE (a Japanese messaging and social app), Alipay (a major Chinese digital payments platform), Telegram (the global messaging app) and Discord (a communication platform popular with gaming and online communities) all offer mini app-style features. ChatGPT also appears to be moving into the space by allowing users to open external services from inside its chatbot, including travel, retail, music and design tools. This trend creates a growing shift in how people discover and interact with digital services.

Apple’s move, therefore, could be seen as a clear signal that Apple intends to support similar patterns on iOS, rather than allowing super apps, AI platforms or rival ecosystems to define this behaviour without Apple’s involvement.

Why Now?

The timing appears tied to several overlapping pressures. For example, regulatory scrutiny has intensified in the United States, Europe and the United Kingdom over Apple’s control of in-app payments and its App Store commissions. Authorities have questioned whether Apple’s rules limit competition. Apple has already faced investigations around super apps, with US regulators arguing that Apple’s policies restricted the growth of app-within-app ecosystems.

There is also a commercial backdrop. Reports have suggested that Apple and Tencent previously agreed a 15 percent commission for purchases made through WeChat’s mini apps. Given the scale of WeChat’s reach, even a small slice of that activity could be extremely valuable to Apple over time.

Apple is also most likely responding to the rise of AI platforms that attempt to reduce reliance on traditional apps. Some developers have speculated that if users spend more time in AI chatbots and transact through them, the App Store’s central role could weaken. Mini apps give Apple a way to reassert influence over this changing landscape.

How The Programme Works

To join the Mini Apps Partner Programme, a developer must operate a host app available on the App Store for iOS or iPadOS. The host app must comply with all Apple Developer Programme conditions and App Review Guideline 4.7, including the requirement to provide a detailed manifest listing every mini app and its metadata.

Participating apps must support specific Apple technologies. These include the Advanced Commerce API, which manages in-app purchase flows for mini apps, and the Declared Age Range API, which helps developers present age-appropriate content. Apple says this provides a safer and more consistent experience for customers.

Developers must also use Apple’s in-app purchase system. Purchases inside mini apps can include consumables, non-consumables, auto-renewing subscriptions and non-renewing subscriptions. If these purchases meet Apple’s criteria and are handled through the Advanced Commerce API, the 15 percent commission applies.

This means that while developers receive a lower fee, they must integrate more of Apple’s commerce and safety tools to qualify. Apple has made clear that the reduced rate is conditional on this deeper technical alignment.

What It Means For Developers

For developers who run mini app ecosystems, such as messaging platforms, digital wallets or gaming communities, the financial impact is pretty straightforward, i.e., a lower fee means more revenue stays within the ecosystem. A host app might use that extra revenue to invest in more mini apps, or to share income with third-party creators more generously.

The programme may also make it more attractive for smaller studios or service providers to build mini apps instead of full native apps. A mini app can be faster to develop and easier to distribute because users do not need to search for or install anything. Host apps with large user bases could become important distribution channels for businesses of all sizes.

At the same time, some developers have voiced concerns about the additional work required to meet Apple’s technical requirements. The Advanced Commerce API and the detailed manifest process introduce extra steps that may be burdensome for small teams.

Competitors And The Wider App Market

Apple’s move directly intersects with the strategies of companies that operate super app-like platforms. WeChat’s mini app ecosystem is the clearest example, but others are emerging. Google already supports Android instant apps, and messaging services worldwide are experimenting with their own in-app experience formats.

The rise of mini apps could gradually change consumer behaviour. For example, if users spend more time inside host apps that contain multiple mini apps, they may download fewer standalone apps from the App Store. This could be an opportunity and a risk for Apple, since it could reduce direct App Store engagement while opening new revenue paths within host environments.

AI platforms also play a role here. For example, mini apps inside AI tools introduce a new layer of app discovery and interaction. Apple’s decision to strengthen the economic and technical framework for mini apps may help keep developers focused on the App Store ecosystem rather than diverting too much attention to alternative platforms.

What Users And Businesses Will Notice

For everyday users, the change will mostly be felt inside the apps they already use. Mini apps can launch quickly, offer simple interfaces and provide focused features without requiring a full install.

For businesses, this change could widen opportunities to appear inside high-traffic apps without committing to a full native app build. For example, retailers, travel companies, financial services, entertainment platforms and many other sectors could use mini apps to reach customers more efficiently. The standardised payment and refund process through Apple may also reassure customers, particularly those making purchases in unfamiliar mini apps.

Challenges And Criticisms

However, some early reactions from developers suggest that Apple’s programme may reinforce, rather than relax, Apple’s control. For example, developers must adopt Apple’s payment tools and age-rating systems to qualify for the lower commission, which critics argue keeps Apple firmly in charge of the revenue chain.

There are also ongoing concerns about App Store competition. Although the fee is lower, developers still cannot use their own payment rails. Privacy groups have questioned whether Apple’s age-rating system will satisfy regulators who are proposing stricter verification measures.

Discoverability remains another challenge for mini apps in general. For example, in large host apps, mini apps risk becoming difficult to find unless the platform provides strong search tools or clear navigation. Apple’s metadata requirements aim to improve transparency and quality, but they also increase the workload for developers who manage large mini app catalogues.

Investors Positive

That said, investors appear to view the programme positively, describing it as a strategic move that supports revenue growth while strengthening Apple’s position as app behaviour evolves. The response from developers and users over the coming months will reveal whether the balance of incentives works in practice.

What This Means For Your Business?

Apple’s decision could reshape how developers think about distributing lightweight digital experiences, since the economic incentive is far stronger than anything Apple has offered around mini apps before. The requirement to adopt Apple’s own commerce and safety tools keeps the company firmly at the centre of the transaction chain, yet the reduced commission makes the trade-off more appealing than previous arrangements. This balance will matter as host apps weigh up whether the increased technical work is justified by the additional revenue and the chance to attract more third-party creators.

The wider market impact could be significant because mini app ecosystems have already changed digital behaviour in other regions. Apple’s move suggests that similar patterns may now emerge more visibly on iOS. If host apps that already attract large audiences begin to expand their mini app offerings, a growing share of daily digital activity could take place inside these environments instead of through standalone native apps. This may alter how services are discovered, how often users browse the App Store and how developers plan their product strategies.

UK businesses may find that mini apps offer an important new route to reach customers who prefer quicker, simpler interactions. A retailer, service provider or travel firm could appear inside a widely used host app rather than relying entirely on its own native app to attract attention. This could mean lower development costs, a broader reach and could give businesses access to an environment where purchases, refunds and subscriptions are handled through familiar Apple systems that many customers already trust.

Others will also be watching the adoption rate closely. For example, regulators may take interest in how Apple links the lower commission to its own technologies, especially in markets where competition and platform control are under scrutiny. Developers will want to see whether Apple’s technical requirements remain manageable as the number of mini apps grows. Host platforms will need to balance the commercial opportunity with the operational responsibility of policing large catalogues of third-party content.

It is likely that the coming months will show whether developers embrace the model at scale or continue to rely on native apps and alternative platforms. Apple has put forward a clearer financial incentive at a moment when the structure of the app ecosystem is evolving, and both the market response and the regulatory environment will shape what happens next.

Rightmove’s share price has fallen by more than a quarter after the UK’s biggest property portal told investors it will slow near-term profit growth in order to fund a major new programme of artificial intelligence investment.

Rightmove Shares Plunge

Rightmove used its early November trading update to outline a significant shift in strategy, announcing plans to spend around £60 million over the next three years on AI-driven upgrades to its platform, tools and internal systems. The company said this investment is central to how it intends to run the business, improve its product suite and position itself for higher long-term growth.

Consequently, the company now expects underlying operating profit to rise by only 3 to 5 per cent in 2026, compared with about 9 per cent growth this year. Revenue growth for 2026 is still forecast at between 8 and 10 per cent. Investors reacted sharply to the reduction in profit expectations, pushing the share price down by as much as 28 per cent during early trading. The stock recovered some ground later in the day, although it still closed more than 12 per cent lower and hit a new 52-week low.

Chief executive Johan Svanstrom said the company was “already working on a wide range of exciting AI-enabled innovations” and that AI is “becoming absolutely central” to how Rightmove operates. He said the investment would “create an even stronger platform and higher-growth business over time”, with the company targeting more than 10 per cent annual revenue growth by 2030.

Why Is Rightmove Investing So Heavily In AI?

Rightmove has framed this new strategy as a deliberate move into an investment phase that runs from 2026 to 2028. The company says the spending will cover three areas, which are:

Consumer-facing improvements, including conversational search tools that allow users to describe what they want in natural language, more personalised recommendations and a virtual mortgage assistant that can guide users through affordability and product options.

Major upgrades to Rightmove’s internal systems, where AI is expected to automate workflows, speed up data processing and improve customer service. This includes what the firm described as re-platforming significant parts of its back-end infrastructure.

Research and development focused on new products and revenue lines. Rightmove has already identified more than two dozen AI projects it wants to develop during this investment cycle.

The company said these tools are designed to help both consumers and agents by improving search accuracy, increasing the speed of listing updates and delivering more actionable insights from the portal’s large data sets.

Why The Market Reacted So Strongly

The sharp fall in Rightmove’s share price reflects a combination of surprise and wider market anxiety. For example, investors have traditionally viewed Rightmove as a highly predictable, low-risk business with very high margins, low capital requirements and steady subscription income from estate agents and developers. Therefore, a sudden fall in forecast profit growth, even if temporary, looks like a significant departure from expectations.

The sharp market reaction reflects a mix of scepticism and uncertainty. For example, analysts have noted that although investing for future growth is usually welcomed, the size and timing of the planned AI spend left investors questioning whether the short-term hit to profit was justified. Commentators have also highlighted that while AI could help Rightmove make better use of its data and improve efficiency, there were concerns the company might be committing substantial funds to technology projects without clear evidence of how quickly they would deliver returns.

Analysts at UBS described the move as a “strategic pivot” that leaves the market with unanswered questions about the timing and return on investment. Others, including RBC (Royal Bank of Canada’s investment banking and research division) and Peel Hunt (a UK-based investment bank and equity research firm), have taken a more positive view, suggesting the sell-off may be overdone and arguing that the investment could help Rightmove maintain its lead in an increasingly competitive market.

The update also seems to have come at a difficult moment for global technology stocks more broadly. For example, fears of an overheating AI sector have triggered sell-offs across US and European markets over the past week, and investors appear cautious about companies committing large sums to AI projects with uncertain payoff periods. Rightmove’s update therefore landed in a market already highly sensitive to any sign of increased spending on AI.

Implications For Agents And The Property Industry

Rightmove’s paying customers are estate agents, lettings agents, developers and other professionals who rely on listings and data tools to win clients and run their businesses. For them, the impact of the new strategy depends on whether the AI improvements genuinely make the platform more effective.

Rightmove has already launched products such as Optimiser Edge, which uses data to help agents target new instructions and improve marketing. Strong take-up has encouraged Rightmove to double down on data-led tools. If the new AI tools deliver as promised, agents could access richer insights into pricing, demand, buyer behaviour and lead quality. That could help them work more efficiently and justify the portal’s subscription fees, which have been a long-running flashpoint in the industry.

Some agents may welcome these updates, whereas others are likely to be concerned that rising investment costs could lead to further fee increases. This tension has been reflected in a new legal claim against Rightmove that accuses the company of unfair pricing. Although the claim is separate from the AI announcement, the perception of rising costs will remain a key issue for stakeholders.

Developers, landlords and corporate property owners may benefit from more accurate pricing tools, better audience targeting and stronger data on local market dynamics. If Rightmove uses its data to launch new B2B products, this could strengthen its position further across the property ecosystem.

Homebuyers, Renters And Landlords

For individual users, the difference will mostly be seen in the platform experience. For example, conversational search could make it easier to find suitable homes without navigating multiple filters. AI-driven recommendations may also surface properties more relevant to specific needs or preferences, while improved data analysis could give buyers and renters clearer insights into pricing trends, local demand and affordability.

These tools could, therefore, save time for renters, first-time buyers and families trying to navigate an often opaque and fast-moving market. More accurate recommendations could also mean fewer wasted viewings.

However, these improvements come with quite a few questions. For example, more personalisation means more data use, and users will want to know how their information is used, stored and analysed. There is also a broader debate over whether automated valuation tools could introduce bias or distort local pricing. As AI becomes more visible in property technology, these issues will attract increased attention.

The UK Property Market

Rightmove dominates online property search in the UK by a wide margin. This means that any shift in its operating model actually has potential implications for the wider housing market. Better search tools could, in theory, improve matching between buyers and sellers, shortening transaction times and reducing friction.

More accurate pricing tools may help reduce the difference between asking and achieved prices, particularly in slower markets. Improved analytics could help developers understand demand patterns and assist landlords in managing rental portfolios.

However, that said, a more advanced platform could also strengthen Rightmove’s position. For example, competitors such as Zoopla and OnTheMarket already invest heavily in technology, and they may now face pressure to respond with their own accelerated AI programmes. If Rightmove’s AI tools become significantly more advanced than its rivals’, agents may feel increasingly locked in. This raises questions about competition, pricing power and how much choice agents and landlords can realistically exercise.

The Portal Landscape

It’s worth noting here that Rightmove’s rivals have been developing their own AI tools for some time now, particularly in automated valuation, user search and agent dashboards. The size of Rightmove’s latest programme may lead competitors to increase their own investment or reposition themselves more aggressively on pricing.

For estate agents, the announcement could signal a future in which portals compete less on the volume of listings and more on the intelligence and value added by their technology. The next few years are likely to be defined by how well AI helps agents attract vendors, manage leads and handle day-to-day operations.

For investors, the key question here may be whether this investment pays off in the second half of the decade. Rightmove believes its operating profit will begin to rebound after 2028 and that higher growth rates will follow. The stock market will, no doubt, be watching pretty closely to see whether those expectations translate into real performance, stronger user engagement and a clear competitive edge.

What Does This Mean For Your Business?

The immediate challenge for Rightmove is proving that these investments will deliver practical improvements rather than simply increasing costs. Investors will want to see evidence that AI tools can streamline operations, deepen user engagement and support new revenue lines without undermining the predictability that has defined the business to date. Estate agents will also be watching closely because AI driven workflows and data products will only be seen as valuable if they genuinely help them win instructions, price properties more accurately and run leaner operations. For those users, the question is less about the scale of the investment and more about whether it translates into tools that make day to day work faster and more effective.

Homebuyers, renters and landlords face a different set of considerations. For example, if conversational search and personalised recommendations improve accuracy and reduce wasted time, the platform could feel more intuitive and more useful during a move. Concerns around data use and algorithmic fairness will still need addressing as AI driven products expand, but the potential for clearer pricing insight and better matching remains significant. The wider property market will also feel the effects because more precise analytics and smarter discovery tools could influence how quickly homes sell, how properties are valued and how landlords plan their portfolios.

The broader implications for UK businesses are based around adoption, competition and capability. Many firms are assessing how quickly they should modernise their own digital systems, and Rightmove’s shift illustrates how even established, highly profitable businesses are accelerating their AI strategies. It provides a useful signal to UK decision makers that AI investment is increasingly being treated as a long term infrastructure requirement rather than an optional upgrade. For companies that supply or rely on property insights, there may be new opportunities to integrate richer data streams into planning, risk assessment and market forecasting.

Competitors will now need to decide whether to match this level of investment or differentiate more clearly on price and service. The risk for the wider portal landscape is growing concentration if Rightmove’s AI programme strengthens its lead, although the response from rivals is likely to shape how the market evolves. If they produce credible alternatives with strong AI features of their own, agents and landlords may benefit from greater choice and lower pressure on fees.

For regulators and policymakers, the developments highlight a sector where data, pricing power and platform dominance intersect. The balance between innovation and competition will be important because the benefits of AI will only be felt widely if the market remains open, transparent and fair for users. The next few years will reveal whether this investment cycle creates a more efficient and more dynamic property ecosystem or whether it intensifies existing concerns about market concentration and rising costs.

Association of British Insurers (ABI) figures show that cyber insurance payouts in the UK have tripled, reaching £197 million in 2024 as businesses face increasingly costly cyber-attacks, particularly from ransomware and malware.

The number of cyber insurance policies has also risen, with 17 per cent more businesses taking out coverage in 2024. However, experts warn that not all claims are guaranteed to be paid. Insurers are tightening requirements, and failure to meet security standards or maintain effective recovery plans may limit payouts.

Businesses must ensure they implement robust cybersecurity measures, including secure backups and effective recovery plans. Cybersecurity should be a core business priority, with regular risk assessments and a proactive security culture to mitigate risks and safeguard against costly attacks.

An Essex couple have become the first in the UK to heat their home using a mini data centre in their garden shed, in a trial designed to cut energy bills and support low income households through the transition to net zero.

Pilot Scheme

Terrence and Lesley Bridges live in a modest two bedroom bungalow near Braintree in Essex. Their home is owned by Eastlight Community Homes, a social housing provider, and they are part of a pilot run jointly by UK Power Networks and Thermify through an innovation project called SHIELD. The couple were selected for the pilot because they rely heavily on their heating, especially as Lesley lives with spinal stenosis and is in significant pain when temperatures drop.

Thermify HeatHub – Huge Savings

Since the installation of the Thermify HeatHub, their monthly energy costs have fallen from around £375 to between £40 and £60. Terrence said: “It truly is brilliant. I’m over the moon that we got picked to trial this out. You can’t fault the heating system, it is a 100 per cent improvement on what we had before.” Lesley added: “You don’t need to go to a sauna after coming here.” Their experience is one of the first real world demonstrations of a heating concept that blends clean energy, digital infrastructure and social support.

Who Is Thermify?

The heating unit in the Bridges’ shed is called a HeatHub. It is developed by the British company Thermify, which offers cloud computing services to businesses. Instead of housing its servers in a single large data centre, Thermify installs small clusters in people’s homes, where the heat generated by data processing is captured and used as low cost domestic heating.

The wider programme is actually part of SHIELD, which stands for Smart Heat and Intelligent Energy in Low income Districts. SHIELD is run by UK Power Networks through the Strategic Innovation Fund. Its aim is to help people who would normally be excluded from the shift to low carbon technologies because of high upfront costs. The project brings together Thermify, Eastlight Community Homes, community energy groups and technical partners to develop what they describe as a Social ESCo model. Under this model, equipment such as solar panels, batteries and HeatHubs is funded upfront by an energy services company and repaid over time through the value created by the technologies.

How The Data Centre In Their Shed Works

Inside the HeatHub are around 500 Raspberry Pi Compute Modules, all submerged in a special oil. As these computers run cloud tasks for Thermify’s business clients, the electricity they use becomes heat, which raises the temperature of the surrounding oil. That heat is then transferred into a heat store and the home’s central heating and hot water systems.

The principle is pretty simple. For example, computers turn electricity into information but all the electricity eventually becomes heat. Traditional data centres spend significant amounts of extra electricity on cooling systems that remove the heat and release it into the air. Thermify’s approach uses that unavoidable heat twice by turning it into a resource for the household.

A dedicated network line is installed so the unit can send and receive data without affecting the resident’s broadband. From the resident’s point of view, it behaves much like a boiler, controlled through familiar heating settings. The Bridges’ shed also contains a solar inverter and a battery, meaning their HeatHub is part of a small integrated energy system that stores and manages electricity through the day.

Why It Cut Their Bills

In the Bridges’ case, the combination of the HeatHub, solar panels and battery storage has transformed their energy use. Thermify pays for the electricity needed to run the computing tasks because this is part of its service to business clients. The heat produced from this process is supplied to the home at a low or no cost because the energy is already being paid for. SHIELD tenants who receive HeatHubs also pay a small standing charge for heat, although UK Power Networks expects this to be significantly lower and more predictable than the cost of gas for many low income families.

Thermify points to independent modelling that suggests this kind of distributed computing could reduce carbon emissions from data centre operations by about 75 per cent on average. SHIELD’s own modelling suggests combining HeatHubs with solar and batteries could reduce household energy costs by 20 to 40 per cent and cut heating related emissions by more than 90 per cent.

Data, Energy And Tech Companies

The concept has clear implications for cloud and data centre operators. For example, data centres already account for roughly 2.5 per cent of the UK’s electricity consumption and the sector’s demand is forecast to grow rapidly in the next five years. As more companies expand into artificial intelligence and digital services, pressure is rising to reduce the environmental impact and find practical uses for the heat that data centres produce.

Distributed systems like Thermify’s also offer an alternative to building ever larger centralised facilities. Although HeatHubs cannot handle the heavy workloads required for advanced artificial intelligence, they can run many common tasks such as analytics, apps or batch processing. If rolled out at scale, the model could create a network of tens of thousands of small data nodes that serve business customers while heating homes. SHIELD itself has a long term ambition to deploy up to 100,000 such systems a year by 2030.

The approach may also interest energy companies and grid operators. For example, embedded assets such as HeatHubs can help manage peaks and troughs in local demand and provide flexibility services to the grid. SHIELD is exploring how these devices might be combined with peer to peer energy trading and other smart local energy systems.

Sustainability Advantages

There’s clearly an environmental case for improving overall energy efficiency and reducing reliance on fossil fuels. With up to 30 per cent of a data centre’s electricity used solely for cooling, capturing that heat and using it to warm homes can replace the need for gas and reduces the total energy wasted.

There are also potential social benefits to consider here. For example, many low income households cannot afford the upfront investment needed for heat pumps or solar installations. SHIELD’s Social ESCo model aims to solve this by funding the equipment and repaying costs through the value generated by the assets. Early stages of the project show strong interest among tenants who are worried about energy bills but keen to adopt cleaner solutions.

Not A Totally New Idea

It should be noted here that the idea of using data centre heat in buildings is not new. For example, in Devon, a startup called Deep Green operates a washing machine sized digital boiler at a local swimming pool. The servers inside the unit warm the mineral oil surrounding them and the captured heat is used to heat the pool. Reports indicate that the installation has reduced the pool’s gas use by more than half and cut emissions by dozens of tonnes of CO₂ each year. A recent investment from Octopus Energy aims to expand similar units to more than one hundred pools across the UK.

Also, another British company, Heata, attaches small servers to domestic hot water tanks. Homeowners earn a payment for hosting cloud workloads and the heat from the servers warms their water. In mainland Europe, district heating networks in cities such as Odense, Paris and Stockholm already capture heat from large data centres to supply nearby homes and offices.

Key Challenges And Criticisms

Although the Bridges’ results are positive, there are ongoing questions about reliability and long term performance. For example, HeatHubs depend on a steady demand for cloud computing. If business workloads fall or move to other locations there could be uncertainty about how much heat is produced and how backup systems would operate. Trials like SHIELD allow operators to test these scenarios before any wider rollout.

There are also some practical issues to consider. HeatHubs need secure network connections, scheduled maintenance and clear communication so residents understand how the system works. Social landlords also have to consider noise, space and safety. Early feedback from SHIELD has highlighted the importance of strong support and simple user experience.

There is also a broader debate about whether heat reuse can keep pace with the rapid growth in data centre energy demand. Artificial intelligence training and inference use far more electricity than the kind of workloads Thermify deploys. Even with heat capture, growing numbers of data centres will place pressure on local electricity networks. Policymakers and regulators are increasingly encouraging heat reuse but stress that it must be combined with wider grid planning and efficiency measures.

For now, however, the Bridges’ warm bungalow in Essex has become a test case for how computing and heating might come together, offering an early glimpse of a model that could reshape how data centres are built and how homes are heated in the years ahead.

What Does This Mean For Your Organisation?

The trial highlights how digital infrastructure and domestic energy systems can support each other, which is why it is gaining interest across the UK. Data centres are expanding rapidly as businesses adopt artificial intelligence and cloud services, yet their rising electricity use and waste heat are becoming harder to manage. A system that captures this heat and delivers it as affordable, low carbon warmth offers clear benefits for households and creates a more efficient model for the tech companies that rely on constant processing power.

There are important implications for UK businesses here. For example, a distributed network of small data hubs could give companies access to computing capacity with a lower environmental impact, supporting sustainability commitments while easing pressure on the wider grid. Energy providers and local authorities may also see value in systems that help stabilise local demand and offer predictable heating costs for low income residents.

The Social ESCo model is another key part of the story, as it removes the upfront cost barrier that prevents many households from adopting low carbon technologies. If the model proves reliable at scale, it could influence how social housing providers, councils and developers approach retrofit programmes and new energy installations.

Heat reuse is likely to become more common as the UK works towards decarbonising heat. Projects like SHIELD show how data processing, renewable generation and home heating can be combined in a practical way, although long term questions remain around reliability, workload availability and system management. Even so, the Bridges’ experience demonstrates how an integrated approach can reduce bills, cut emissions and provide a template that could be adapted for both homes and businesses in the years ahead.

You can now extract text and see other relevant information about images within your OneDrive folder, without even having to open them. Furthermore, you can pretty well extract all the relevant meta information about all the files you have in there, thereby saving time having to open them up each time, just to see what’s there.

[Note – To Watch This Video without glitches/interruptions, It may be best to download it first]

Did you know you can attach a reminder to any message you send, so you never forget to follow up if you don’t hear back? This tiny step guarantees important emails stay top‑of‑mind without leaving your inbox.

To set a reminder:

– While composing a message, click Follow Up → Add Reminder.
– Choose the date and time you want Outlook to nudge you, or pick a preset like “Tomorrow”.
– If you need to adjust the reminder later, open the sent email, click Follow Up again, and edit or cancel it.

Why it’s handy: This Outlook 365 feature will pop up a reminder if you haven’t received a reply, turning a simple email into a task that never slips through the cracks.
Give it a try and see how it streamlines your follow‑ups!

Shopify says artificial intelligence (AI) is now driving record levels of shopping activity, with traffic to its merchants’ stores up sevenfold since January and AI-attributed orders rising elevenfold, claiming it marks the start of a new “agentic commerce” era.

Shopify’s AI Milestone Announced Alongside Strong Financials

These latest figures were unveiled on 4 November 2025 during Shopify’s third-quarter earnings call for the period ending 30 September. The Canadian-based e-commerce software company, which powers millions of businesses in more than 175 countries, reported revenue of around US $2.84 billion, a 32 per cent rise year on year, with gross merchandise volume (GMV) climbing to US $92 billion, also up 32 per cent. Free cash flow margin (the profit left after expenses and investments) stood at about 18 per cent, marking nine consecutive quarters of double-digit free cash flow margins.

Operating income reached US $434 million, slightly below analyst expectations, but executives emphasised that AI-driven performance was the real story of the quarter. “AI is not just a feature at Shopify. It is central to our engine that powers everything we build,” said president Harley Finkelstein during the call, describing AI as “the biggest shift in technology since the internet.”

Shopify and Its Role in Global Commerce

Founded in Ottawa in 2006, Shopify provides digital infrastructure that allows merchants to start, scale and run retail operations online and in-store. For example, the company’s tools cover web hosting, checkout, payments, logistics, marketing, analytics and third-party app integrations. Its reach includes major brands such as Estée Lauder and Supreme, as well as small independent businesses.

The Value of Its Data Network

Shopify’s value essentially lies in its vast data network. For example, with millions of active merchants generating billions of transactions each year, the company can analyse patterns across product categories, price points, consumer behaviour and regional trends. Finkelstein said this data scale provides a distinct edge in the AI era, allowing Shopify to “turn our own signals — support tickets, usage data, reviews, social interactions or even Sidekick prompts — into fast, informed decisions.”

AI Traffic and Orders See Explosive Growth

The most striking statistics from the earnings call were that traffic from AI tools to Shopify-hosted stores is up seven times since January 2025, and that orders attributed to AI-powered search are up eleven times over the same period. Although Shopify did not provide absolute numbers, the growth rate suggests that AI chatbots and conversational assistants are starting to play a meaningful role in how customers find and purchase products.

The company’s internal survey found that 64 per cent of consumers are likely to use AI during the Christmas holiday shopping season, which is a sign, it says, that shoppers are already comfortable relying on digital assistants for product discovery and comparison.

Finkelstein has framed this change as more than a short-term sales boost. “We’ve been building and investing in this infrastructure to make it really easy to bring shopping into every single AI conversation,” he told analysts. “What we’re really trying to do is lay the rails for agentic commerce.”

What Does ‘Agentic Commerce’ Mean?

Shopify’s term “agentic commerce” refers to a model where AI agents act on behalf of consumers, guiding them through discovery, evaluation, checkout and even post-purchase stages such as returns and reordering. For example, rather than searching through multiple sites, a user can simply describe what they want to a conversational AI assistant, which can then query databases, compare prices, and complete the transaction.

The “Commerce for Agents” Stack

To support this model, Shopify has built what it calls its “commerce for agents” stack. This includes a product catalogue system designed for AI queries, a universal shopping cart that lets consumers buy across multiple merchants, and an embedded checkout layer using Shop Pay for one-click transactions. These features are being integrated into platforms such as ChatGPT, Microsoft Copilot and Perplexity through formal partnerships announced earlier this year.

This infrastructure means that AI assistants can browse Shopify merchants’ catalogues and complete purchases directly within chat interfaces. As AI-driven discovery becomes more conversational, Shopify aims to position itself as the primary retail backbone behind these agent-led interactions.

The Scout System

Shopify is also deploying AI internally. For example, its “Scout” system analyses hundreds of millions of pieces of merchant feedback to help employees make product and support decisions more effectively. “Scout is just one of many tools we’re developing to turn our own signals into fast, informed decisions,” Finkelstein said.

Sidekick

Another major tool is Sidekick, an AI assistant embedded within Shopify’s merchant dashboard. Sidekick can analyse sales trends, suggest pricing adjustments, generate marketing copy, or create reports on command. In the third quarter alone, more than 750,000 shops used Sidekick for the first time, generating close to 100 million conversations. Shopify says this helps merchants operate more efficiently and spend less time on routine administrative work.

Shop Pay

Shop Pay is the company’s one-click checkout solution and remains a cornerstone of its AI ecosystem. In Q3 it processed about US $29 billion of GMV, a 67 per cent increase year on year, and accounted for around 65 per cent of all transactions on the platform. This integration ensures that when AI agents complete orders, Shopify retains control of the payment flow and associated data.

Global Impact and European Opportunity

Finkelstein told investors that consumer confidence “is measured at checkout,” adding that shoppers on Shopify “keep buying” and “keep returning.” He noted that demand has remained resilient across categories, even as economic uncertainty persists. Europe appears to be a particular bright spot, with cross-border GMV (the total value of all sales made through Shopify’s platform) steady at 15 per cent of total sales and growth in sectors such as fashion and consumer goods.

For UK and European merchants, this could present a new phase of opportunity. For example, businesses already using Shopify can benefit from being automatically visible to AI-driven discovery systems without developing custom integrations with each platform. By ensuring that product listings are detailed, structured and machine-readable, merchants can increase their chances of being recommended by AI agents.

There is also a potential opening for agencies and developers to specialise in optimising “agent-ready” storefronts, designing catalogues and metadata that AI systems can interpret accurately. For smaller retailers, this could be an efficient route into AI commerce without the high cost of in-house development.

How AI Is Changing the Competitive Landscape

Shopify’s emphasis on AI-driven commerce poses strategic questions for competitors. For example, Amazon and major regional marketplaces already use AI recommendation engines, but Shopify’s model offers decentralised access: independent merchants can collectively benefit from the same AI infrastructure without surrendering control of their brands.

If agentic commerce grows as Shopify predicts, discovery and purchasing could increasingly occur inside chat platforms rather than traditional websites or search engines. That would reshape marketing and customer acquisition strategies, pushing retailers to focus more on structured data, integration quality and conversational optimisation.

For Shopify itself, the rise of agent-driven traffic could also reinforce its role as the connective tissue of global retail, potentially deepening its partnerships with large AI providers and securing a share of new sales channels that bypass traditional web search entirely.

Opportunities and Challenges for Businesses

For merchants, the potential benefits include higher-quality leads, faster conversions, and less reliance on paid advertising. AI-powered assistants can surface relevant products to users who are ready to buy, reducing friction in the path to purchase. The integration of Sidekick also promises time savings through automation of everyday tasks like inventory monitoring and campaign planning.

However, the challenges are equally significant. For example, attribution remains a key question, i.e., determining which sales are truly “AI-driven” is difficult when customers interact across multiple devices and channels. There is also the issue of discoverability. As AI agents narrow recommendations to just a few products, competition for visibility may intensify, potentially favouring larger brands that can afford dedicated AI-optimisation strategies.

Data privacy and regulatory compliance are further concerns, especially in the UK and EU. For example, agentic commerce depends on detailed user data to personalise results, and any sharing of this data between Shopify, AI partners and merchants will attract scrutiny under GDPR and related frameworks. Businesses will need clear consent processes and transparent data handling to maintain consumer trust.

Critics also warn of overreliance on automated systems that can misinterpret queries or produce inaccurate results. Large language models are known to “hallucinate”, and shopping assistants could recommend inappropriate or unavailable items. Shopify’s claim that AI represents autonomy rather than mere automation raises questions about accountability if an agent completes a transaction incorrectly or processes returns without oversight.

Despite these uncertainties, Shopify’s strategy and apparent success with it could be seen as a signal that conversational and agentic shopping will become a defining feature of global retail. The company’s 7× rise in AI-driven traffic and 11× increase in orders could be seen as providing the clearest evidence yet that the technology is beginning to translate from hype into measurable commerce.

What Does This Mean For Your Business?

Shopify’s results appear to show that AI-driven shopping is no longer an abstract concept but a tangible factor reshaping how consumers buy and how merchants sell. The company’s data and partnerships give it a strong early foothold in this emerging space, yet they also highlight the scale of change underway across the entire retail ecosystem. For merchants and technology partners, particularly in the UK, the lesson appears to be that conversational and agent-led shopping channels are likely to become a growing part of how customers discover and complete purchases. Those who adapt their product data, content and customer engagement models early will be better placed to capture new demand as AI assistants become a standard entry point to commerce.

At the same time, the risks are becoming more visible. For example, the concentration of traffic within a handful of AI platforms introduces new dependencies and competition for visibility that could prove as intense as traditional search engine optimisation. Data protection and transparency will remain major issues, especially in the UK and EU where regulators are tightening scrutiny on how consumer data is shared between AI systems and third-party platforms. Businesses will need to ensure that automation enhances customer experience without removing human accountability or trust.

For Shopify, the early surge in AI-related sales provides some validation of its long-term investment in agentic commerce, but the road ahead will depend on whether AI tools can sustain accuracy, reliability and fairness at scale. For retailers, investors and consumers alike, the company’s current momentum highlights the fact that AI is already changing commerce in practice, not just in theory, and the balance between innovation, control and transparency will define who benefits most from this new era.

In this Tech Insight, we look at how to get the most from WhatsApp groups by using all their key features to make chats more organised, productive, and secure for both organisers and members.

Why WhatsApp Groups Matter More Than Ever

WhatsApp has become the world’s primary messaging platform, used by over 2.9 billion people each month and handling around 130 billion messages every day. For families, clubs, workplaces, and local communities, it has evolved into an essential coordination tool rather than just a place to chat.

Groups now include a wide range of built-in tools designed to help organisers manage communication more effectively. For example, features such as polls, events, message reactions, and Communities have turned WhatsApp into a structured environment capable of handling everything from social groups to large-scale organisational networks.

Groups, Communities And Channels Explained

WhatsApp currently offers three main ways to reach groups of people, i.e., Groups, Communities, and Channels. Understanding the difference helps users decide which best fits their purpose.

1. Groups are the most familiar format. Everyone can send and receive messages, share files, and react to posts. Each group can now include up to 1,024 members, according to WhatsApp’s official documentation.

2. Communities sit above ordinary groups, acting as an umbrella structure that links several related groups together under one theme. They include a dedicated announcements group that lets admins share key updates with all members across every linked group. This is useful, for example, for schools, businesses, or local organisations that want to keep large audiences connected without merging everyone into one chat.

3. Channels work as a one-way broadcast tool. For example, followers can receive updates from public figures, brands, or organisations, but they can’t reply within the channel itself. Channels are therefore designed for updates rather than conversations.

Setting Up A Group The Right Way

Creating a WhatsApp group is quick, but setting it up thoughtfully helps it run smoothly in the long term. The organiser starts by naming the group, adding an image, and setting a short description that explains what the group is for.

Within the Group Settings menu, admins can control who is allowed to send messages, change the group name or picture, and add new members. The Approve New Members feature lets admins review join requests before they are accepted, helping to prevent unwanted participants or spam.

For example, a workplace coordinator might use this setting to restrict a project group to approved team members, while a community organiser could use it to make sure only verified residents join a neighbourhood group.

How To Create A Well-Structured WhatsApp Group

– Open WhatsApp and tap New chat, then New group.

– Add your chosen members and set a clear, descriptive group name.

– Write a short description outlining the group’s purpose and any basic rules.

– In Group settings, decide who can post messages or add new members.

– Enable Approve New Members if the invite link might be shared beyond your core group.

Admin Tools That Keep Groups Organised

Admins now have more control than ever before. For example, they can appoint multiple admins to share responsibilities, remove members, reset invite links, or change permissions without deleting the group.

When people are invited to groups they don’t recognise, WhatsApp now displays a context screen showing who created it, how many members it has, and options to leave immediately. This reduces confusion and protects users from scams or spam invitations.

Day-To-Day Tools That Keep Conversations On Track

The simplest features often make the biggest difference. Message reactions let users acknowledge posts quickly without sending separate replies, keeping chats more concise.

Users can press and hold a message (or hover over it on desktop) to bring up a reaction bar and choose an emoji that fits their response. Others can tap the same emoji to agree without cluttering the chat.

WhatsApp is also introducing threaded replies to group messages, allowing related responses to be grouped together for easier reading. Editing messages after sending has already been rolled out, offering more flexibility in busy conversations.

Polls And Events Simplify Group Decisions

Polls and events are now standard tools for coordination. For example, polls allow users to ask a question with multiple options and gather votes directly within the chat. They are ideal for deciding on meeting times, event themes, or team preferences.

Events, by contrast, let organisers create a calendar-style entry with a date, time, location, and optional call link. Members can mark whether they are going, maybe, or not going. Any changes made by the organiser update automatically for everyone.

How To Create A Poll Or Event

– Open the group chat and tap the attachment icon.

– Select Poll to create a multi-option question or Event to schedule an activity.

– Add the details, then send it to the group for members to respond.

File Sharing, Calls And Productivity Features

WhatsApp’s file-sharing limit is now 2GB, which allows large presentations, videos, or PDFs to be shared directly. These files are automatically sorted under Media, links and docs in the group information screen, making them easy to find later.

Also, voice and video calls within groups now support up to 32 participants, a feature that has made WhatsApp a lightweight alternative to traditional meeting apps. For example, the platform says more than two billion calls are made every day, many of them through group chats.

Pinned messages, currently being rolled out more widely, help keep key updates visible at the top of busy chats.

Creating Smaller Groups Within Groups

While WhatsApp doesn’t allow true “sub-groups” within a single chat, it now offers two ways to achieve similar organisation.

1. Communities link multiple related groups under one structure. For example, a school might have one community containing groups for each class plus an announcements group for all parents. A business could create groups for each department within a single community for consistent communication.

2. There is also the ability to set up smaller side groups manually for focused discussions. For example, a large club might have a main group for all members and a smaller Committee Planning group for organisers. The key is to keep communication transparent by mentioning in the main group when side discussions are taking place.

In general, Communities are best for structured, ongoing organisation with clear roles, while smaller groups work well for short-term collaboration or private planning.

Privacy, Security And Safety Features

WhatsApp also still offers end-to-end encryption across all group messages and calls, meaning only participants can read or hear what’s shared. The platform also provides safety screens that warn users before joining unknown groups, and clearer options for reporting suspicious content.

Admins can reset invite links at any time, preventing them from spreading publicly. Reporting tools now allow users to flag specific messages instead of entire chats, helping WhatsApp review potential scams more accurately.

These measures, combined with the Approve New Members setting and improved admin controls, make groups safer and easier to manage even as they grow larger.

Power Features For Large Organisations

For schools, clubs, or businesses managing multiple groups, the Communities feature provides top-level organisation. Each community includes an announcements group for updates, while linked sub-groups handle topic-specific discussions.

WhatsApp has also begun rolling out built-in message translation, allowing users to translate posts directly inside chats without switching apps. This is especially valuable for international organisations or multicultural teams.

Quick Checklist For Group Organisers

– Use clear names and descriptions to define the purpose of each group.

– Set permissions carefully to control who can post or add new members.

– Turn on Approve New Members to reduce spam.

– Replace long discussions with polls and events for better organisation.

– Encourage message reactions rather than repetitive replies.

– Consider Communities for managing multiple related groups.

– Keep shared files accessible under Media, links and docs.

– Reset invite links if they become public.

Used effectively, these features transform WhatsApp groups from cluttered chats into structured, secure, and genuinely productive spaces that make everyday coordination simpler for everyone involved.

What Does This Mean For Your Business?

When used well, WhatsApp groups can bring order and clarity to communication that might otherwise become fragmented across emails, calls, and messages. The platform’s steady introduction of tools such as polls, events, Communities, and permissions settings reflects a clear move towards more professional and structured group management. For everyday users, these features simplify coordination and decision-making. For organisers, they offer genuine administrative control without adding unnecessary complexity.

For UK businesses in particular, WhatsApp’s evolution into a full-featured collaboration space has many practical benefits. For example, many small firms already rely on it informally to connect remote staff, contractors, and customers, and the new tools make those networks easier to manage. The ability to approve new members, create communities for different departments, or schedule meetings directly within the app offers a low-cost way to keep teams connected in real time. Used responsibly, it could become an accessible alternative to larger, paid communication platforms for smaller organisations.

The same features also have value beyond business. Local groups, volunteer networks, and schools can all benefit from Communities that link separate discussions together while maintaining privacy and control. The addition of safety screens and end-to-end encryption keeps users protected while helping organisers maintain accountability.

WhatsApp is now becoming less of a casual messaging app and more of an organised communication environment where structure, transparency, and security define how people interact. For businesses, community organisers, and individual users alike, understanding and applying these group features effectively could turn WhatsApp into one of the most useful everyday coordination tools available.

Amazon has accused AI startup Perplexity of illegally accessing its e-commerce systems through its agentic shopping assistant, Comet, marking one of the first major legal tests of how autonomous AI tools interact with major online platforms.

Perplexity and Comet

Perplexity is a fast-growing Silicon Valley AI company valued at around $18 billion and known for its “answer engine”, which competes with Google and ChatGPT by providing direct, cited responses rather than lists of links. Its newest product, Comet, extends this model into what’s known as “agentic browsing”, which is software that not only searches but acts.

Comet can log into websites using a user’s own credentials, find, compare and purchase products, and complete checkouts automatically. The user might, for example, tell Comet to “find the best-rated 40-litre laundry basket under £30 on Amazon and buy it”. Comet then navigates the site, checks prices and reviews, and completes the order.

Perplexity says Comet is private, with login credentials stored only on the user’s device. It argues that when users delegate tasks to their assistant, the AI is simply acting as their agent, meaning it has the same permissions as the human user.

Amazon’s Legal Threat And Allegations

On 31 October 2025, Amazon sent Perplexity a 10-page cease-and-desist letter through its law firm Hueston Hennigan, demanding it immediately stop “covertly intruding” into Amazon’s online store. The letter essentially accuses Perplexity of breaking US and California computer misuse laws, including the Computer Fraud and Abuse Act (CFAA) and California’s Comprehensive Computer Data Access and Fraud Act (CDAFA), by accessing Amazon’s systems without permission and disguising Comet as a Chrome browser.

Amazon’s counsel, Moez Kaba, wrote that “Perplexity must immediately cease using, enabling, or deploying Comet’s artificial intelligence agents or any other means to covertly intrude into Amazon’s e-commerce websites.” The letter says Comet repeatedly evaded Amazon’s attempts to block it and ignored earlier warnings to identify itself transparently when operating in the Amazon Store.

According to the letter, Perplexity’s unauthorised behaviour dates back to November 2024, when it allegedly used a “Buy with Pro” feature to place orders using Perplexity-managed Prime accounts, a practice that Amazon says violated its Prime terms and led to problems such as customers being unable to process returns. After being told to stop, Amazon says, Perplexity later resumed the same conduct using Comet.

The company also alleges that Comet “degrades the Amazon shopping experience” by failing to consider features like combining deliveries for faster, lower-carbon shipping or presenting important product details. Amazon claims this harms customers and undermines trust in the platform.

Security Risks And Data Concerns

Amazon’s letter also accuses Perplexity of endangering customer data. For example, it points to Comet’s terms of use, which it says grant Perplexity “broad rights to collect passwords, security keys, payment methods, shopping histories, and other sensitive data” while disclaiming liability for data security.

The letter cites security researchers who have identified vulnerabilities in Comet. For example, The Hacker News reported in October that a flaw dubbed “CometJacking” could hijack the AI assistant to steal data, while a Tom’s Hardware investigation in August found that Comet could visit malicious websites and prompt users for banking details without warnings. Amazon says such flaws illustrate the dangers of “non-transparent” agents interacting directly with sensitive e-commerce systems.

Must Act Openly and Be Monitored, Says Amazon

While Amazon insists it is not opposed to AI innovation, it argues that third-party AI agents must act openly so their behaviour can be monitored. “Transparency is critical because it protects a service provider’s right to monitor AI agents and restrict conduct that degrades the shopping experience, erodes customer trust, and creates security risks,” the letter states.

Amazon warns that Perplexity’s actions violate its Conditions of Use, impose significant investigative costs, and cause “irreparable harm” to its customer relationships. It has demanded written confirmation of compliance by 3 November 2025, threatening to pursue “all available legal and equitable remedies” if not.

What Is Agentic Browsing?

Agentic browsing describes AI systems that can autonomously act on users’ behalf, e.g., from finding products and booking travel to filling forms and making payments. The concept represents a step beyond traditional automation, potentially turning AI from a passive search tool into an active personal assistant.

The appeal is that these systems can save time, reduce manual effort, and make repetitive digital tasks simpler. For consumers and business users alike, agentic assistants could automate procurement, research, and routine purchases.

However, it seems that this new autonomy also challenges the rules of engagement between users, AI developers, and online platforms. For example, when a human browses a site, the platform can track preferences, display promotions and tailor recommendations. When an AI agent acts in their place, it may bypass all those mechanisms and, crucially, any monetised placements or advertising.

Perplexity’s Response

Perplexity quickly went public with its response, publishing a blog post entitled Bullying is Not Innovation. It described Amazon’s legal threat as “aggressive” and claimed it was an attempt to “block innovation and make life worse for people”.

The company argued that Comet acts solely under user instruction and therefore should not be treated as an independent bot. “Your AI assistant must be indistinguishable from you,” it wrote. “When Comet visits a website, it does so with your credentials, your permissions, and your rights.”

Perplexity’s blog also accused Amazon of prioritising advertising profits over user freedom. It cited comments by Amazon CEO Andy Jassy, who recently told investors that advertising spend was producing “very unusual” returns, and claimed Amazon wants to restrict independent agents while developing its own approved ones.

Chief executive Aravind Srinivas added that Perplexity “won’t be intimidated” and that it “stands for user choice”. In interviews, he suggested that agentic browsing represents the next stage of digital personalisation, where users, not platforms, control their experiences.

Previous Allegations Against Perplexity

Amazon’s claims are not the first to question Perplexity’s web practices. For example, earlier this year, Cloudflare (a web infrastructure and security company) published research showing that Perplexity’s AI crawlers were accessing websites that had explicitly opted out of AI scraping. Cloudflare alleged that the company disguised its crawler as a regular Chrome browser and used undisclosed IP addresses to avoid detection.

Perplexity denied intentionally breaching restrictions and said any access occurred only when users specifically asked questions about those sites. However, Cloudflare later blocked its traffic network-wide, citing security and transparency concerns.

The startup is also facing ongoing lawsuits from publishers including News Corp, Encyclopaedia Britannica and Merriam-Webster over alleged misuse of their content to train its models. Together, those disputes portray a company pushing at the legal and ethical boundaries of how AI interacts with the web.

Why The Amazon Clash Matters

The dispute with Amazon is really shaping up as an early test case for how much autonomy AI agents will have across the commercial web. For example, Amazon maintains that any software acting on behalf of users must still identify itself, follow platform rules, and respect the right of websites to decide whether to engage with automated systems.

However, Perplexity argues that an AI assistant used with a person’s consent is part of that person’s digital identity and should have the same access as a regular browser session. The company believes restricting that principle could undermine the emerging concept of user-controlled AI and set back progress in agentic browsing.

For Amazon, the matter is tied to the customer experience it has spent decades refining, and one that depends on data visibility, targeted recommendations and carefully managed fulfilment. For AI developers, the case signals the likelihood of tighter scrutiny and the potential for conflict if agents interact with online platforms without explicit approval.

Businesses experimenting with autonomous procurement or digital assistants will also be watching closely. Tools that can buy or book on behalf of staff offer obvious productivity benefits, but only if those agents operate within clear contractual and technical limits.

Regulators are beginning to take interest too. For example, questions are emerging over where accountability lies if an agentic system breaches a website’s terms or handles personal data incorrectly, and whether users, developers or platforms should bear responsibility. How these questions are answered will influence how agentic AI evolves, and how openly such systems are allowed to participate in the online economy.

What Does This Mean For Your Business?

The outcome of Amazon’s confrontation with Perplexity will set a practical benchmark for how far autonomous AI agents can go before platforms intervene. What began as a dispute over one shopping assistant now touches the wider question of how digital power is distributed between users, developers and global platforms. If Amazon succeeds in forcing explicit disclosure and control over third-party agents, it could consolidate platform dominance and slow the development of independent AI tools. If Perplexity’s position gains support, the web could see a surge of user-driven automation that bypasses traditional commercial gateways.

For UK businesses, companies already exploring AI tools to handle purchasing, market research or logistics will need to ensure those systems act within recognised platform rules and data protection standards. The eventual precedent could shape how British firms integrate AI agents into supply chains, e-commerce systems and customer service platforms. It may also affect costs and compliance responsibilities, depending on whether platforms like Amazon begin enforcing stricter access requirements on all autonomous systems.

For consumers, the promise of convenience from agentic browsing is balanced by legitimate concerns about data security and transparency. For regulators, the case underscores the urgent need to clarify who is accountable when AI systems act independently. For AI companies, it highlights that technical innovation alone is no longer enough; transparent cooperation with platform owners and adherence to existing legal frameworks will now be part of the competitive landscape.

The Amazon–Perplexity dispute has, therefore, become more than a legal warning. In fact, it looks like marking the start of a global debate over how automation, commerce and trust can coexist online, and one that every business and policymaker will have to engage with as agentic AI becomes part of everyday digital life.

Microsoft has built a synthetic online marketplace to stress test AI agents in realistic buying and selling scenarios, but the early results appear to have revealed how fragile even the most advanced models remain when faced with complex, competitive environments.

Why Microsoft Built A Fake Marketplace

Magentic Marketplace is Microsoft’s new open source simulation environment for what the company calls “agentic markets”, where AI systems act as autonomous customers and businesses that search, negotiate and transact with each other. The project, developed by Microsoft Research in collaboration with Arizona State University, is designed to explore how AI agents behave when placed in a simulated economy rather than isolated single agent tasks.

The initiative reflects growing excitement across the tech sector about so-called agentic AI, systems capable of taking actions on a user’s behalf, such as comparing products, booking services or handling customer enquiries. Microsoft’s researchers argue that while such systems promise major economic efficiency gains, there is still little understanding of what happens when hundreds of agents operate simultaneously in the same market.

The Value of Studying AI Agents’ Behaviours

Ece Kamar, corporate vice president and managing director of Microsoft Research’s AI Frontiers Lab, has said that understanding how AI agents interact, collaborate and negotiate with one another will be critical to shaping how such systems influence real world markets. Microsoft describes the project as part of a broader effort to study these behaviours safely and in depth before agentic systems are deployed in everyday economic settings.

The work sits alongside a broader research programme at Microsoft exploring what it calls the “agentic economy”. The associated technical report, MSR-TR-2025-50, was published in late October 2025, followed by a detailed blog post and open source release on 5 November.

How Magentic Marketplace Works

Instead of experimenting with real online platforms, Microsoft built a fully synthetic two sided marketplace. One side features “assistant agents” representing customers tasked with finding products or services that meet specific requirements, for example ordering food with certain dishes and amenities. The other side features “service agents” acting as competing businesses, each advertising their offerings, answering questions and accepting orders.

The marketplace environment itself manages all the underlying infrastructure, from product catalogues and discovery algorithms to transaction handling and payments. Agents communicate with the central server via a simple HTTP/REST interface, using just three endpoints for registration, protocol discovery and action execution. This minimalist architecture allows the researchers to plug in a wide range of AI models and keep experiments reproducible.

The Experiment

Microsoft ran its initial experiments using 100 customer agents and 300 business agents. The test scenarios included synthetic restaurant and home improvement markets, allowing the team to control every variable and analyse outcomes in detail. The study compared a range of proprietary and open source models, including GPT 4o, GPT 4.1, GPT 5, Gemini 2.5 Flash, GPT OSS 20b and Qwen3 variants, and measured performance using standard economic metrics such as consumer welfare (the perceived value of purchases minus prices paid).

What Happened When Microsoft Let The Agents Loose

When given a simplified “perfect search” setup, where only a handful of highly relevant options were available, leading models such as GPT 5 and Anthropic’s Claude Sonnet 4.x achieved near optimal performance. In these ideal conditions they consistently selected the best options and maximised consumer welfare.

However, when Microsoft introduced more realistic challenges, such as requiring the agents to form their own search queries, navigate lists of results and choose which businesses to contact, performance dropped sharply. While most agents still performed better than random or cheapest option baselines, the advantage over simple heuristics often disappeared under realistic levels of complexity.

A Paradox of Choice Revealed

Interestingly, the study also revealed an unexpected “paradox of choice”. For example, when the number of search results increased from three to one hundred, most agents failed to explore the wider set of options. In fact, it was found that many simply picked the first “good enough” choice, regardless of how many alternatives existed. Also, consumer welfare fell as more results were shown, particularly for models like Claude Sonnet 4, which saw average welfare scores drop from around 1,800 to 600. GPT 5 also showed a steep decline, from roughly 2,000 to just over 1,000, suggesting that even large models struggle to reason across large decision spaces.

Collaboration Tested

The researchers also tested how well multiple AI agents could collaborate on shared tasks, such as dividing roles in joint decision making. Without clear instructions, most agents became confused about who should do what. When researchers provided explicit step by step guidance, performance improved, but Kamar noted that true collaboration should not depend on such micromanagement.

Manipulation, Bias And Behavioural Failures

One of the most striking findings came from experiments testing whether business side agents could manipulate their AI customers. Microsoft tested six tactics, ranging from standard persuasion techniques such as fake credentials (“Michelin featured” or “award winning”) and social proof (“Join 50,000 happy customers”) to more aggressive prompt injection attacks that directly tried to rewrite a customer agent’s instructions.

The results varied widely between models. For example, Anthropic’s Claude Sonnet 4 resisted all manipulation attempts, while Google’s Gemini 2.5 Flash showed mild susceptibility to strong prompt injections. By contrast, GPT 4o and several open source models, including Qwen3 4b, were easily compromised, with manipulated businesses successfully redirecting all payments towards themselves. Even subtle tactics such as fake awards or inflated review counts could influence purchasing decisions for some systems.

These findings appear to highlight a broader concern in AI safety research, i.e., that large language models are easily swayed by adversarial inputs and emotional framing. In a marketplace context, such weaknesses could enable dishonest sellers to exploit customer side agents and distort competition.

Bias

The experiments also appear to have uncovered systemic biases in agent decision making. For example, across all tested models, agents showed a strong “first proposal” bias, accepting the first seemingly valid offer rather than waiting for additional responses. This behaviour gave a ten to thirty fold advantage to faster responding sellers, regardless of quality. Some open source models also displayed positional bias, tending to pick the last option in a list regardless of its actual merits.

Together, these findings seem to suggest that agentic markets could replicate and even amplify familiar real world problems such as information asymmetry, bias and manipulation, only at machine speed.

Microsoft And Its Competitors

Microsoft is positioning itself as a leader in agentic AI, building Copilot systems that can act semi autonomously across Office, Windows and Azure. However, publishing this research about Magentic Marketplace that exposes major limitations in current agent behaviour shows not just scientific transparency, but also an acknowledgement that current systems remain brittle.

At the same time, releasing Magentic Marketplace as open source code on GitHub and Azure AI Foundry Labs gives Microsoft significant influence over how the next phase of AI evaluation is conducted. The company has effectively created a public benchmark for testing AI agents in market like environments. This may shape how regulators, researchers and competitors such as Google, OpenAI and Anthropic measure progress towards safe deployment.

It is worth noting here that the agentic AI race is on and competitors are pursuing their own versions of agentic systems, from OpenAI’s Operator tool, which can perform real web tasks, to Anthropic’s Computer Use feature, which controls software interfaces on behalf of users. None has yet published a similarly large scale testbed for multi agent markets. Industry analysts suggest that Microsoft’s decision to expose failures so openly may also be strategic, helping the company frame itself as a responsible actor ahead of tighter global regulation on AI autonomy.

Businesses, Users And Regulators

For businesses hoping to integrate agentic AI into procurement, sales or customer support, the message from this research is that these systems still require close human supervision. Agents proved capable of making simple transactions but were easily overloaded by large product ranges, misled by false claims and prone to favouring the first acceptable offer. In high stakes contexts such behaviour could lead to financial losses or reputational harm.

The findings also raise new competitive and ethical questions. For example, if agentic marketplaces reward speed over accuracy, or if certain models are more vulnerable to manipulation, companies that optimise for aggressive tactics could gain unfair advantages. Microsoft’s economists warned that such structural biases could distort future digital markets if left unchecked.

For regulators, Magentic Marketplace offers a rare tool to observe how autonomous agents behave before they enter real economies. The ability to run controlled experiments on transparency, bias and manipulation could inform emerging AI safety standards and consumer protection frameworks.

Challenges And Criticisms

While widely praised for its openness, the Magentic Marketplace research has also drawn some scrutiny. For example, the test scenarios focus mainly on low risk domains like restaurant ordering, which may not reflect the complexity or stakes of sectors such as healthcare or finance. Also, because the data is fully synthetic, it avoids privacy issues but may underrepresent the messiness and unpredictability of human driven markets.

The current experiments also study static interactions rather than dynamic markets, where agents learn and adapt over time. Real economies evolve as participants change strategy, something Microsoft plans to explore in future iterations. Some researchers have also pointed out that focusing mainly on “consumer welfare” may overlook broader measures of fairness, accessibility and long term market stability.

That said, at least the findings so far give researchers a clearer view of how AI agents behave when placed in competitive settings. Microsoft’s approach could also be said to provide a fairly structured way to observe these systems under controlled market conditions and to identify where improvements are most needed before they are applied more widely in real commercial use.

What Does This Mean For Your Business?

For all the progress in developing intelligent assistants, Microsoft’s Magentic Marketplace experiment has exposed how far current AI models still are from handling the unpredictability of real markets. The failures observed in decision making, collaboration and manipulation resistance point to weaknesses that could directly affect trust and reliability if similar systems were deployed commercially. For UK businesses exploring automation through AI agents, this research is a reminder that the technology is not yet capable of making independent purchasing or negotiation decisions without oversight. The risks of bias, misjudged choices and exploitability remain significant.

At the same time, the study shows why testing environments like Magentic Marketplace will be vital for regulators, developers and investors as agentic AI moves closer to practical use. For example, controlled simulations can reveal hidden biases and security flaws before these systems handle real financial transactions. For policymakers in the UK and elsewhere, the findings reinforce the need for standards that ensure accountability and human control within automated decision systems.

For Microsoft, this project strengthens its image as a company willing to expose and study AI limitations rather than conceal them. For its competitors, the research sets a benchmark for transparency and evaluation that others will be expected to meet. For businesses and public institutions, it highlights the importance of using AI agents as supportive tools rather than autonomous decision makers until reliability, fairness and resilience can be proven in real economic conditions.