AI Commerce Status Report Q1 2026

Reality, Readiness and the Gap Between Hype and Execution

This piece is intended as a status report rather than a manifesto. It draws solely on publicly documented and verifiable initiatives and is meant to provide a calm, factual reading of where AI in a commerce context actually stands at the start of 2026, as opposed to where keynote slides would prefer it to be.

Executive Summary

By early 2026 artificial intelligence in digital commerce has comfortably outgrown the speculative phase. It now sits inside numerous initiatives, pilot programmes and a respectable number of genuinely productive applications. The accompanying public discourse, by contrast, has not aged quite as gracefully. The vocabulary of the moment, "Agentic Commerce", "AI-First", "Autonomous Shopping", is deployed with admirable confidence and rather less precision, often blurring the line between what is shipping today, what is being built, and what remains a slide in someone's strategy deck.

This report tries to put that line back where it belongs. It covers only publicly documented and verifiable use cases, alongside initiatives whose technical maturity is described in terms one can actually scrutinise. Hypothetical scenarios and uncosted visions are left, politely, for another occasion.

The picture that emerges is consistent. Several productive building blocks now exist that tend to be filed under "Agentic Commerce", chiefly conversational interfaces, in-chat checkout flows, and new commerce surfaces accessible through AI platforms. These mark a structural shift in how customers interact with commerce systems. They do not, however, amount to autonomous purchasing agents operating at scale. The agents most often invoked in conference programmes are still, for the most part, somewhere between proof of concept and press release.

The applications in production today rest almost entirely on explicit human intent. AI supports search, selection and checkout within new surfaces, but does not make purchasing decisions on its own account. The genuine progress lies less in machine autonomy than in the decoupling of interface from commerce logic. API-based architectures allow new modes of interaction to be plugged into existing transaction systems without rebuilding the underlying decision-making fabric.

Publicly disclosed initiatives from major retail and fashion businesses underline this. AI platforms are increasingly used as additional sales channels or as dialogue-led experience layers. In parallel, large groups continue to invest in AI for decision support across product development, demand forecasting, inventory and marketing. Useful, scalable, occasionally transformative, but pursuing rather different objectives from the vision of fully autonomous commerce agents.

What is striking is that the bulk of the current momentum is happening not at the customer touchpoint but at the level of plumbing. Across the industry, work is under way on protocols, standards and authorisation mechanisms intended to let AI systems execute transactions safely and within the law. Payment providers, platform operators and the larger marketplaces are quietly investing in identity, fraud detection and approval workflows. These are preconditions for greater autonomy rather than autonomy itself, and they tend not to feature on opening keynote slides.

The contrast between rhetoric and operational reality is hard to miss. Where keynotes and op-eds favour a binary in which one is either "AI-First" or shortly to be irrelevant, the documented use cases tell a more textured story. Progress is incremental, scoped to clearly defined applications, and shaped by regulatory, organisational and technological constraints that have not been suspended for the occasion.

The sober conclusion, for organisations actually running commerce businesses, is that the central task is not to deploy autonomous agents but to build the conditions under which new forms of interaction can work at all. Consistent data models, API-ready systems, intelligible decision logic, and governance arrangements that make AI deployment controllable, all of it less photogenic than launching an "agent", and considerably more useful.

The state of AI Commerce in the first quarter of 2026 can therefore be summarised without much drama. The direction is visible, the infrastructure is being built, and the autonomy on offer is bounded. The difference between successful and unsuccessful initiatives lies less in speed than in the ability to distinguish between what is verifiable, what has been announced, and what is still, for now, a hopeful projection.

This report is offered as a snapshot. It documents the current position, situates the more interesting initiatives, and aims to provide a base for substantive conversation that does not require either alarm or oversimplification.

Clarifying Terms: What "Agentic Commerce" Currently Means

In early 2026 the term "Agentic Commerce" is used with notable enthusiasm and very little consistency. In debates, product launches and keynote speeches it stretches to cover everything from simple conversational interfaces to visions of autonomously acting AI agents. That elasticity has made objective assessment of progress unusually difficult and has done a great deal to inflate the perceived significance of individual initiatives.

A more dependable reading requires breaking the term into functional layers. These should not be read as a maturity model but as analytical categories describing the role AI actually plays in a given commerce context.

At the lowest layer is agent-assisted interaction. Here AI functions as a conversational interface through which users search, compare or retrieve information about products. Decision-making remains entirely with the human. AI does not select; it structures information and reduces complexity. Most of the production-grade commerce applications visible today belong here.

Above that sits agent-assisted selection. AI proposes products based on context, preferences or earlier interactions. Selection is prepared rather than executed autonomously, and explicit user consent remains a precondition for anything that follows. Technically, this is generally a recombination of recommendation systems with conversational front-ends.

Then comes agent-integrated checkout. The entire purchasing process unfolds inside the AI interface, with search, selection and payment occurring without redirection to a conventional shop. AI orchestrates the flow without making independent purchasing decisions. The transaction is initiated by the user. Several of the more visible current implementations sit at this layer.

Quite distinct from these is agent-based decision support in back-office operations, where AI is used to prepare or simulate internal decisions in pricing, demand forecasting, assortment planning or marketing orchestration. These applications have a direct effect on commercial outcomes, but they are not part of the purchasing dialogue with the customer. They are organisationally and strategically important, and they are routinely conflated with agentic commerce.

Only at a further layer, one with very few publicly documented production examples, would autonomous agentic execution begin. In such a setting AI would not merely advise or orchestrate but would independently choose products and trigger transactions. The preconditions, properly defined authorisation mechanisms, legally robust accountability and very high system reliability, are precisely the conditions that the current generation of deployments is still being built to satisfy. As of early 2026, scaled production examples of this form of autonomy are essentially absent.

The conflation of these layers means that very different developments end up being discussed under a single label. An in-chat checkout is implicitly equated with autonomous purchasing agents, even though the two differ fundamentally in objective, risk and complexity. This leaves businesses with the impression that decisions must be taken today which are technically, organisationally and regulatorily not yet ready to be made.

This report therefore uses the term "Agentic Commerce" carefully and with some precision. Throughout the chapters that follow, it is made explicit at which layer a given use case operates and whether it represents production deployment, a limited rollout or an announced intention. The aim is to make initiatives genuinely comparable rather than uniformly impressive.

With the conceptual ground prepared, the next chapter turns to what can actually be observed in the wild.

What Is Publicly Verifiable as Production-Ready in Early 2026

Once the terminology is in order, the central question becomes which forms of AI-supported commerce are demonstrably in production in early 2026. The relevant distinction is between live functionality, regionally limited rollouts and announcements still searching for a deployment date. This chapter restricts itself to applications whose deployment is either already in production or whose mechanics are transparently described.

One clearly verifiable area is transaction completion within conversational interfaces. Several major platform providers now allow users to discover and purchase products inside a dialogue, without opening a traditional shop interface. The flow remains user-driven. AI structures the interaction, guides the process and removes friction without making independent decisions.

Among the better-known examples is in-chat checkout inside AI assistants, where users can ask about products, compare offers and complete a purchase without leaving the conversation. Selection is explicitly confirmed and payment is handled through integrated flows. The functionality is intentionally bounded, often to specific products or selected merchants, less because the technology cannot do more than because the regulatory, legal and security framework would prefer that it did not, just yet.

A defining feature of these implementations is that they ride on top of existing commerce and payment infrastructure. AI acts as an orchestrating surface, not as a standalone commerce system. Product data, pricing, availability and transactions remain the responsibility of conventional backends. The advance is in the seamless integration of those backends into new modes of interaction.

A second, closely related area is the use of conversational interfaces for product selection without an embedded purchase mechanism. Many retailers deploy AI-led dialogues to help users navigate ranges, surface recommendations or unpack complex product information. These applications go beyond conventional search and filters without entirely replacing them. They sit alongside existing structures, particularly where the user's needs are not yet sharply defined.

Here too, decision-making remains with the human. AI aggregates information, asks clarifying questions and offers options. The transition from suggestion to decision is explicit. In practice, this is often the next iteration of established recommendation systems, augmented with natural language and contextual processing.

Alongside these customer-facing applications there is a third category of production use, frequently grouped under the Agentic Commerce banner despite pursuing rather different objectives. This is AI-supported decision support in internal commerce processes. Companies use AI to simulate pricing decisions, forecast demand, optimise assortments and prioritise marketing investment. The commercial impact is direct, and the systems do not operate autonomously in any meaningful sense of the word.

Such internal systems are often more advanced than public discussion would suggest. They are productive, measurable and embedded in existing decision-making processes. Their scope, however, is clearly bounded. They prepare decisions; they do not take them. Responsibility and approval remain with people.

What stands out across all of these productive applications is a common discipline. They operate within explicitly defined limits. Human intent is always required. Authorisation, responsibility and liability are unambiguously assigned. These boundaries are not the absence of innovation but the result of conscious risk management and an alignment with what customers and regulators reasonably expect.

The picture is consistent. By early 2026 several productive building blocks now exist which together support real structural change in commerce. That change is primarily about how users interact with commerce systems, not about how autonomously those systems decide. Today's visible applications shift the surface; the underlying decision architecture is not yet being rewritten.

The next chapter takes this further by looking at concrete public initiatives that treat AI-based interfaces as new commerce channels and at the architectural conditions those channels require.

Channel Shift Rather Than Autonomy: AI as a New Access Point to Commerce

The applications described above show what is productively possible in early 2026. The next question is how those possibilities are being put to structural use. This becomes most visible in initiatives that treat AI-based interfaces not as ancillary features but as standalone access points to commerce. In such cases, the locus of interaction shifts even though the fundamental decision logic of the transaction is left intact.

At the heart of this development is the decoupling of interface from commerce backend. Traditional e-commerce architectures were tightly bound to specific frontends. A website, an app or a marketplace defined how products were presented, selected and bought. AI-based interfaces break that binding. They sit as a new interaction layer on top of existing commerce systems without replacing their core logic.

A prominent example is the initiative by JD Sports Fashion with commercetools and connected payment partners. The objective is to allow customers to discover and purchase products directly within AI platforms. The AI environment takes on the role of the frontend. Product data, pricing, availability and transaction processing remain entirely with the commerce backend.

A more recent initiative makes a related but architecturally distinct development visible. The Debenhams Group, whose portfolio includes PrettyLittleThing, boohoo and Karen Millen, announced in mid-February 2026 a partnership with PayPal to offer an AI-powered shopping assistant directly inside the PayPal app. Customers can discover products, receive personalised recommendations and complete the purchase without leaving the app. The assistant is initially available to selected users in the United States, with a broader US and UK rollout planned for later in 2026.

What is interesting here is less the AI itself than the location of the interaction. Where JD Sports positions the AI platform as a new frontend over a commerce backend, this case sees the payment platform itself become the gateway to commerce. PayPal is moving from payments service provider to a venue where discovery, advice and transaction converge. The company has created a dedicated organisational function for this and is happy to label it "Agentic Commerce".

A degree of nuance is in order. The assistant is dialogue-led, it recommends products and it guides the user through checkout. The purchasing decision still belongs to the user. This is conversational commerce inside an existing app, not an autonomous agent making purchases on its own initiative. In this context, "Agentic Commerce" describes the channel rather than the degree of autonomy. That is a perfectly reasonable position; it is simply not the one the label tends to imply.

In both cases, the architecture makes clear that this is not a step towards autonomous AI agents but an extension of the channel model. AI platforms become additional touchpoints, comparable to mobile apps or marketplaces. The difference is the texture of the interaction. Instead of navigation, filters and product listings, an intent-led dialogue takes over.

The purchasing process still requires explicit consent. AI suggests, structures and guides. Selection and completion are deliberate human acts. That boundary is technical, legal and organisational by design. It keeps responsibility, liability and control unambiguously assigned.

The architectural value of these initiatives lies less in the AI itself than in the consistency of an API-first orientation. Only systems that expose their core functions cleanly through interfaces can be integrated into new modes of interaction. The channel shift turns out to be a useful diagnostic for which commerce stacks are ready for the next decade and which begin to creak as soon as the frontend changes shape.

It is also notable that these approaches tend to be regionally and functionally constrained. Most of the known implementations launch in selected markets with bounded functionality. That is not a lack of ambition but a recognition that payment processing, fraud prevention, regulation and user expectations vary considerably from one market to the next. A simultaneous global rollout is the exception, not the rule.

Public coverage of these initiatives is often inflationary. Phrases such as "AI-Powered Shopping" or "Agentic Commerce" suggest a clean break with established models. In practice, the step is incremental, even if it is strategically significant. The way users engage with commerce systems is changing. The decision-making mechanisms underneath are not.

For organisations, this distinction matters. Reading these developments as evidence of an immediate need for autonomy risks setting the wrong priorities. The actual task is not to deploy autonomous agents but to build the architectural conditions that make new access points possible. That means consistent data models, dependable pricing and inventory logic, and robust interfaces.

The channel shift is genuine progress in digital commerce. It shows how commerce can adapt to new modes of interaction without abandoning its foundations. Autonomy remains a prospect, not yet the present condition.

The next chapter adds a further dimension by looking at companies that deploy AI not primarily as a new sales channel but as a means of redesigning digital experiences and internal decision-making.

Dialogue-Based Experiences and Organisational AI: Two Distinct Development Paths

Beyond the channel shift discussed above, there are further approaches in which AI plays a central role in commerce without functioning as an independent sales channel. These approaches pursue different objectives and address different layers of value creation. The contrast is most visible in initiatives focused either on the design of digital experiences or on the support of internal decision-making.

A current example of the first path is the e-commerce initiative recently unveiled by the Italian luxury house Brunello Cucinelli. With the new platform developed internally under the project name Callimacus, the company adopts a dialogue-led mode of interaction that consciously moves traditional navigation and filter structures to the back of the stage. Users engage with the platform in natural language and receive curated suggestions oriented around context and occasion rather than around formal product categories.

Here AI is used primarily to translate intent into experience. The objective is not to automate or delegate decisions but to make access to the range feel closer to ordinary human conversation. The purchase remains fully under conscious control. AI acts as a mediating layer rather than an active agent. In the luxury segment in particular, that restraint is part of the proposition. Autonomy in this context would not be perceived as added value but as a loss of control and a softening of brand identity.

Brunello Cucinelli illustrates an important point. AI is not always deployed to accelerate or automate. In certain contexts, it serves to refine, personalise and emotionally enrich the digital experience. Productive, visible and strategically anchored, but materially different from the visions of agentic autonomy that tend to dominate the headlines.

A second, clearly distinct path is visible in large enterprise initiatives such as those announced by the fashion conglomerate PVH. The focus here is not the customer dialogue but the systematic integration of AI into internal processes. The aim is data-driven decision support across the value chain, from design and product development through demand forecasting and inventory management to marketing orchestration.

These applications have substantial impact. They influence assortments, pricing and availability and therefore indirectly affect commerce outcomes. At the same time, they sit outside the purchasing dialogue itself. AI analyses, simulates and recommends. Decision-making remains with people. Autonomy is once again deliberately bounded.

Public discourse is quite content to bundle both paths under the same heading. Dialogue-led experience platforms and internal decision support are routinely cited together as examples of Agentic Commerce. That conflation does neither justice. Both draw on similar technological foundations, but they address different problems on different planes.

What they share is a design optimised for maturity rather than speed. Both initiatives presuppose that data, processes and accountabilities are clearly defined. They are not the product of an "AI-First" doctrine but of concrete strategic questions. In each case AI is used where it supports the logic of the business, not where it tries to overtake it.

Taken together, these two paths show that the current state of AI in commerce is shaped less by a single objective than by context-dependent decisions. Companies are choosing very deliberately what role AI should play. Brand, sector, audience and risk appetite all influence those choices.

The next chapter explains why this deliberate restraint is not merely caution but a response to structural conditions that must still be put in place before more advanced autonomy becomes responsible.

Infrastructure Before Autonomy: Why Protocols, Payments, and Governance Are Decisive

The use cases discussed so far make clear that the current limits of AI in commerce are determined less by models or compute than by infrastructural prerequisites. Autonomous agentic systems in early 2026 are not failing because they are not clever enough. They are bumping up against questions of authorisation, responsibility, security and standardisation.

Commerce is not a purely informational system. It is transactional, legally bound and risk-laden. Every purchasing decision implies payment, liability, returns, fraud prevention and regulatory compliance. None of that is solved by larger models. It requires clear rules, reliable protocols and controllable interfaces.

Against that backdrop, it is unsurprising that the bulk of current momentum sits not at the application layer but in the foundations. Platform providers, payment processors and large marketplaces are investing in the standards that would allow AI systems to interact safely with commerce systems in the first place.

A central topic is the authorisation of action. Where human users make purchasing decisions consciously and bear unambiguous legal responsibility, the equivalent attribution for AI systems is genuinely complex. Who is responsible if an agent makes the wrong call? Who authorises payment, and on what conditions? Which interventions are permitted and which are not? These are not theoretical questions. They are preconditions for any productive form of autonomy.

A second bottleneck is in payments and fraud prevention. Payment providers are working on the rules under which AI-powered systems may trigger transactions: multi-stage approvals, limits, identity checks, continuous monitoring. These mechanisms must work reliably before autonomous systems can operate at scale. The cautious tempo of the current rollouts reflects exactly that.

In parallel, technical protocols are emerging that aim to enable standardised communication between AI systems and commerce infrastructure. Initiatives such as the Universal Commerce Protocol intend to provide shared interfaces for product data, availability, pricing and transactions. These efforts are less visible than customer-facing features but strategically decisive. Without standards, autonomy remains fragmented and structurally unscalable.

A further dimension is internal governance. Autonomous systems require explicit rules about which decisions may be delegated and which may not. These rules are not purely technical; they are organisational. They concern accountability, escalation and control. The companies working productively with AI today are quietly investing significant resources in those structures, often well before they expose any new external capability.

What this trajectory shows is that the infrastructural agenda is not being treated as a side project. It sits at the centre of the activity of the most credible market participants. It also explains why so many publicly aired visions remain unrealised. Autonomy is not a feature that can be introduced in isolation. It only becomes possible when the legal, technical and organisational frameworks line up.

For the purposes of assessing the status quo, the absence of autonomous agentic commerce systems is not a failure of innovation. It is a sign of deliberate progress. The industry is putting in place the conditions under which more advanced autonomy could be pursued responsibly. That work is complex and time-consuming, and entirely necessary.

The next chapter takes this insight and uses it to read the public discourse, asking why simplifying narratives such as "AI-First or lose relevance" do not do justice to the actual situation, and what risks they create.

Rhetoric Versus Reality: Why Public Discourse Is Distorted

The analysis so far offers a coherent reading of what is actually happening in AI-supported commerce. That reading sits in marked contrast to the rhetoric shaping public discourse in early 2026. In a great many articles, keynotes and commentaries, a binary narrative dominates: organisations must reorient themselves immediately and comprehensively around an "AI-First" strategy, or expect to lose relevance in the foreseeable future.

It is an understandable framing. It is also insufficient. It collapses different stages of development into one, ignores structural prerequisites, and generates pressure to act that rarely produces better decisions. The sense of existential threat owes less to verifiable market developments than to a tidy compression of future scenarios into the present tense.

A defining feature of this rhetoric is the equating of direction with current position. Scenarios that follow plausibly from technology trends are reported as already achieved reality. Dialogue-based interfaces become autonomous agents, pilot projects become industry-wide standards, announcements become mass-market deployments. The shift in tense is subtle but powerful. It reshapes the perception of what is necessary and how urgently.

To this is added the overuse of labels. Terms like "Agentic Commerce", "AI-Powered Shopping" and "Autonomous Retail" increasingly serve as catch-all phrases for radically different approaches. They generate attention while obscuring meaningful differences in maturity, risk and effect. As the use cases above show, what these labels often describe are architectural extensions or new interfaces that explicitly avoid implementing autonomy at all.

This kind of simplification has consequences. Companies come under pressure to launch initiatives without clearly defined objectives, benefits or organisational prerequisites. The question shifts from where AI can be meaningfully deployed to how quickly visible activity can be generated. The risk of reactive action goes up; strategic clarity goes down.

It is also notable that the discourse rarely distinguishes between sectors, business models and target audiences. Developments relevant to global platforms or highly standardised consumer goods are applied without amendment to mid-sized organisations or niche providers. The structural differences are quietly disregarded. Autonomy being painstakingly prepared in a heavily regulated enterprise environment appears in the discourse as a universal imperative.

Real development proceeds with considerably less drama. Companies actually working productively with AI tend to pursue incremental approaches. They experiment within clearly bounded use cases, evaluate results and adapt their structures gradually. Progress arrives where technology is embedded in existing decision logic, not where it tries to leap over it.

The contradiction between rhetoric and reality is therefore less a sign of missing innovation than an expression of divergent incentives. Public attention follows different mechanics from operational execution. Discourse rewards pointed messaging; productive systems require patience, precision and restraint. That tension is one of the defining features of the current state of AI in commerce.

A sober reading therefore requires separating the alarmist tone from the actual position. The state of early 2026 is marked neither by stagnation nor by imminent disruption. It is marked by structural development, bounded production readiness and clearly defined limits. Naming those limits is not a refusal to engage; it is a precondition for sound decisions.

The concluding chapter draws practical implications from this assessment. They are not operational prescriptions but guiding principles to help organisations distinguish between hype, expectation and actual maturity.

AI and Agentic Commerce Initiatives at a Glance (As of Q1 2026)

Before drawing final conclusions, a complementary look at concrete initiatives publicly visible in early 2026 is useful. The list below makes no claim to capture every relevant programme. It is updated regularly and is meant as a curated orientation rather than an exhaustive market survey.

Zalando uses AI to analyse user behaviour, fit data and return patterns in order to generate personalised recommendations and outfit suggestions inside the purchasing flow. The rollout has been running since 2024/2025, covering Germany, Austria, Switzerland, Benelux, Scandinavia, and gradually France, Italy and Spain. (Source: TechCrunch, November 2025)

Shopify offers a suite of AI features under the name "Shopify Magic", including automated copy generation for product descriptions, segment predictions and AI-supported recommendations at checkout. The rollout began in 2024 and is now largely active worldwide. With the Winter Edition 2026, Shopify substantially evolved its AI assistant Sidekick from a reactive chatbot into a proactive "Agentic Copilot" that analyses business data and surfaces prioritised actions as Sidekick Pulse directly inside the admin. (Sources: Shopify Newsroom, Shopify Winter Edition 2026)

Amazon combines automatic checkout systems ("Just Walk Out"), AI-based smart trolleys ("Dash Cart") and the AI shopping assistant Rufus. Rufus has reached around 250 million users by early 2026 and, on Amazon's own figures, contributes roughly $10 billion in annual sales growth. Recent additions include the AutoBuy function, which lets users set target prices and have Rufus complete a purchase automatically when the threshold is met. Rufus can now also shop across third-party merchants. (Sources: About Amazon, Omise AI, January 2026)

IKEA combines AR visualisation with AI-led product recommendations inside the shopping context through "IKEA Place". The app has offered AR since 2017, with AI enhancements added in 2024/2025. Active markets include the US, Germany, the UK, France, Scandinavia, Canada and Australia. (Source: IKEA Place App)

Klarna integrates AI-driven product recommendations, dynamic offer presentation and behaviour-based trigger signals inside the shopping flow. Continuous expansion has been under way since 2023/2024 across Sweden, Germany, the UK, the US, the Netherlands, France and Spain. (Source: Finextra)

Lowe's uses VisionAI for visual product search, category navigation and AI-led discovery assistants. The technology has been in use since 2021, with extensions in 2024/2025. The primary market is the US, with selected features also available in Canada. (Source: Lowe's Newsroom)

H&M uses AI for personalised product recommendations based on user behaviour, session data and micro-segmentation. Continuous expansion has been under way since 2022/2023 across Europe (notably Germany, Sweden and the UK), the US and selected Asian markets. (Source: FashionUnited)

JD.com uses AI for visual product recognition, automated recommendations and AR support across cross-channel touchpoints. Broad integration took place in 2024/2025, focused on China and Southeast Asia through partner setups. In March 2026, JD.com announced the construction of what it describes as the world's largest Embodied Intelligence Data Collection Center, drawing on its supply chain footprint across retail, logistics, healthcare and food delivery to train AI models for physical interaction. (Sources: Reuters, Gasgoo, March 2026)

Additions March 2026

Microsoft Copilot Checkout & Brand Agents were unveiled at NRF on 8 January 2026 and have been rolling out since March 2026. Copilot Checkout lets users discover, compare and purchase products inside the AI assistant without leaving the Copilot interface. Brand Agents allow brands to deploy their own AI representatives inside Copilot, intended to turn conversations into conversions. With this, Microsoft positions itself as a meaningful additional player in the agentic commerce arena alongside OpenAI and Google. (Source: Microsoft Advertising, January 2026)

Perplexity "Buy with Pro" has offered a one-click purchasing function inside the AI search engine since late 2025. In February 2026 Perplexity expanded the shopping experience further while simultaneously stepping away from advertising as a business model. For merchants, that creates a new discovery and conversion channel inside an AI-native search interface. (Source: AdsX, February 2026)

Alibaba International "Accio Work" was unveiled on 23 March 2026. The platform deploys enterprise AI agents as "virtual employees" handling tasks such as market analysis, design, sourcing, inventory monitoring and online shop creation. According to Alibaba, functional shops can be built in 30 minutes. The initiative targets primarily small and medium-sized enterprises in cross-border commerce. (Sources: SCMP, March 2026, TechNode, March 2026)

Otto Group: AI Assistants, Logistics Robotics and Content Platform. Germany's largest online retailer OTTO unveiled two AI assistants on 5 March 2026: a conversational shopping assistant for product discovery, and an automated customer service agent. Both operate inside the OTTO platform, complementing existing search and navigation with dialogue-led interaction. In parallel, the Otto Group announced a collaboration with NVIDIA in January 2026 to advance AI-driven robotic coordination in warehouse logistics, with the aim of intelligently orchestrating autonomous systems across fulfilment centres. The Otto Group also now offers its AI content platform Movex Virtual Content Creator (VCC) to external retailers. According to Otto Group, the platform delivers average cost savings of 60 per cent, and up to 90 per cent in fully automated scenarios. (Sources: E-Commerce Germany News, March 2026, Otto Group Newsroom, January 2026, Retail Optimiser)

Assessment

The initiatives above show that AI in e-commerce already extends well beyond personalisation and is increasingly used for decision support. Commerce flows are not merely accelerated; they are structurally altered. Checkout and purchasing decisions now sit alongside assistive, in some cases agentic, components.

That said, Agentic Commerce in the strict sense, namely the autonomous purchasing decision on behalf of the user, is still rarely fully operational. What can be observed today is a systematic move towards self-learning, context-sensitive assistance mechanisms that occupy the space between personalisation and semi-autonomous interaction.

While this report focuses specifically on AI adoption inside e-commerce environments, many of the observed differences in implementation maturity reflect broader structural conditions that extend well beyond the retail sector. A cross-industry structural analysis of those drivers, including infrastructure density, capital logic, governance architecture and data maturity, is set out in our research paper Beyond Geography: The Structural Drivers of Productive AI Adoption, published on ThePromptMonk.

Conclusions Q1 2026: Orientation Rather Than Alarmism

The reading offered in this report resists tidy summary. By early 2026 one can speak neither of a breakthrough in autonomous commerce agents nor of stagnation. The market is in a phase of structural preparation in which the conditions for later development are quietly being put in place.

The applications in production focus on clearly bounded functions. Conversational interfaces support orientation and selection. Integrated checkout flows reduce friction. AI-supported systems guide internal decision-making across the value chain. In each case, the human remains the decision-maker. Autonomy is not being replaced; it is being prepared.

That reality stands in contrast to a discourse that often prefers speed over assessment. The notion that organisations must reorient themselves wholesale around an "AI-First" logic in the short term misjudges the real dependencies. Neither technology nor organisation can be transformed in isolation. Progress arrives where new capabilities are integrated into existing structures, not where they try to leap past them.

A few overarching insights follow from the use cases observed. First, architecture matters more than individual features. Companies with consistent data models, clean interfaces and clear accountabilities can absorb new modes of interaction without compromising stability. Second, governance is not an obstacle but a precondition. The deliberate limitation of autonomy builds trust and enables scale. Third, context is decisive. Industries, brands and business models differ considerably in the role AI can credibly play.

For decision-makers, the present moment is therefore characterised less by pressure to act than by a need to assess. The central question is not how quickly autonomous systems can be introduced, but which forms of support are sensible today and which preconditions are still missing. In many cases the greatest leverage lies less in new features than in clarifying objectives, processes and accountability.

The status of AI Commerce in the first quarter of 2026 is best described as a transitional phase. The direction is visible, the infrastructure is being built, and the autonomy on offer is bounded. Companies that use this phase to lay foundations and build experience will be in good shape for what follows, without being driven by short-term narratives.

This report is offered as a snapshot. It does not claim completeness and cannot replace individual judgement. It does suggest that between hype and reality there is a space in which thoughtful decisions remain entirely possible. Orientation arises where developments are assessed soberly. Alarmism, by contrast, narrows the field of view.

Looking towards the coming quarters, what matters is how infrastructure, regulation and organisational maturity evolve. Only in their interplay will it become clear which forms of agentic systems can move beyond a supporting role. Until then, restraint is not a disadvantage but an expression of strategic clarity.

A continuation of this assessment is provided in the AI Commerce Status Report Q2 2026, which documents developments between April and May 2026. An overview of all reports published so far is available on the Status Reports hub.

Andreas Rieger

Digital Commerce & AI Consulting

As of: Q1 2026. This report represents a point-in-time assessment and will be updated as significant developments occur.

Glossary

API-first An architectural approach in which all functions of a system are primarily exposed through programmatic interfaces.

Agentic Commerce Commerce scenarios in which AI agents research, compare and initiate transactions on behalf of consumers or businesses, with degrees of autonomy ranging from assisted selection to independent purchasing decisions.

In-Chat Checkout Purchase completion within a conversational AI interface without redirection to a conventional shop website.

UCP (Universal Commerce Protocol) An open standard initiated by Google for communication between AI agents and commerce systems, featuring dynamic capability negotiation between agents and merchants.

Sources

¹ CNBC: "OpenAI revamps shopping experience in ChatGPT after instant checkout" (March 2026).

² PYMNTS: "Walmart Shuts Down Agentic Commerce With OpenAI" (March 2026).