Webinar recap: agentic commerce is changing discovery, not checkout ownership

Agentic commerce arrived faster than most of the industry expected. In the space of about eighteen months, AI agents went from answering shopping questions to acting on them: browsing, comparing, adding to basket and, increasingly, checking out, all on a shopper's behalf rather than at their hand.

Perplexity built an entire browser, Comet, around the idea. Visa and Mastercard have spent the past year building payment rails specifically so agents can transact without a human present at the point of sale. What was a research project two years ago is now a live capability, and retailers across every category are racing to work out what it means for them.

Then Amazon sued Perplexity AI, accusing Comet of disguising AI agents as human shoppers to buy on its site. The case has been read as the opening shot in a war over who controls the checkout button when AI does the shopping, and it has crystallised a fear now spreading through merchant boardrooms: that agents are about to cut retailers off from their customers, the way aggregators once threatened to cut airlines and insurers off from theirs.

That's the anxiety PSE Consulting set out to explore. Across more than 4,250 interviews with consumers in the UK, the US, France, and Germany, all of them active users of AI for online shopping, we tested what people are actually doing, not what the industry assumes they're doing.

We discussed our findings in The Paypers' latest webinar on agentic commerce, ‘The Shortlist Economy - How AI Is Rewiring Buying Habits (and How Merchants Can Respond)’, with the full findings to follow in a whitepaper of the same name.

The panel for the webinar, which took place on July 28, included Chris Jones, Managing Director at PSE Consulting (moderator); Christopher Uriarte, Partner at Glenbrook Partners; Andrew O'Connor, Senior Payments Consultant at PSE Consulting; Yasemin Merve Inceer, Director of Product Development at Mastercard; and Saad Saeed, Founder and CEO of Layla.ai.

Here are some of the key takeaways.

The window, not the shop

Let’s begin with the fundamentals: why are people using AI to shop at all? Overwhelmingly, for price. 32% of consumers cite price as their primary reason for choosing between AI recommendations, more than double the 14% who simply defer to the AI's top pick. The ratio is instructive. Most consumers are treating AI assistants as comparison engines, not as proxies empowered to decide on their behalf.

This is, in truth, an old instinct wearing new clothes. Price comparison sites performed an identical service for insurance and travel two decades ago: they narrowed the field, and consumers still chose the underwriter or the airline they recognised. AI shopping assistants are repeating the manoeuvre at scale. 89% of respondents say recognising the seller's brand is important or very important when acting on an AI recommendation, and 92% say reviews are decisive. Two-thirds (68%) would consider a seller they had never encountered before, but only once they had checked the reviews first.

AI is enhancing consumer decision-making, not replacing it. Consumers are happy to delegate discovery and shortlisting to an AI assistant. But trust remains critical when it comes to the final purchase. If the brand is not recognised and there are no reviews to validate it, the AI recommendation does not close the sale.

This is not a passing habit merchants can wait out. Nothing in the data points to a buyer’s remorse or retreat back to old habits - consumers are engaging carefully (checking reviews, weighing price, verifying brands) rather than experimenting half-heartedly and dropping off. That is typically the behaviour of people building a new habit, not testing a novelty.

There are national accents worth noting. UK consumers are the most brand-loyal of the four (15% cite brand recognition as their primary factor, against 8% elsewhere) and the most protective of their loyalty points: 36% say losing them would deter them from AI shopping tools altogether, against 14% in Germany. France is the most AI-trusting market surveyed, with over half of French consumers (53%) reporting a mostly positive experience and the most evenly distributed set of decision factors of any country in the study.

The free lunch, priced by algorithm

The second question was one of economics: will consumers pay for an AI shopping assistant they trust to be neutral, or take a free one and accept that its advice is shaped by whoever is paying for placement? The answer echoes the last twenty years of search and social media rather than departing from it. 43% are content to use a free, advertising-influenced AI assistant; only 27% would rather pay for a fully impartial one.

That is not the same as consumers switching off their judgement. The data shows no sign of what psychologists call cognitive surrender, a wholesale, uncritical deference to whatever the AI suggests. What it shows instead is a more calculating trade-off: people understand that advertising shapes the results, and they are pricing that understanding into how much they trust the answer.

However, 40% still say advertising reduces their trust in AI recommendations, rising to 48% in the UK. The US is the outlier on willingness to pay: 34% would fund a fully impartial assistant, the highest figure of any market. There are also some substantial differences by age, with half of over-55s saying advertising reduces trust, against a third of under-35s, who have grown up with algorithmically curated feeds and expect little else.

Two jobs, one journey

If price and trust explain how people choose, the third piece of the puzzle is where they choose to do it. Here the data complicates a good deal of the ‘AI will disintermediate everything’ commentary. 74% of consumers prefer an independent AI assistant for shopping, either a universal tool such as ChatGPT or Gemini (41%) or a category specialist for travel, finance, or healthcare (33%). Only 10% want AI built into a single retail, travel, or delivery platform.

But wanting an independent guide to the shop window is not the same as wanting an independent till. 93% of consumers expect their use of retail marketplaces to hold steady or increase as AI adoption grows; 89% say the same of travel marketplaces and food delivery platforms. Only around one in ten expect to rely on either less.

Set those two findings side by side and the shopping journey resolves into two distinct jobs that used to be blurred together: discovery, increasingly conducted by an independent AI acting as a well-briefed advisor, and execution, still handled by the marketplaces, merchants, and payment rails consumers already trust.

There has been a narrative that agentic AI would make marketplaces less relevant by allowing consumers to curate the internet on their own terms, rather than relying on marketplaces. What the research suggests instead is that consumers increasingly see discovery and execution as distinct stages of the shopping journey, and they still rely on established brands such as marketplaces when it comes to fulfilment, payments, logistics, customer service, and operational trust.

The competitive consequence follows directly: the competitive battleground is moving upstream into the AI discovery layer. Success will depend on whether products and services are surfaced by AI assistants before consumers ever enter a marketplace environment.

The good news for merchants is that this discovery layer is not a black box beyond their reach: with 43% of consumers already content to use free, advertising-funded AI assistants, merchants have a genuine opportunity to shape their visibility through the same well-tested advertising models now being built into these new AI platforms.

Is the Amazon-Perplexity fight actually about consumers?

Bring those three threads together, price-led discovery, trust-gated decisions, and a journey dividing cleanly into discovery and execution, and the anxiety behind the Amazon-Perplexity case starts to look overstated. Across all four waves, 90% of consumers expect to use retail and travel marketplaces the same or more as AI adoption grows. Just 14% say they simply follow an AI's top recommendation, against 32% who say price drives their choice. That is not a population handing its loyalty to an algorithm; it is a population using one to negotiate harder.

The fear that AI agents will simply cut merchants out of the transaction is, on the evidence of our research, a hallucination. Consumers are using AI to build a shortlist, not to transfer their loyalty. The Amazon-Perplexity case is really a fight about who owns the discovery layer, it is not evidence that shoppers have stopped caring who they buy from. That is not licence for merchants to relax, however: the competitive battle has moved upstream, into the AI discovery layer itself. But the idea that agents are about to replace the merchant relationship altogether isn't what consumers are telling us.

The question that keeps surfacing in conversations with merchants is some version of: will the AI simply take our customers? The research suggests that is the wrong question. The right one is whether the AI knows you exist when it draws up the shortlist in the first place.

What this means for the payments industry?

For merchants, the danger is not losing the sale at checkout; it is never making the shortlist. With 74% of consumers preferring independent AI over platform-embedded tools, visibility now depends on how faithfully your product, pricing, and trust data are represented to third-party AI systems, not merely on how well you present to a human on your own site. Brand recognition and reviews remain your strongest lever once you are on the list, and loyalty programmes still carry weight, particularly in the UK.

For AI platforms, a tiered market is forming: free, advertising-supported tools for the mainstream, and impartial, paid alternatives for a smaller, quality-conscious segment. That model has a ceiling: enterprise and higher-value users in particular are wary of a free, advertising-funded assistant that relies on their data to shape recommendations, and will increasingly expect platforms to be transparent about how their information is used and who else it is shared with. Litigation of the Amazon-Perplexity kind will keep fair representation, and data ownership, firmly in the public eye.

For PSPs, acquirers, issuers and schemes, the more immediate question isn't delegated authority - it's who is accountable today, while a human is still nominally in the loop, but bots and bad actors are already exploiting that grey area. The Amazon-Perplexity dispute is a live example: agents disguised as human shoppers, fraudulent merchants, and listings gaming their way past verification, and consumers potentially hiding behind agentic platforms to dispute purchases they authorised.

Authentication and dispute resolution frameworks need to catch up with that reality first. Only once genuine delegated purchasing emerges at scale will roles and liability need to be redrawn for a buyer that may be an algorithm acting under a person's authority. Volume is not migrating off existing rails, but the rules governing accountability on those rails are being tested now, with the harder questions about delegated authority still to come.

For the shape of commerce, the throughline across four markets is that the shopping journey is dividing into two distinct layers rather than collapsing into one. AI is becoming the window; large seller brand such as marketplaces remain the shop. Businesses treating this as a single threat to be repelled are likely defending the wrong ground, several moves too late.

Want to learn more? Watch the full webinar, including the Q&A session at the end.

About the author

Chris manages PSE Consulting’s business and is well known in the UK and EU for his regular insights into payments innovation.  He has spent the last 20+ years leading assignments for major clients during his time at PSE and Accenture. His specialisations include customer proposition development, market entry strategies and enterprise value creation.  Chris is a highly effective communicator, with very strong analytical skills and able to deliver recommendations to C-level clients and company Boards. Chris regularly supports major enterprises, corporates and global digital merchant’s payments bringing new market perspectives and identifying fresh opportunities for innovation and expansion. He has also delivered payment assignments for Fintechs, banks and processors on topics such as: regulatory impacts, new acceptance methods, open banking, BNPL, gateway/acquirer convergence, and opportunities for M&A and inorganic growth.

 

About PSE Consulting 

PSE Consulting is a leading global provider of payment advisory services to players across the payments landscape. PSE’s expertise has enabled it to deliver actionable market insights and operational optimisation to senior payments leaders for over 30 years. 

To learn more, visit: https://pseconsulting.com/