Winning the AI shortlist: visibility, trust, and marketplace loyalty in agentic commerce

We sat down with Chris Jones, Managing Director at PSE Consulting, to discuss PSE Consulting's recent consumer research into agentic commerce and what the findings mean for merchant visibility, PSPs, agents, consumer loyalty, and where the discovery layer is heading next.

Nearly half of consumers are happy to use a free, ad-funded AI shopping assistant over a paid, impartial one. What does that tell you about how the commercial model for agentic commerce is going to shake out?

It tells you the model isn't going to be one thing – it's going to split. Most consumers will end up using free services funded by monetising their data to advertisers, in much the same way search engines operate today: if it's free, you're the product. Our research backs this up directly – 43% of consumers are content with an advertising-influenced AI assistant, against just 27% who'd rather pay for something impartial.

Alongside that, I'd expect a complementary tier to emerge on the enterprise side, where businesses pay specifically for unbiased answers, free of any advertising influence. What we're heading towards is a genuinely complicated billing landscape for agents – advertising-linked models sitting next to outcome-linked models, sitting next to enterprise contracts with their own tiers, opt-ins and discounts. Arguably only Google has anything resembling this complexity today. Agentic commerce is about to build its own version from scratch.

As advertising within AI assistants starts to matter as much as search advertising once did, does that play out differently depending on whether the agent sits on-platform – built into a retailer's own app or site – versus off-platform, a third-party assistant shopping across multiple merchants? What does that split mean for how payments need to be built into the experience?

This is where it gets interesting for the big platforms. Companies like Amazon have built enormous revenue streams from customer eyeballs on their own surfaces. If agents start transacting on a consumer's behalf without the consumer ever landing on that surface, that revenue is potentially at risk. I'd expect merchants to start shifting ad spend away from traditional search and onto agentic platforms – though exactly how much, and how fast, will come down to conversion rates we simply don't have data on yet.

On payments infrastructure specifically: the evidence we've seen so far suggests agents don't actually want a payments role. They don't want to own fraud, chargebacks, logistics or delivery – that complexity is going to stay with merchants, which means the Merchant of Record remains the buyer of payments infrastructure, much as it is today. What will emerge, though, are new services around securing customer data, verifying identity, and passing reliable information to merchants for their own risk management. I'd expect fraud and risk tools to be bought both by agents – to protect their own reputation and by merchants, as they already are.

Marketplaces remain the destination even as AI becomes the starting point. Does that push more individual merchants toward becoming marketplaces themselves – aggregating other sellers or categories – and if so, what does that mean practically?

It's one of the more interesting second-order effects we're likely to see. Consumers have historically transacted through a fairly small number of sites, and in an agentic world that number may shrink further. That's a strong incentive for large brands with established customer relationships to broaden what they sell and lean harder on the trust they've already built.

You can see the pattern already in Amazon's move into adjacent categories, and it's not confined to the giants – Target's invite-only Target Plus marketplace in the US is growing quickly, and Tesco has moved the same way in the UK with Tesco Marketplace. As the agentic layer curates more of the consumer's choices, I'd expect more merchants to follow that logic.

Loyalty points still carry real weight in whether consumers trust an AI shopping tool. What do these figures tell us about the consumer priorities that standards like UCP need to design around as they mature?

What it tells us is that consumer influence extends well beyond price and the transaction itself – loyalty, rewards and reviews all matter, and any standard that ignores them is designing for half the picture. In our research, 89% of consumers said recognising the seller's brand mattered when acting on an AI recommendation, and loyalty points specifically are a real deterrent risk – over a third of UK consumers said losing their points would put them off using AI shopping tools altogether.

We're already seeing companies differentiate around this. Booking.com makes its inventory and pricing available to third-party agents but keeps its reviews and platform experience to itself – I'd expect a lot more of that: a deliberately different quality of service via third-party agents versus going direct with the merchant.

For UCP specifically, that means building in flexibility around things like FX to accommodate consumer preference during the transaction, and thinking hard about the post-transaction environment too – if a consumer starts a purchase inside an agentic environment, they're likely to want to manage queries and disputes about it there as well.

Bringing it all together - if a merchant only takes one thing from this research, what should it be?

That agentic commerce is going to be a genuinely significant channel – at least as important as Google search is to merchants today, and quite possibly on a par with their own website and app. This isn't a niche experiment.

The practical takeaway is that merchants need a low-risk environment to test and learn in, and the ability to adapt as standards and liability structures evolve and very few will be able to do that alone. They'll likely need to lean on their PSPs, or on new orchestration platforms that are starting to emerge to fill that gap. Right now, the cost of experimentation is relatively high, though I'd expect that to fall as PSPs build out their own intermediation layers. But merchants do need to start now – they need real data on how their own customers want to use agentic infrastructure, and the only way to get that is to be in the market testing it.

About the author

Chris Jones 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/