Inside Nuvei's Customer Advisory Board: 7 things the smartest merchants think about payments in 2026

Beaverbrook Estate sits in the Surrey Hills, about an hour outside London, and it carries more history than its gardens let on. As the home of Lord Beaverbrook, Churchill's Minister of Aircraft Production, it was once a hub of wartime decision-making. Under his direction in 1940, British fighter production, including the famous Spitfire, tripled in the months leading into the Battle of Britain. I thought about that story more than once over the two days I spent there in late June: when the right people gather with a shared purpose, progress compounds.

That was the idea behind Nuvei's Customer Advisory Board, which brought some of the industry's most advanced payments leaders to Beaverbrook for two days of honest conversation about where the sector is heading. I have sat in plenty of rooms where merchants and providers talk past each other. This was the opposite of that. People compared notes on what is working, admitted what is failing, and pushed us, and each other, on what should come next.

What follows are the seven things that room made clear to me about where payments infrastructure is headed.

1. Merchants are getting ready for agentic commerce before the volume arrives

Let me start with an honest read of the market: consumer demand for agentic commerce is not here yet. In fact, Nuvei's How America Pays survey found that 60% prefer to wait and see how others use AI for purchasing before trying it themselves. And when OpenAI pulled its Instant Checkout feature in March 2026, just five months after launch, citing near-zero sales and operational complexity, it gave skeptical merchants all the permission they needed to sit back.

But the payments leaders we spoke with accept that while the volume for agentic commerce is not here today, the gap before it arrives is precisely the window in which business readiness gets built. And they are right to do so, with Juniper Research forecasting agentic transaction value growing from around 8 billion dollars in 2026 to 1.5 trillion dollars by 2030.  

At the same time, payments leaders want a say in how agent identity and liability get defined. That is a meaningful shift for an industry that has historically treated protocol design as something that happens to merchants rather than with them, and a reminder that trust works best when networks, merchants, and payment providers build it together rather than any single party writing the rules alone.  

Which leads me to my next point.  

2. The trust gap that stalled agentic's first wave is starting to close

The first attempt at agentic commerce, built around AI-assisted product discovery and purchase, ran into a wall that had little to do with the technology. The commercial and liability logic had not been resolved, and consumers were unwilling to delegate the checkout moment to a system with no clear accountability.

Merchants carry a parallel fear, and they were candid about it. When the entire journey happens inside an AI interface, the merchant may never interact with the customer directly. Who owns the relationship, the data, and the loyalty? How do you market to a customer you never meet, or even know who your customers are? For brands that have spent decades building direct relationships, that is an important question.

What has changed in recent months is the emergence of trusted agent protocols and Know Your Agent (KYA) verification: mechanisms that let a merchant extend the same confidence to an autonomous buyer that it extends to a verified human one.

3. The basics are back at the center of the conversation

Well ahead of anything to do with AI or new rails, what remains as the biggest sources of operational friction for payments leaders is reporting accuracy, settlement timing, and the daily mechanics of reconciling across providers.  

A finance team processing at scale across multiple providers typically faces a different settlement rhythm from each one: one acquirer settles in two business days, another follows a rolling cycle, a regional partner operates on entirely different cut-off rules. At any given moment, a share of revenue is sitting in transit across systems that report it differently, in different formats, against different transaction IDs.

Merchants processing across multiple providers routinely lose between 8 and 12 basis points to hidden fees and reconciliation gaps they cannot explain. For a business processing $10 million a month, that is up to $144,000 a year leaking before anyone asks a question about the next frontier.  

What this tells us is that while much of the provider market chases the next innovation cycle, what heads of payments and finance teams are asking for first is a partner with the fundamentals in place: reliable transaction processing, accurate and timely reporting, consistent settlement, and enough operational transparency to actually act on the data. In 2026, at least for the people in that room, the basics are back at the center of the conversation.

4. AI already runs inside the payments function, and data readiness decides whether it works

The visible AI story in payments is agentic commerce. The more widespread one is how payments and finance teams across the industry already use AI every day to reconcile transactions, interpret performance data, and build the dashboards their own leadership depends on. But when the AI workflows draw data from many systems and providers at once, you get a model that will confidently produce the wrong answers. A Dun & Bradstreet study found that 54% of payments companies and 51% of fintechs report AI projects that have already failed as a direct result of poor data, with duplicate records, siloed systems, and distrust in underlying datasets cited as the dominant causes.

The quality, consistency, and transparency of the data a payments partner delivers is becoming a selection criterion, because that data now feeds the models a merchant's leadership relies on. The providers who treat their reporting as a product will be the ones whose merchants can put AI to work.

5. Payments still has an internal marketing problem

This one is close to my heart as a Chief Revenue Officer at Nuvei. Payments rarely gets described inside a business as what it actually is: one of the most financially impactful functions a company runs. Persuading finance, product, and executive leadership that payments infrastructure is a growth lever rather than a cost line remains one of the most persistent challenges for a head of payments. It is a strange dynamic. The function processing the largest single flow of a company's revenue often has to fight hardest for a seat at the strategic table.

As payments data becomes one of the richest sources of commercial insight a business holds, showing where customers convert, where they abandon, and which markets are underperforming, it is becoming a vital lever for finance teams. With average cart abandonment rates of 70% according to Baymard Institute, it also exposes friction at the point of purchase. And approval-rate improvements land directly in revenue, which makes payments performance increasingly hard for a CFO to ignore.

Payments providers have a role to play in accelerating that shift, and it is one we at Nuvei take seriously. The most useful thing a partner can do for a head of payments is arm them with evidence: reporting that expresses infrastructure performance in revenue terms, from approval uplift and recovered revenue to cost per transaction by market.  

6. Fraud tooling is under more honest scrutiny than it has been in years

Merchants are asking pointed, specific questions about the maturity of fraud tooling across the market, rather than accepting vendor claims at face value. Nasdaq Verafin's 2026 Global Financial Crime Report puts global losses to bank fraud and scams at $579.4 billion in 2025, with scam losses growing at more than 19% a year, fueled in large part by fraudsters deploying the same generative AI tools that payments companies use to defend themselves.

The most advanced providers are building AI-native fraud detection that can keep up with modern attacks: real-time anomaly scoring, behavioral biometrics, and layered identity checks that go far beyond static rules. Because fraud tactics change in real time, detection has to change with them.

Fraud is also highly industry-specific. The abuse patterns of a gambling operator are very different from those of a travel marketplace, a digital goods platform, or a subscription business. Merchants are no longer willing to accept generic tools that ignore those differences. Risk teams now want to know how the model is trained, what signals it uses, and whether it reflects the chargeback patterns, abuse loops, and regulatory exposure of their own vertical.

7. Design partnerships are replacing vendor relationships

The shift I find most energizing is that the industry's more advanced payments teams are actively seeking design partnerships, working alongside providers to shape tooling as it is built rather than adopting whatever ships next. In fraud especially, merchants want their vertical's reality reflected in the models from day one.

That is a different commercial relationship than the one most payments vendors are used to offering, and it rewards providers willing to treat their largest merchants as collaborators rather than customers. An advisory board like the one we hosted at Beaverbrook is precisely this kind of design partnership, where merchants' lived operational reality shapes what gets built.

What two days at Beaverbrook made clear

One throughline ran across everything I heard: readiness. Readiness to participate in standard-setting rather than simply comply with it. Readiness to extend trust to autonomous agents once identity is verifiable and liability is assigned. Readiness to pursue agentic infrastructure and new rails once the reporting underneath that ambition is actually reliable.

What stood out most to me, especially from returning members, was how much of that readiness is relational. The strongest ideas came from people who have been in dialogue with us and with each other for years, challenging assumptions, comparing failures and successes, and gradually building the trust to share what is really happening in their businesses.

In that sense, the real lesson of Beaverbrook is that progress belongs to the people who keep showing up to do the work together. I left the Surrey Hills convinced that the future of payments will be shaped in rooms like that one.