The customer has already decided, more or less, before they open your app.

This is the part of the acquisition journey that challenger banks are not tracking, and it is happening at scale. Consumers are using AI search tools – ChatGPT, Perplexity, Google’s AI Overview – most heavily at the beginning of a purchase decision: the research phase, the comparison phase, the narrowing-down phase. Research from Cognizant is unambiguous on this. The Learn phase is where AI has the deepest penetration in financial services. By the time someone reaches the Buy phase, the shortlist is already formed.

Clients are evaluating products, comparing institutions, verifying credibility and narrowing their options before ever visiting a website. They are not arriving at your homepage with an open mind. They are arriving to confirm a view the AI already helped them build.

Why AI search is never neutral

This is not a neutral process. The way a question gets answered shapes the answer that feels right. An AI system that describes one bank as “the go-to for mass-affluent savers” and another as “popular with younger customers” has already done the segmentation work – and the person reading it will likely accept those categories without examining where they came from. The framing is invisible, which makes it more powerful, not less. Comparison sites understood this twenty years ago: the review that calls a product “reliable and straightforward” is making an argument, even when it reads like a summary. AI search has industrialised that effect. A prospective customer who has asked three questions about savings rates and received three answers citing the same institution has not been sold to. They have simply arrived at a conclusion that felt like their own.

Janine Hirt, CEO of Innovate Finance, argues that 2026 will be the year UK challenger banks cement their status as a new backbone of the financial system – institutions that are “no longer just secondary accounts but the primary banking relationships for millions.” She is right about the destination. The question is whether the sector appreciates what is required to get there, and information arbitrage is a large part of the answer.

Where challenger banks start at a disadvantage

The incumbents, counterintuitively, hold a structural advantage in this new terrain. They have decades of coverage in outlets that AI systems treat as authoritative. They have Wikipedia entries with citations, analyst reports, regulatory filings, and the kind of institutional paper trail that signals credibility to a large language model. Trusted publishers like NerdWallet and Bankrate appear repeatedly as source material across the major AI engines, shaping not only consumer perception but also the information those systems use when assembling guidance. The editorial decisions of a handful of financial comparison sites are currently more influential on AI-generated recommendations than most banks’ own communications output.

Challenger banks built their early reputation through channels AI search cannot easily ingest – out-of-home advertising plastered across the London Underground and the New York subway, a thousand flashes of a salmon-orange card, viral app store moments. These are expensive vectors. A wraparound at Bank station costs serious money. So does a takeover of the L train. They build recognition, and recognition matters. They do not, however, determine what the AI says when someone asks which savings account is worth opening.

Philip Belamant, CEO of Zilch, put it plainly earlier this month: “AI will fundamentally change how people interact with money, shifting payments from something consumers actively manage to something that is intelligently managed and optimised in the background.” He was talking about payments execution. The same logic applies a step earlier, to payments discovery. If AI is increasingly managing and optimising financial decisions in the background, then the information those systems draw on is not a communications afterthought. It is the competitive ground.

What gets a bank into the answer

What occupies that ground is substantive, specific, authoritative content: genuinely argued pieces in respected trade publications, data that other outlets cite, commentary on regulatory developments that positions a bank as a participant in the industry conversation rather than a product looking for customers. A challenger bank that does something genuinely notable in a narrow space – and gets it covered properly across credible sources – will find that moment working for it long after the press release is forgotten. The AI system has a long memory for authoritative citations and a short one for promotional noise.

Most challenger banks do not think about content this way. They think about it as brand building, which means it gets treated as discretionary. That framing is a legacy of the first disruptive moment in this sector, when the product was the story and the communications were commentary on it. The second disruptive moment is different. It is about winning the consideration phase before that experience begins, in a part of the customer journey the bank never sees.

The second disruptive moment will not be decisively won on the side of a bus.

Jon Schubin runs content for Cognito