For the fintech or RegTech CEO who has funded two years of sponsorships, stands and content without a deal anyone can trace to it. Marketing in regulated financial services has been rebuilt around signals, ecosystems and a new reader of your reputation, and the firms winning attention are not the ones talking loudest about themselves.
Somewhere in your budget is a line for marketing, and if you are like most founders I meet, it hurts to look at. Two years of conference stands, sponsored roundtables and a content calendar that never misses a week, and not one deal in the pipeline that anyone can honestly attribute to it. The instinct at that point is either to cut the spend or to double it, and both instincts miss the diagnosis. The problem is rarely the volume. It is that the money is buying broadcast in a market that has stopped listening to broadcasts.
The marketing innovation chapter of my book is built around that shift: from static campaigns to intelligent engagement, from pushing products to reading signals. And it contains the sentence this whole article hangs on: brand equity is not what a firm says about itself, it is what others say about using it. Last week I wrote that your first meeting now happens without you, inside the answer ChatGPT or Claude gives when your buyer asks who does what you do. This piece is about what fills that window, and the answer is other people's words about you, which means your marketing job has quietly changed from speaking to being spoken about.
The voices that sell for you
Brand equity is what others say about using you. Four voices to invest in, and one to retire. The rest of this piece is how to earn each one.
Testimonials, named case studies, peer reviews. Evidence with numbers attached, curated into onboarding and campaigns.
Why it works: peer validation now outweighs brand messaging in the buying decision.
Marketplace listings, integration guides, co-branded documentation, sandbox access. Proof you fit the buyer's existing systems.
The model: Stripe's documentation sold more than its sales team ever did.
Trade press, industry bodies, working groups, podcasts and speaking slots your buyers already trust.
Why it works: a buyer consults seven sources before deciding. Be one of them.
AI answers your buyer consults first. Fed by quotable proof pages, consistent company facts, and coverage in sources the models trust.
The proof: Ramp went from 3% to 22% AI-answer visibility in three months.
Broadcast messaging to anonymous segments, measured in impressions.
The evidence: NatWest found 80% of its campaigns were rated not relevant by the customers receiving them.
The discipline starts with signals. A CIO at a Nordic pension fund downloads a white paper on SFDR reporting; that is not a lead to be added to a newsletter, it is a buying signal with a name attached, and account-based marketing exists to answer it specifically. In regulated financial services your addressable market is not millions of anonymous clicks, it is a few hundred named institutions, which makes ABM less a tactic than the only honest description of the job. The buyer, meanwhile, is doing their own listening: Gartner found this year that 45 per cent of B2B buyers used AI in a recent purchase, and that a buyer consults an average of seven information sources before deciding. If your firm is not among the seven that a particular buyer reads, then for that deal your marketing might as well not exist.
And when you do speak, speak to the person, not the segment. The evidence for what generic messaging costs comes from an unexpectedly large witness: NatWest found that 80 per cent of its campaigns were rated not relevant by the customers receiving them. Its answer was to rebuild engagement around what it calls customer missions: offers triggered by moments in the customer's own life rather than by the bank's campaign calendar. The operational gains are startling. A new engagement now goes live in one day where it previously took 60 to 100 days, and needs four or five people where it previously took forty. If a high-street bank can personalise at that speed, a fifty-person fintech sending the same email to its entire prospect list has run out of excuses.
Forrester's 2025 Buyers' Journey Survey found that 94 per cent of business buyers now use AI in their buying process. Yet visits arriving from a link in an AI answer make up only around one per cent of most firms' website traffic. Almost every buyer is asking. Almost none of it shows in your analytics.
A fund COO asks ChatGPT who the credible providers are, reads the answer and settles on three names without clicking anything. If you are one of them, she searches for you a week later and your dashboard credits Google. If you are not, you lose a deal you never knew existed. The shortlist forms silently. So SEO is not dead; it has a new employer. The models cite pages that are crawlable, structured and authoritative, which is SEO's old checklist serving a new reader, and writing for that reader is what generative engine optimisation means.
The playbook is practical, and there are now firms proving it works. Ramp, the US spend-management fintech, grew its visibility in AI answers from 3 per cent to 22 per cent in three months, moving from nineteenth to eighth among fintech brands in its category, by building precisely the assets the models want to quote: comparison pages, proof pages, plainly structured answers to the questions buyers actually ask. Three hundred citations in a single month, faster than traditional SEO ever compounded. The test for your own firm costs nothing: ask the models who the leading providers are for your buyer's problem, and see whether you exist. That is the new share of voice, and it is measurable this afternoon.
The ecosystem is where "what others say" becomes a distribution strategy. A modern financial firm markets through marketplaces, ERP app stores, developer forums and partner catalogues as much as through its own site, and every integration listing, co-branded guide and sandbox is a marketing asset: third-party proof that your platform works with the systems your buyer already runs. Stripe remains the canonical case, a payments firm whose documentation did more selling than its sales team, because developers read it and then marketed Stripe to their own employers. Each partner page describing your integration is one more independent voice vouching for you, and, usefully, the machines read partner catalogues too.
My book's advice on visibility still holds, and in 2026 it pays twice. Coverage in the trade publications your buyers read, a seat on an industry working group, a slot on a conference agenda, an appearance on the podcast your prospects listen to on the way in: each is another voice saying you know what you are talking about. None of it is advertising, which is precisely why it works. A fund COO discounts what a vendor says about itself; they do not discount what their own industry body chose to put on its stage. And those same publications are the sources a buyer's AI draws on when it assembles a shortlist, so every credible mention is read twice, once by a person and once by a machine.
If you sell to consumers rather than institutions, the equivalent voice is the creator, and there the regulator has drawn a hard line. The FCA's 2025 crackdown on illegal finfluencer promotions produced more than 650 takedown requests and three arrests, and the first finfluencers it charged face trial in 2027. Borrowed credibility has to be borrowed properly: disclosure, vetting and risk warnings built into the content before it is published, not after the regulator complains.
None of this earns its budget line without attribution, and my book is blunt about vanity metrics. Impressions flatter; pipeline decides. The measurement that matters connects marketing to the stages from last week's piece: which content moved a named account into qualification, which proof page shortened the paper process, which partner listing sourced the meeting. Pricing, incidentally, has become a marketing lever in its own right, freemium tiers and usage models that let the product market itself, and it deserves its own piece later in this series. But the scoreboard here is simple: marketing that cannot point to named accounts moving through named stages is decoration, and 2026 has no budget for decoration.
So the fix is the same as it was for the revenue engine, a rebuild rather than a louder version of the old thing. Listen for signals and answer them personally. Build the pages the machines quote and the buyers trust. Let your clients, your partners and your industry speak for you. Then measure everything against pipeline, because the spend was never the problem; the silence about you was. It is not how good you say you are. It is how good others say you are, and in 2026 the others are talking to your buyer before you ever get the chance.
Sources: Gartner; Bain & Company; Forrester; Previsible; Profound; FCA; One Blackwater. All linked inline.
Whether your buyer’s AI names you is the first of five stages a Revenue Engine Review examines, from the machine shortlist through to renewal. One Blackwater runs it in three weeks. Detail on the Services page.