Guest speaker: Mike, CEO and Co-Founder at Financial Marketing, a specialist City of London agency built around supporting financial organisations, financial services businesses and corporate finance firms since 2012. Mike has spent more than 35 years building and differentiating financial services brands, was formerly Head of Marketing and Vice-President at Bank of New York, and is a Freeman of the City of London.
Host: Yaser Ayub, Founder of Rayze. Yaser has 18 years of marketing experience across fintech and financial services, including as VP of Marketing at Rocket Internet in Berlin and Head of Marketing for a challenger bank. His areas of expertise are SEO, AEO and PPC, and he started out as a web designer and developer.
There are two separate but closely linked problems behind "being found and interesting."
The first is visibility. A firm might have brilliant research and real insight, but if that content sits behind a gated PDF, or isn't accessible to a large language model, it can't be found — no matter how good it is. Boardrooms are increasingly asking whether their business appears in AI search. The honest answer often comes down to something most marketing teams haven't checked: whether an LLM can actually reach the content in the first place.
That visibility gap has drawn in a wave of self-styled experts offering shortcuts. Some of the tactics used to force quick visibility in AI search are little more than the black-hat SEO of a decade ago — publishing "listicle" articles that rank your own brand first and list competitors below it is one example. Google has already started penalising it.
A study by Lily Ray found that Google cited a brand's listicle but didn’t recommend that brand. Competitors got recommended — a reminder that shortcuts like this tend to backfire. The parallel with old tactics like keyword stuffing and invisible text was deliberate: those mistakes were eventually recognised and penalised, and the same is likely to happen with today's AI-visibility shortcuts.
The second problem only shows up once the first is solved.
Financial services marketing has long leaned on the same handful of words — trusted, experienced, client-focused, flexible — and that language has become wallpaper. It's easy for a human to skim past, and just as easy for an LLM to treat as generic and forgettable. The point isn't only to be visible. It's to give people, and the systems that represent brands to people, a real reason to remember you.
Several clients running AI-optimisation work recently reported a marked improvement in lead quality. That result led to a SimilarWeb study on the downstream impact of being mentioned in AI. It found that a brand mention in AI made a visit up to three times more likely, with roughly double the engagement time once someone arrived. That matters more, not less, as AI-driven traffic gets harder to track through conventional analytics.
It also reflects a shift in how AI search works. People increasingly ask open-ended questions rather than typing keywords, so ranking on page one still helps — but becoming the trusted source cited within the answer matters just as much, if not more. The sources LLMs draw on and prefer are often the hardest to get featured in. Wikipedia (kept genuinely current, not stuffed with promotional text) and YouTube were both raised as examples.
One of the clearest differences between traditional SEO and AI search is that LLMs work at the level of topics and entities, not individual keywords.
This matters strategically. As more buyers get a synthesised answer rather than a list of links, the goal shifts from ranking to becoming one of the sources that inform the answer. Any go-to-market strategy now needs to account for how a brand is represented in LLM-driven search, not only traditional search, including sentiment as well as presence. The underlying goal isn't to write for machines. It's to make real expertise accessible enough that, when a buyer asks AI a genuine question, a brand's own content does the answering.
The challenge facing financial marketers breaks down into three integrated dimensions, not three separate lanes:
On distinctiveness specifically, a line like "we're a trusted partner for businesses seeking growth finance" no longer does enough work. The more specific version — naming the deal sizes, sectors or problems a firm actually specialises in — is what both LLMs and humans respond to.
Being vague about your audience is one of the most common problems, and one of the easiest to fix. Saying you're an invoice finance provider, an asset finance provider or a mortgage provider isn't enough on its own. Getting specific about the actual problem you solve — helping a growing business manage a cash-flow challenge, for a particular deal size or market — gets much closer to how a real buyer thinks and searches.
That thinking should shape the content itself: write as if answering a genuine question from a customer, because that's effectively what's happening. It also needs to be visible, not buried behind gated PDFs, non-indexable formats, video without transcripts, or poorly structured pages. There's nothing wrong with gating content for a good reason, like capturing a lead or gauging interest from a target audience. But if the goal is search or AI visibility, content needs to be open. IP that stays locked away indefinitely isn't doing anything for the business anyway.
One example found a useful middle ground: a company kept a report gated as a PDF, but also published a separate, fully crawlable article covering its key points. That "create once, publish everywhere" approach — adapting the same material for the medium and audience it's reaching, whether that's an LLM or a human reader — let the brand satisfy both channels at once.
None of this matters if the technical basics aren't in place, and here traditional SEO and AI search overlap almost completely. The checklist looks familiar: can crawlers actually render the site (heavy JavaScript-driven sites are a common problem), is the HTML clean, is the robots.txt file configured correctly, and is the CDN set up so it isn't accidentally blocking AI crawlers.
Page speed matters more than many teams assume. One study found that slow-loading sites are disadvantaged because LLMs are trying to retrieve information quickly. Internal linking still signals which pages matter most, and schema markup does the job it's always done: telling both search engines and LLMs what a page is actually about.
In one real example, a client's blog pages weren't being indexed by Claude at all. Solving it meant working systematically through a list of possible technical causes — including asking Claude directly for help identifying the blocker.
Understanding which sources an LLM pulls from is worth checking directly: run your own searches, see which sources appear, and assess how accurate that information is. PR is a particularly effective way to influence this, because it's current. One client had a new board appointment picked up and represented accurately in AI results almost immediately after a press release went out.
Wikipedia is another strong lever, though it comes with strict inclusion criteria — typically genuine brand recognition — and isn't something that can be gamed easily. Attempts to spam it tend to get caught and penalised.
For any financial services brand starting from scratch, the process typically begins the same way: run a set of real questions through the major AI tools, see what they currently say about the brand, and map the gap between that and what's actually true.
Working with a fintech client serving banks, that process — auditing what AI already knew about the brand and its core topics, then closing the content gaps around those topics — increased ChatGPT-driven sessions by 217% and Gemini-driven sessions by 130%. In the same body of work, AI Overviews visibility for a set of top-three listings improved by 255% over a defined period. The method stayed consistent throughout: choose the right core topics, not keywords, test how LLMs currently answer around them, check what competitors are saying, and build out content to close the gap.
View the case study here.
Once ICP-led, question-based content is live, the work doesn't stop. It's easy to create content nobody engages with, and pages that sit with very low engagement over time can hold back a site's ability to rank for its more competitive terms. Regular zero-click analysis, followed by consolidating or improving underperforming pages, is a standard part of ongoing work.
The same discipline applies to chasing every possible keyword. Clients who focus on a specific ICP and message tend to see both better lead quality and stronger AI visibility, because a distilled brand essence works harder than a broad one. Less genuinely is more, for search rankings and brand recall alike.
A short story makes the point well. A craftsman carving elephant figurines from scrap wood, asked how he does it so well, explains that he simply cuts away anything that doesn't look like an elephant. Strong brands work the same way — BMW is built around joy, Volvo around safety, Disney around magic, Harley-Davidson around a sense of freedom. Each has cut away anything that doesn't support that one association.
Most financial services brands do the opposite. They reach for the same safe, generic language as every competitor — trusted, experienced, client-focused — and end up sounding identical to everyone else in the category. Byron Sharp's point that brands rarely compete on a genuinely unique, meaningful difference, even though they need to, came up here, alongside the idea — often credited to Simon Sinek — that people buy based on trust and values as much as on functional features. Being the best at the functional basics is now the price of entry, not a real differentiator.
This connects directly to trust in AI search too. If an LLM represents a brand as trustworthy, that's effectively a third-party endorsement passed on to the buyer. For financial services specifically, a "Your Money, Your Life" category for Google, visible trust signals like author bios, publish and review dates, accreditations, and content-refresh frequency support compliance and AI or search performance at the same time.
The clearer and more consistent a brand's message, the more likely it becomes an automatic association for a person — and, by the same logic, an easier pattern for an LLM to recognise and represent consistently.
In practice, this means replacing generic claims with specific, provable ones. Instead of describing a business as "a trusted finance provider," describe an actual deal: what was funded, for whom, and what it enabled. Quantifiable results, named case studies, testimonials, white papers and industry recognition all count as meaningfully specific content, and all add real value for search engines, LLMs and humans alike.
Prompt tracking in AI search is still a young, imprecise discipline. Getting it right depends on a few principles: use open-ended, non-branded prompts, since branded prompts inflate share-of-voice numbers and paint an unrealistically strong picture. Build long-tail prompts around how the ICP actually asks questions, drawing on forums, "people also ask" data and AI-specific call-tracking tools. Track share of voice against competitors over a sustained period, rather than expecting a single snapshot to mean much.
Being upfront with clients that this is new territory — agreeing the ICP, doing the keyword research, and agreeing a working set of prompts before optimisation begins — is the most honest way to run this kind of work.
For a financial services brand ready to act on this, the priorities stay consistent. Audit what AI currently knows about the brand and its core entities, and run that alongside a wider brand and messaging audit to spot where competitors are better understood on a given topic. From there: fix technical and crawlability gaps (this should be routine, not exceptional), correct inaccurate sources, build genuinely question-led content and case studies, identify a working set of non-branded ICP prompts, and track share of voice against competitors over time.
We can run the same audit on your brand: what AI already knows about you, where the gaps are, and what it will take to close them.
Book a 1-2-1 with our team. Talk to a senior consultant.