‘Number One on Google and Invisible in AI’: How Brokers Get Shortlisted by Chatbots

The economic bargain of the internet for the past two decades was that keyword-heavy search engine optimisation (SEO) was enough to secure visibility. Now, users pose long questions to AI chatbots and receive direct answers rather than links to follow.

Pepperstone Growth Consultant Andy Francos notes that AI assistants have become a meaningful part of how new customers discover and consider financial services brands.

Zay Yang, Founder & CEO of Wordflow, a company helping brands track their recommendations across Generative Engine Optimization (GEO) platforms, says that most visibility is not on a brand's own site.

Pepperstone now focuses on getting cited or recommended in AI questions across all major engines, including ChatGPT, Google's AI Overviews and AI Mode, Gemini, and Claude.

For regulated brands, when someone asks an AI to establish a broker's safety or licensing catalogue, the most-cited source is the relevant regulator, not the broker's homepage.

If a trader wants specific information only the business will know, such as deposit bonuses, the AI will crawl the broker's website, but it requires the information to be in plain text and cannot handle JavaScript well.

Yang stresses that the things that decide a recommendation—such as who regulates a firm, what it charges, and which accounts it offers—must sit in plain text, not buried in a script, a slider, or a PDF.

Homepages that were exclusively designed for human eyes will now need to be redesigned for the AI's crawl, as not adapting to this architecture means being invisible to millions of people.

Francos notes that third-party sources now account for a significant share of the citations associated with Pepperstone, which is backed by Wordflow's broker data.

Beyond regulators and homepages, recurring sources in AI search include independent comparison and review sites, established finance editorial and local personal-finance media, and community.

For the brokers Wordflow tracks, the brand's own website is routinely outranked as a cited source by third parties.

Yang adds that structured FAQ and entity markup can help, as they let the engine place a broker correctly and quote it cleanly.

Yang highlights that "You can be number one on Google and invisible in AI at the same time. The pages that get cited read like a credible source, not a landing page. Thin, promotional, keyword-built content underperforms, and AI spots it faster than Google ever did."

Local Sources and the Hedging Tax

Not everything is equal in AI search. Training data and referenced citations remain heavily weighted toward US and UK English, so searching the same brand in English and German will yield different answers, sources, and framing.

Across the brokers Wordflow tracks, the cited sources barely overlap: in Australia, they include ASIC, MoneySmart, and local comparison sites; in the UK, the FCA, Trustpilot, and BrokerChooser; and in Singapore, MAS, the local exchange, broadsheets, and homegrown personal-finance sites.

A broker can be recommended confidently in one market and hedged as newer in the one next door, purely because the local citation core is different.

Yang says part of Wordflow's job is to minimise the so-called hedging tax, where a broker can be recommended half the time and still lose deals because a third of those recommendations come caveated.

There is an upside in all these, as local sources are less contested, allowing a broker to build authority faster there than in a crowded US market.

Reddit's Declining Role

Reddit has a special place in retail trading lore, having been the birthplace of the GameStop saga. In 2024, the platform signed deals with both Google and OpenAI, giving them access to its massive library of human conversations.

In one market Wordflow tracks, a single Reddit comparison thread is among the top sources cited in Google's AI answers and Perplexity for brokerage questions.

However, the platform's weight within AI search appears to be diminishing. Pepperstone previously held Reddit among its primary source citations, but that changed abruptly when GPT-5.6 was rolled out to general users, causing the site to fall off the reference list completely.

Francos notes that even as total citations quadrupled, Reddit's prominence did not recover.

The platform is also now considering not extending its deal with Google.

Francos says that Reddit is the most over-generalised data point in AI search, and in their UK broker prompts, Reddit ranks 32nd, behind the FCA, gov.uk, and a half-dozen competitor sites.

One cannot simply buy their way to an organic Reddit threat the same way they can buy a citation elsewhere; those are earned over years by being useful.

Francos adds that a well-placed advertorial on a trusted publisher works differently and can earn citations.

He says the honest story is that Pepperstone tried paid Reddit and it didn’t pay back, so the budget is a small, shrinking line now.

What is more important on Reddit is community engagement, answering practical questions transparently rather than treating the platform purely as an advertising channel.

James Buckley-Thorp, founder and CEO of UK-based AI company Atlian, notes that brokers and prop firms seeding Reddit threads to get ChatGPT recommendations are financial promotion.

He equates it with fincluencers, which the FCA and other regulators have been trying to rein in.

Buckley-Thorp says a planted thread that an AI then repeats to millions is the same thing with better camouflage, for products where the FCA says 80% of customers lose money.

He adds that AI is the best tool for catching this, as it can spot a hundred accounts praising the same firm in the same week far faster than any human moderator.

The Next Shift

As the rules of discovery are rewritten, the industry faces an even deeper shift, with the immediate headache being attribution, as the customer journey has gone completely dark.

A trader might read a debate on a forum, ask ChatGPT to generate a shortlist of regulated brokers, independently google the chosen brand, and finally open an account.

To the broker's internal metrics, this registers simply as an organic brand search, with every channel that actually influenced the decision remaining invisible.

But this marketing puzzle pales in comparison to what comes next. Yang warns that "the ground is still moving."

The next shift is agents that don't just answer but act, comparing and even opening accounts on someone's behalf.

This is not a distant hypothetical, as the industry has already begun experimenting with Model Context Protocol (MCP) integrations, allowing traders to connect their accounts directly to AI agents.

While current safeguards prevent these bots from executing trades autonomously, the direction of travel is unmistakable.