Young Investors Are Giving AI the Benefit of the Doubt

The financial-services industry has spent years worrying about the influence of finfluencers. It now has a more diffuse and potentially more consequential source of influence to contend with: the answer generated when an inexperienced investor types a question into an AI tool.

The FCA’s research published on 27 August found that 56% of UK adults aged 18 to 40 who own, or would consider buying, investments trust AI tools for investment information. That puts AI ahead of television and radio, the press and social-media influencers in the regulator’s comparison. Four in five less experienced investors had already used AI for investment help.

The figures should not be read as evidence that young consumers have suddenly developed blind faith in algorithms. Most respondents understood that AI can be wrong, and a large majority said sources should be checked. The commercial issue is more subtle: AI has become a credible first-stop interface for financial curiosity.

That matters because the first explanation a person receives often sets the frame for every decision that follows. It can shape which product they search for, what level of risk feels normal, whether they view an investment as regulated and whether they believe they have recourse when things go wrong.

The first answer in an investment journey increasingly arrives before a regulated firm is even in the conversation.

Trust is being transferred, not earned

The FCA’s findings expose a significant gap between confidence in AI and understanding of its limits. Almost half of those surveyed believed AI-generated financial information was regulated. Nearly a third thought compensation arrangements might apply if an AI-led recommendation proved harmful. More than a third considered it acceptable to make an investment decision solely from AI output.

This is not simply a consumer-education problem. It is a question of how authority is perceived online.

A fluent interface can make a response feel impartial, current and tailored even when it is drawing on stale, incomplete or poorly ranked information. Unlike a human influencer, an AI system does not announce a visible commercial motive, personality or tribe. That apparent neutrality may be part of its appeal. It also makes its limitations less intuitive.

For firms, the result is an awkward reversal. The heavily regulated provider may have more reliable information, clearer disclosures and accountable staff, but still enter the customer journey after an unregulated general-purpose tool has established the basic narrative.

The task, then, is not to mimic chatbot confidence with more promotional content. It is to make regulated expertise easier to reach and easier to use when a customer has a genuine question.

---

Design for the question before the transaction

Many investment businesses still organise digital journeys around product pages, campaign landing pages and application funnels. Those assets are necessary, but they assume the customer already knows what they want.

AI changes the more important stage: the messy, early conversation. People ask whether a stock is overpriced, whether a cash ISA is enough, how much risk is sensible, what a fund fee means or whether a trend on social media is credible. Those questions do not map neatly to a product taxonomy.

A useful response from a regulated firm should therefore have three characteristics.

First, it should be specific without pretending to be personalised advice. Generic educational content can be accurate yet fail to help a consumer move forward. Firms need plain-language explanations tied to real decisions, accompanied by clear boundaries around what can and cannot be concluded without knowing the individual’s circumstances.

Second, it should show its workings. Dates, sources, assumptions and uncertainty should be visible rather than buried. This is particularly important in markets where prices, tax treatment, eligibility and product terms can change. A confident answer that cannot identify its basis is a poor substitute for a slower but verifiable one.

Third, it should provide a sensible route onward. A customer who starts with a simple research question may need regulated advice, guidance, a risk warning, an affordability check or merely time to reflect. The experience should not force every enquiry towards a trade.

This is where good consumer outcomes and commercial sense align. A firm that helps people understand the difference between information, guidance and advice is more likely to acquire customers with realistic expectations — and less likely to inherit avoidable confusion later.

The evidence trail becomes part of customer experience

AI-assisted services inside financial firms face a related but distinct challenge. It is not enough for a model to produce a plausible answer. A firm needs to know which question was asked, what information was used, whether the answer was within approved parameters and when the customer was directed to a human or a protected journey.

That is often described as governance. Customers experience it as competence.

If a digital assistant cannot explain why an answer changed, distinguish a market fact from an opinion or recognise that a question signals vulnerability, the weakness becomes visible in the interaction itself. The reputational cost will not be confined to a compliance report.

This requires investment in content operations as much as model selection. Product teams, compliance teams, investment specialists and customer-service leaders need a shared process for maintaining source material, reviewing high-risk answers and identifying questions that repeatedly produce misunderstanding.

The most valuable data may not be the chatbot’s headline usage total. It may be the unresolved questions: where users abandon a journey, request reassurance, misunderstand protection or ask versions of the same question after reading an answer. Those are signals of both consumer risk and unmet demand.

---

A better competitive response than warning labels

Warnings will remain necessary. The FCA is right to stress that general-purpose AI tools do not carry the protections attached to regulated advice. But warnings alone cannot restore the old order of financial discovery.

Consumers will continue using AI because it is immediate, conversational and available at the moment of uncertainty. Firms should assume that many prospective customers will arrive having already received an answer elsewhere. Their job is to add what a general-purpose tool cannot reliably provide: current product knowledge, accountable judgement, a view of the customer’s wider position and a regulated relationship.

That begins with a practical question for leadership teams: if a customer asks an AI tool about our category tonight, what will they misunderstand — and what would make it easy for them to verify the answer tomorrow?

The firms that answer that question well will not try to monopolise the customer’s research. They will make themselves the place where research becomes a defensible decision.