People are increasingly talking to AI chatbots about their finances [1][2]. Services like ChatGPT (by OpenAI) and Claude (by Anthropic) can research, recommend, and take action.
Financial intelligence is also offered by:
- Advice firms: Human-powered, advanced, but costly.
- Fintech firms (consumer apps, challenger banks): Digital-first, basic, with very low marginal costs.
Both segments are investing heavily in AI:
- Advice firms: Boost efficiencies and reduce costs.
- Fintech firms: Increase intelligence to recommend tied products, retain customers, and increase AUM.
As UK fintech firms increase the intelligence they offer, they quickly find they may need to become authorised by the regulator. The FCA broadly categorises services as:
- Information (unregulated): Generic educational content
- Tools (unregulated): Calculators to explore goals and scenarios
- Guidance (unregulated): Personal prompts that can suggest product types
- Targeted support (regulated): Group-based recommendations, i.e. “people like you”
- Financial advice (regulated): Personal recommendations that use specific products, either in a focused area, or multiple areas [3].
AI chatbots disclaim “this isn’t financial advice” when users prompt advice-like behaviour. Many question whether this cuts it, however in its April 2026 report the FCA announced that general-purpose AI chatbots fall outside its “perimeter” — they don’t need to facilitate their supervision, evidence suitability, or put governance frameworks in place, amongst other things [4][5]. The perimeter may change in future [6], and AI chatbots may purposefully step into it. OpenAI acquired Hiro then announced financial guidance (“Personal Finance in ChatGPT”) in the US market [7][8]. This offers little consolation to UK firms finding themselves regulated and in competition.
It’s impossible to ignore the wider context of the global AI revolution. The FCA’s perimeter is set by a government that wants the UK to be an attractive place for frontier AI firms to build and commercialise [9]. Anthropic has been courted by the London Mayor to expand operations, Google DeepMind has over 1,000 London-based employees, and OpenAI has just signed a lease in King’s Cross [10][11][12]. This is a critical period — any perceived regulatory friction could risk momentum.
Convergent evolution is seen in nature when different organisms in similar environments find similar solutions. An internet-famous example is “carcinisation”; five different groups evolved into the shape of a crab [13]. Our environment is causing different firms to evolve; AI chatbots, fintech apps, challenger banks, and advice firms, are racing towards a future that integrates:
- A total, live view of client financial data
- Whole-of-market access to financial products
- Advanced intelligence across different areas, across the client lifecycle
- A personalised, accessible, delightful interface
- Scalable governance systems
Here’s how transformation might look:
- Client data: Human-led fact finds, digital fact finds, and Open Banking connections → OS-native, agentic AI that can access apps and inboxes, and share data with trusted parties.
- Financial products: Tied or preferred products, platform constraints, and high-effort transfers → whole-of-market catalogues, dynamic personalised products, and agentic execution.
- Intelligence: Human advice and digital guidance → digital advice with seamless human-agent collaboration for complex cases.
- Interface: Meetings, advice documents, and digital apps → personalised, dynamic UIs, conversational interaction, and access to human advisers or coaches.
- Governance: Human-led checks and automated boundary checks → scaled automated checks, intelligent client classification, end-to-end audit trails, and human oversight.
It’s all to play for. Each player has tailwinds:
- Challenger banks: Core services give live client data, trust to move money, and a flexible business model.
- Fintech firms: Serving focused needs means they’re well positioned to move the fastest and build the most delightful experience.
- AI chatbots: Likely the first place to ask financial questions. They also benefit from early access to frontier lab technology.
- Advice firms: High-end intelligence will resist disruption. Collaboration opportunities are emerging to provide hybrid offerings.
But they have headwinds, too:
- Challenger banks: Change at such a scale is unlikely to be quick. Their products may become entries in someone else’s catalogue.
- Fintech firms: Serving focused needs has been a winning strategy, but if they can’t solve the full picture they risk becoming just a feature.
- AI chatbots: They’re new to financial regulation. Research, money movement, and product usage has no space for error — agents still need oversight.
- Advice firms: Entry-tier services, such as workplace offerings, are particularly vulnerable to the increased intelligence of other players.
And so, if you find a discarded shopping list in London, here’s what you might find on it: OS-level context-as-a-service, Open Finance connectivity, product and wrapper rails, cash and savings rails, transfer infrastructure, research tools, modelling systems, automated advice engines, and governance technology.
The FCA has a strong reputation; it has drawn a perimeter that gives fair notice to frontier labs and provides certainty. The grey area is if “financial intelligence" isn’t packaged as single-purpose. As intelligence increases, it will start to look, swim, and quack like financial advice. Regulatory tools may need to evolve — could we see FCA-published eval suites, or agentic mystery shopping in future? The grey area is uncomfortable, and any grumbles are understandable. But the logic is hard to escape: financial intelligence is increasing, consumers use it, and the firms that cannot assemble the pieces to compete may find themselves becoming components in someone else’s system. The only option is to build.
About Multiply
Multiply is a financial advice technology platform that helps regulated firms deliver personalised financial guidance, targeted support, and fully regulated advice at scale. It was the first UK company to provide fully automated financial advice as an authorised firm. Its technology is used by advice firms to make human advisers more efficient, and by fintech firms to power financial intelligence inside their own customer experiences. Multiply's product suite includes a proprietary Advice Engine, Risk System, and agents that can author reports, perform pre-checks, and extract information about a client. It gives personalised recommendations based on client data, product catalogues, and house views. It can give guidance, targeted support, and full-fat advice, across both focused and holistic areas.
About the author
Tim Elder is Chief Product Officer at Multiply AI, where he leads product strategy across regulated financial advice, AI agents, and adviser technology. He has spent his career building products across fintech, startups, scale-ups, marketplaces, and complex regulated environments. Before Multiply, he built products for Google, founded multiple startups, and advised early-stage founders through Product Surgery. He was also a Director at Startup Grind London, and attended Stanford GSB’s Emerging CPO programme. On weekends, he forages mushrooms and tends his allotment.
Sources
[1] Anthropic analysed 1M Claude conversations from Mar–Apr 2026 to find ~6% involved personal guidance, with personal finance making up 11% of those guidance conversations. Source: Anthropic, “How people ask Claude for personal guidance”, 30 April 2026, https://www.anthropic.com/research/claude-personal-guidance (Accessed: 19 May 2026).
[2] Ofcom used outgoing-referral data to determine that finance accounted for 3.8% of UK ChatGPT outgoing referral topics in the 12 months to August 2025. Source: Ofcom (2025), “Online Nation Report 2025”, 10 December 2025, https://www.ofcom.org.uk/siteassets/resources/documents/research-and-data/online-research/online-nation/2025/online-nations-report-2025.pdf (Accessed: 19 May 2026).
[3] The FCA and HM Treasury’s Advice Guidance Boundary Review (AGBR) explores the boundary between guidance, targeted support, simplified advice, and holistic regulated financial advice. Source: FCA (2025), “Targeted support and simplified advice: proposals for closing the advice gap”, 30 June 2025, https://www.fca.org.uk/publications/consultation-papers/cp25-17-targeted-support-simplified-advice-proposals-closing-advice-gap (Accessed: 19 May 2026).
[4] The FCA’s Perimeter Report says consumers are increasingly asking general-purpose LLM platforms for help making financial decisions, but that they are not receiving regulated advice and do not currently have Financial Ombudsman or FSCS protections. It also notes that an LLM deployed specifically to provide financial advice would likely fall within the FCA’s perimeter. Source: FCA (2026), “FCA perimeter report”, 29 April 2026, https://www.fca.org.uk/publications/corporate-documents/fca-perimeter-report (Accessed: 19 May 2026).
[5] Firms authorised to give regulated advice must comply with the FCA Handbook, including requirements on suitability, governance, systems and controls, customer outcomes, and cooperation with the regulator. Source: FCA Handbook, https://www.handbook.fca.org.uk/ (Accessed: 19 May 2026).
[6] Source: FCA (2026), “Mills Review to consider how AI will reshape retail financial services”, 27 Jan 2026, https://www.fca.org.uk/news/press-releases/mills-review-consider-how-ai-will-reshape-retail-financial-services (Accessed: 19 May 2026).
[7] Source: The Economic Times, “OpenAI acquires AI personal finance startup Hiro”, 14 April 2026, https://m.economictimes.com/tech/startups/openai-acquires-ai-personal-finance-startup-hiro/articleshow/130251086.cms (Accessed: 19 May 2026).
[8] Source: OpenAI, “A new personal finance experience in ChatGPT”, 15 May 2026, https://openai.com/index/personal-finance-chatgpt/ (Accessed: 19 May 2026).
[9] The FCA explains and enforces the regulatory perimeter, but does not decide where it sits. The legal boundary is set by the Government and Parliament, with HM Treasury responsible for shaping the legislation and working with regulators to update the framework over time. Source: FCA (2026), “FCA perimeter report”, 29 April 2026, https://www.fca.org.uk/publications/corporate-documents/fca-perimeter-report (Accessed: 19 May 2026).
[10] Source: The Times, “Sadiq Khan woos CEO of AI firm blacklisted by Trump”, 6 March 2026, https://www.thetimes.com/uk/london/article/sadiq-khan-anthropic-dario-amodei-trump-03mv0zp9c (Accessed: 19 May 2026).
[11] Source: Business Insider, “UK Google DeepMind employees vote to unionize over its AI defense deals”, 5 May 2026, https://www.businessinsider.com/google-deepmind-employees-unionize-vote-ai-military-contract-uk-2026-5 (Accessed: 19 May 2026).
[12] Source: The Times, “OpenAI signs lease on new London office”, 12 April 2026, https://www.thetimes.com/business/technology/article/openai-double-uk-workforce-london-office-hxqjj702f (Accessed: 19 May 2026).
[13] Source: Wolfe, J.M., Luque, J. and Bracken-Grissom, H.D. (2021), “How to become a crab: Phenotypic constraints on a recurring body plan”, BioEssays, 43(5), https://pubmed.ncbi.nlm.nih.gov/33751651/ (Accessed: 19 May 2026).
