This Week in Brief

This week’s news is less about new AI tools and more about how firms and clients are actually using — or holding back from — the AI that’s already out there. One adviser has built himself an AI “digital twin” for round-the-clock client contact, new research shows most investors still aren’t using AI for their investments despite advisers embracing it, and a wider survey finds firms are deliberately pumping the brakes on new tech rather than chasing every development.

Key Developments

An Adviser Cloned Himself With AI — and Clients Are Using It

Dubai and Japan-based adviser Adrian Rowles has spent 12 months building “Virtual Adrian,” a self-hosted AI video/voice clone of himself that will be available to clients 24/7 once it launches later this year. The clone pre-screens enquiries and answers existing clients’ follow-up questions, but Rowles is clear that it stops short of anything regulated: fact-finds, recommendations and implementation still happen with him in person. He built it on his own server specifically to avoid sending client data to third-party AI systems — worth noting for any firm considering something similar, given data protection and Consumer Duty obligations don’t disappear just because the “adviser” on screen is a clone. (FT Adviser, 24 September 2026)

Most Investors Still Aren’t Trusting AI With Their Money — Even as Advisers Lean In

M&G’s Investor Compass report found 63% of investors are not using AI to help with their investments, and only 13% do so regularly. Advisers, by contrast, are already using it heavily (75%), mainly to save time on admin rather than to change what they tell clients. M&G’s Ross Liston made the point plainly: “faster is not always better when people are making major financial decisions” — a useful line for any adviser explaining to a client why the human conversation still matters alongside the tech behind the scenes. (FT Adviser, 25 September 2026)

Advice Firms Are Deliberately Slowing Down on New Tech, Not Speeding Up

New research from NextWealth (based on over 900 advice professionals) found just 13% of firms plan to add a new technology partner in the next 12 months, down sharply from 34% in each of the last two years. NextWealth’s Heather Hopkins says this isn’t firms falling behind — it’s a deliberate strategy, with smaller firms in particular choosing to let bigger firms “make the mistakes” first. Separately, the same report found perceived value for money — not integration or service — has the strongest link to a platform’s asset growth, with Fundment rated top overall. (Professional Adviser, 24 September 2026)

From the Trade Press

  • Update: AI Still Gets Money Questions Wrong More Often Than Right — FT Adviser’s deeper look at Saturn’s AI-accuracy research (first covered in a previous digest) found 17 leading AI models answered only 43% of money-related questions correctly overall, dropping to 12% on hard questions. Free models performed markedly worse (37% accurate) than paid ones (51%). FT Adviser, 21 September 2026

  • Wealthier Clients More Likely to Trust AI With Their Investments — Entrepreneurs and higher-net-worth clients are more inclined to trust AI-generated investment views than other investors, this week’s coverage found — a segmentation angle worth watching for firms targeting that client base. FT Adviser, 23 September 2026

What to Watch

Rowles’ self-hosted “digital twin” is one of the first concrete UK examples of an AI avatar operating in the pre-advice space — worth watching whether other advisers follow suit, and whether the FCA’s ongoing review of the regulatory perimeter has anything to say about AI tools that sit just in front of, but not inside, the regulated advice process.

Sources: FT Adviser, Professional Adviser, web search. Compiled 25 September 2026.