This Week in Brief

This week’s news is less about flashy new launches and more about the foundations that make AI worth trusting. Pensions giant Aon put AI retirement guidance head-to-head with real adviser recommendations and found a genuinely mixed picture — useful, but capable of confident-sounding mistakes. Separately, new research shows many advice firms still aren’t satisfied with the data they get from platforms, which matters because shaky data makes for shaky AI. And on the security side, a sharp rise in cyber attacks on platforms is fuelling calls for advisers to question every contact and confirmation, not just the obviously suspicious ones.

Key Developments

Aon Research: AI Could Help Close the Advice Gap — But Comes With Real Risks

Aon has published a detailed comparison of AI-generated retirement guidance against the advice real members actually received. The motivation is stark: only 8.6% of UK adults received regulated financial advice in 2024, yet 73% now use AI tools daily and as many as 80% say they’d consider using AI for pensions or investment decisions. AI is free, always on, and instantly available, which is exactly why it’s filling a gap advisers haven’t been able to reach.

The results were mixed. At its best, Aon found AI gave clear, intuitive explanations that sat well alongside a human conversation. At its worst, it produced incorrect calculations, inconsistent recommendations, and a tendency to present speculation as fact — delivered with the same confident tone whether it was right or wrong. Aon’s advice to pension schemes applies just as well to advice firms: encourage clients to verify anything AI tells them against scheme or adviser resources, and engage people earlier, before they’ve mentally settled on a decision an AI tool talked them into.

Data Quality Is Holding Back Advisers’ AI Ambitions

A new NextWealth report finds that 61% of advice firms are only “somewhat satisfied” with the data they get from platforms, and 54% feel the same about pension providers — proof, as the report’s author put it, that the data is “being tolerated, not celebrated.” The recurring complaints are messy transaction data, inconsistent formatting, and patchy management information, problems familiar to anyone who has tried to get a clean client view across multiple providers.

This matters well beyond admin convenience. Any AI tool — from a suitability-report checker to a meeting-note summariser — is only as good as the data it’s working from, and firms are increasingly turning to third-party data aggregators to plug the gap rather than waiting for platforms to fix it themselves. The largest, fastest-growing firms now treat data quality as a non-negotiable when choosing who to work with, and the Platforms Association is launching a data charter later this month aimed at addressing exactly this.

Advisers Urged to Rethink Security as AI-Powered Scams Get More Convincing

An FT Adviser opinion piece this week makes the case for “zero trust” — verifying every request rather than assuming a familiar voice or email is genuine — pointing to increasingly sophisticated AI-driven scams that can convincingly impersonate clients or colleagues. It lands alongside separate figures showing 55% of platforms have seen cyber attacks increase over the past year.

Neither story is about a specific new product, but together they’re a practical prompt to check your firm’s verification habits: are you confirming unusual instructions through a second channel, and would your team spot an AI-generated voice or message designed to look exactly like the real thing?

From the Trade Press

  • “Will pension firms become subject to a robot tax?” — A thought-provoking look at proposals, backed by figures like Bill Gates, to tax the “labour” performed by AI as it replaces human roles. If adopted, it could blunt some of the cost savings firms are currently banking on from AI adoption. Professional Pensions, 18 June 2026

  • “Tatton: ‘our in-house coders use AI to deliver upgrades in weeks’” — The investment platform says using AI in its own software development and investment management teams has sharply cut the time needed to ship upgrades. A reminder that AI’s impact isn’t limited to client-facing tools. FT Adviser, 16 June 2026

  • “Pensions industry urged to adopt ‘pub fight approach’ to AI” — Delegates at the Pensions UK Local Authority Conference were told not to let AI distract from the fundamentals of good member service — engage with it deliberately, not defensively. FT Adviser, 16 June 2026

What to Watch

The FCA’s AI Input Zone survey closed today, 19 June — its findings will help shape whatever guidance comes next. The Mills Review remains on track to report to the FCA Board this summer, and the Platforms Association’s new data charter is due later this month. Intelliflo’s agentic AI features, covered here last week, are still expected to begin rolling out later this year.

Sources: Professional Pensions, Professional Adviser, FT Adviser, Aon, NextWealth, web search. Compiled 19 June 2026.