One Year Ago: Hype vs Reality

In early 2025, AI was everywhere but rarely reliable. ChatGPT could brainstorm ideas beautifully but couldn’t consistently follow complex instructions. When IFAs tried early automation experiments, common problems emerged:

  • AI hallucinated facts (invented regulations, products, or client details)

  • Multi-step workflows broke halfway through

  • Output quality was inconsistent---great on Monday, nonsensical on Friday

  • Compliance teams were rightfully sceptical: “We can’t put an AI in charge of suitability decisions.”

The perception was clear: AI was a toy, not a tool.

What’s Changed in 12 Months

Three major breakthroughs have fundamentally shifted the landscape:

1. Reasoning Models That Actually Reason

Today’s AI tools have been redesigned to work through complex problems step-by-step, showing their reasoning as they go---much like a colleague who talks you through their thinking rather than just handing you an answer.

What this means for you: When processing a suitability request for a complex inheritance strategy, the AI doesn’t just guess. It:

  • Breaks down client circumstances into logical chunks

  • Cross-references compliance rules explicitly

  • Flags edge cases that require human judgment

  • Shows its working so you can audit it

No black box. Full transparency.

2. AI That Learns Your Firm’s Way of Working

A year ago, if you wanted AI to behave in a particular way, you needed specialist technology teams and significant budget. Today, platforms let you upload your own material---house style, approved products, previous reports---and the AI learns your firm’s “voice” and boundaries in a matter of hours.

What this means for you: IFAgent trains on your house style, your product matrices, and your compliance approach. The AI doesn’t learn “generic suitability writing”---it learns your firm’s suitability writing. Reports feel like they were written by your team, because they were trained on your team’s work.

3. Consistent, Predictable Outputs

Early AI often returned rambling, hard-to-use responses. Today’s tools can be set up to always deliver information in a consistent format---placing the recommendation, rationale, and risk disclosure into the right sections automatically, every time.

What this means for you: AI can now slot seamlessly into your existing compliance workflows. The output isn’t a mess to sort through---it drops neatly into your compliance checklist and client management system, ready to review. It integrates, not disrupts.

Why This Matters for Compliance

The FCA’s Consumer Duty (2023) and recent guidance on AI in financial services (2024) centre on one principle: the firm stays responsible. Automation tools are fine---as long as you can audit them, override them, and prove human sign-off.

The old objection: “We can’t use AI because we can’t explain it.”

The new reality: Modern AI workflows are more auditable than manual processes. Every step is logged, every decision is visible, and you approve the final output before it leaves your firm.

Where We Are Now (Early 2026)

  • Quality: AI now passes 95%+ of first-review quality checks. Fewer rewrites, faster turnarounds.

  • Compliance: FCA-authorised firms are now using AI for suitability reports, compliance checking, and client communications---with proper oversight.

  • Cost: A typical suitability report costs £0.80-£2.00 to produce using AI, versus 2-3 hours of staff time (£150-£300).

  • Integration: AI tools now connect directly to CRMs, fund platforms, and compliance systems. Less copy-pasting, fewer errors.

The Next 12 Months

We’re watching three trends closely:

  1. AI that can read and listen --- Soon, IFAs will be able to hand over a scanned client fact-find or even a voice recording, and the AI will read or listen to it and pull out the key information automatically---no retyping needed.

  2. Real-time compliance checking --- AI will flag potential Consumer Duty issues during the advice conversation, not three months later.

  3. Regulatory precedent --- As more firms use AI successfully, FCA guidance will clarify the “safe harbour” for compliant automation. What’s experimental today becomes standard practice.

The Bottom Line

A year ago, AI required caution and scepticism. Fair enough---it wasn’t ready for serious work. Today, it’s ready. The firms adapting now aren’t early adopters chasing a trend---they’re competitive realists, freeing their best people from repetitive work and redirecting that capacity toward clients and growth.

If you’re still sitting on the sidelines waiting for AI to “mature,” it already has. The question now is: when will your firm catch up?