Will agentic AI finally shrink the advice gap?

People’s finances are becoming more complicated, from saving enough into a pension and making it last through retirement, to helping children and grandchildren meet their financial needs. There’s a real opportunity for the advice profession to improve the nation’s long-term financial resilience. But the profession can only do that if advisers have room for more clients.

The debate about how to close the gap often focuses on new ways to help consumers with their finances, like the FCA’s targeted support regime. It’s too early to say what that will do, but I think the best way to support more people is to give existing advisers more time to serve them.

The debate about how to close the gap often focuses on new ways to help consumers with their finances, like the FCA’s targeted support regime. It’s too early to say what that will do, but I think the best way to support more people is to give existing advisers more time to serve them.

In our 2026 Advice Efficiency Survey, two-thirds (66%) of advice professionals said they’d serve more clients if they had more time. A similar proportion (68%) cited manual data entry as their biggest challenge, and 30% spend three or more hours on a single report.

The impact of AI

AI is already winning some of that time back. Our survey found AI use increased from 43% in 2025 to 74% in 2026, with note-taking, transcription and report writing the most common uses.

The time savings created are worth having. Our analysis shows the AI meeting assistant we’ve built into our software can cut time spent on fact-finds and meeting reports by up to 85%. But generative tools work one task at a time. The adviser still has to start, check and move each case along. While the checking should stay with the adviser, handing the starting and moving to AI is where the next meaningful block of time will be saved.

That’s what agentic AI is for. It’s designed to run continuously in the background, scanning for anomalies, preparing advisers for their next client review and flagging where a decision is due. Because it works across a firm’s systems, it also takes on the job of moving information between them, so staff don’t have to do it manually.

Advisers describe what they want in plain English, such as preparing an annual review, and the AI agent works out the steps and runs them. If a process needs to change, the adviser just says so without rebuilding anything. The agent can prepare, send and track letters of authority, so nobody has to chase them. It can pull together valuations, performance and changes since the last review and draft talking points. It can also flag out-of-date or missing data before the client walks in.

None of that helps if firms don’t trust AI though, and trust depends on control. At intelliflo, we’re building three levels of control, so the firm decides how much the agent does, task by task. In co-pilot mode, the adviser reviews and actions everything. In supervised mode, the agent prepares the action but needs approval before anything is finalised. Autonomous mode is for low-risk, high-volume steps the firm has approved, with the agent completing the agreed task and reporting back. Firms can start hands-on and move a task up a level as their team gets comfortable with it.