The moments that matter: what not to automate

The best question about automation is not only what a system can do. It is what the firm should still do as a human service. Accounting and bookkeeping depend on trust, interpretation and judgement. Removing repetitive work is valuable. Removing the moments through which people learn and clients feel understood is a different proposition.

The Moments That Matter explores that boundary. A routine request for a missing document and a conversation about an impending cash shortfall may both look like communication tasks. They are not interchangeable. One can follow a clear process. The other may need someone to listen, understand the stakes and take responsibility for the next step.

The paper offers three ways to think about the work. Automate repeatable mechanics with clear controls and exception handling. Augment judgement where technology can organise evidence or prepare a draft, but a person must interrogate it. Keep consequential, ambiguous or emotionally sensitive moments human-led. The framework is a decision aid, not a claim that every task belongs permanently in one category.

Staff development matters too. Re-keying figures is not an apprenticeship worth preserving for its own sake. But learning why a number moved, spotting an unusual assumption and explaining uncertainty are essential. Firms can redesign learning by asking juniors to form a view before reading an AI explanation, discuss exceptions and observe how experienced advisers handle difficult conversations.

This is also what separates advisory from a business owner asking AI about their data. Advisory workers do not merely deliver an answer. They listen for the real concern, challenge a course of action and help the client carry it through. Clients pay for considered judgement, professional accountability and a relationship that remains present when the decision becomes difficult.

Start by identifying a handful of client moments you would regret automating badly. Give each one a named owner, a clear escalation path and enough protected time. A modern firm uses infrastructure and AI to become more available, not more distant. Read this alongside the first two papers: build capacity, use it for meaningful advisory and protect the human work that makes that advisory valuable.