AI anxiety, from the owner's chair: what to do when the work gets cheaper
Learning the tools is necessary, but it isn't a plan. Small-firm owners should decide where the time AI saves goes — and protect the value that doesn't show up in a dashboard.
Highly Distinguish · Published 28 September 2026 · 3 min read
"Learn AI or be replaced" is half an answer. The other half is deciding what your firm does with the time AI gives back.
If you run a small practice, you've heard both versions of the AI story. One says it will make your team twice as productive. The other says it will make your services a commodity. Both can be true at the same time, and that's what makes it uncomfortable.
A story about a successful makeover
The Hong Kong film Night King is set in a nightclub in 2012 Tsim Sha Tsui East. Its long-time manager runs the place on relationships, favours and decades of reading the room. Then a new CEO arrives with systems, targets and performance data.
The surprising part is that the makeover works. Results improve. The new system looks modern and efficient. And that success is exactly what makes the old roles easier to remove.
That's the uncomfortable shape of AI in a small professional firm too. The question isn't whether the new tools work. They often do. The question is what the business becomes once they do.
The question to ask about every task you automate
Pick one task your team is starting to do with AI — drafting client emails, summarising documents, preparing checklists. Then ask:
If this task becomes nearly free, who owns the next decision?
If the answer is "the client could now do it themselves", you've found work that's heading towards a commodity. If the answer is "a qualified person still has to judge, check and stand behind it", you've found where your value is moving.
Clients of accountants, brokers and migration agents rarely pay for typing. They pay for judgement, for someone who knows the rules, and for someone who takes responsibility when it matters. AI makes the typing cheap. It makes judgement and accountability more visible.
What a dashboard can't see
The film draws a sharp line between what an organisation can measure and what actually keeps customers coming back:
| Easy to measure | Hard to measure, easy to lose |
|---|---|
| Emails sent per hour | Whether the client felt looked after |
| Files closed per week | The detail an experienced staff member noticed |
| Turnaround time | Trust built over years |
| Process compliance | Knowing when the standard process doesn't fit |
Neither column is pure virtue. Measurement exposes wishful thinking, and "we've always done it this way" can hide real problems. But when AI pushes everything in the left column towards zero cost, the right column is what's left to compete on.
Four decisions for owners
1. Decide where the saved time goes before it disappears. If AI saves each person five hours a week, that time will be absorbed somewhere. Choose deliberately: more client conversations, better reviews, faster responses — or fewer hours, if that's what the business needs.
2. Keep people on the judgement calls. Let AI draft. Keep a person deciding, checking and signing off — especially on anything that reaches a client, a lender or a regulator.
3. Move your staff up, not out. The team members who learn to use AI well, check its work and explain the results to clients are worth more to you, not less. Train them.
4. Protect the relationships. Automate the reminders and the paperwork, not the moments where a client needs to hear from a person who knows them.
Learn the tools — then locate your leverage
None of this is an argument for avoiding AI. Firms that refuse useful tools on principle will simply be slower and more expensive. Learn the tools, use them enough to see where they're reliable and where they fail, and then make an honest decision about what your firm is for once the routine work gets cheaper.
This article is adapted for business owners from a longer essay, What AI Anxiety Really Means.