AI Transformation for Small Business: What It Means When You Have 20 to 200 People

"AI transformation" is a phrase built for large companies. It comes with operating models, centres of excellence and multi-year roadmaps. In a business of 20 to 200 people, almost none of that applies, but the underlying idea does: the business should work differently because AI can now do a lot of what people did by hand. This is what that looks like at your size.

Stage 1: everyone has a governed tool

Every member of staff has access to a business-tier assistant, under company control, with a clear usage policy and a spending ceiling. Personal accounts for work stop. People have been shown how to use it for their own job.

This is the foundation and it is quicker than people expect. With a clear decision on provider and someone who has done it before, it takes about 30 days. Our 30 day implementation plan covers it step by step.

You are done with stage 1 when at least half your staff use the tool weekly for real work, and nobody is using personal accounts for client material.

Stage 2: a few workflows change

This is where most of the value is, and where most businesses stall. Individual productivity gains are real but scattered. The bigger gains come from picking a workflow that happens many times a week and redesigning it with AI in the middle.

Good first candidates share three features: they are frequent, they are mostly text or documents, and the output is checked by a person anyway. Examples:

• Turning a client meeting into a file note, action list and follow-up email.
• First drafts of proposals or tender responses from a library of past wins.
• Triage of a shared inbox into categories with suggested replies.
• Monthly reporting commentary from the numbers.
• Onboarding packs for new clients or employees.

Pick three. Redesign each one end to end, write down the new way, and measure time before and after. See AI for operations and back-office teams for more candidates.

You are done with stage 2 when three workflows run the new way by default and you can say how many hours a month they save.

Stage 3: the business reorganises

Once several workflows have changed, harder questions follow. Does the junior role that spent half its time on first drafts still look the same? Can you offer a service at a price that was not viable before? Should the next hire be another person for the same work, or someone who builds and maintains automations?

This stage is slower and is driven by the business, not the technology. It is also where agents, which act across systems rather than just drafting text, start to matter. We cover what that means for cost and control in AI agent spending limits and AI agent permissions.

Why businesses stall after stage 1

• Nobody owns stage 2. The rollout project ended, and workflow redesign was never anyone's job.
• The workflows chosen were too big. "Automate finance" is not a workflow. "Draft the monthly variance commentary" is.
• No baseline was measured. Without a before number, nobody can show the after, and the effort loses support.
• The partners or directors are not using it. Staff read the signal. See getting partners to back an AI rollout.

What you do not need

• A centre of excellence. You need one owner with a few hours a week.
• A multi-year roadmap. The tools change too quickly. Plan a quarter at a time.
• A custom AI platform. Business-tier assistants with good configuration cover most of stage 1 and 2.
• A six figure consulting engagement. See AI transformation without a consulting firm.

For the sequence in more detail, read the AI transformation roadmap for a 20 to 200 person company. For budget, see how much AI transformation costs.

Frequently asked questions