The Smallest Automation Project Worth Doing

Small is the correct way to start

The first automation is not really about the money it saves. It is about whether your business ends up believing automation works. A visible, mechanical win produces a team that suggests the next five; an ambitious half-finished project kills the subject for two years.

So we scope first projects to be survivable if they are wrong: weeks rather than quarters.

What a small project includes

• One process, end to end, actually working
• One intake channel — the highest volume one
• The top three exceptions handled automatically, the rest routed to a person
• Monitoring, so silent failure is impossible
• Documentation and handover, not skipped because the project is small

What we do not cut, however small the budget: exception handling, testing and the ability to see when something has failed. Those are the difference between a system you trust and a demonstration.

What we cut to keep it small

1. Additional intake channels — phase two
2. The admin interface — a query does the job for the first months
3. Additional user roles — each one multiplies testing
4. Historical data migration — usually nobody looks at it
5. Integrations that are not on the critical path

Make it fund phase two

Measure for two weeks before. Automate the single most repetitive path completely. Measure again after four weeks. Use the difference to justify the next phase.

Businesses that do this find the second conversation is arithmetic rather than advocacy, which is why it happens.

Under £5,000

Below that, buy configuration rather than development: make the tools you already pay for work properly, and do the small automations that need no supplier at all.

We will tell you when that is the honest recommendation, and which twelve things to do first.

Frequently asked questions