The work worth automating,
and the work worth deleting.

Most automation projects fail before any code is written: a process gets sped up that should not exist, or a number gets automated that two departments still disagree about. We work out the order first, then build the one job that pays back.

Talk about a job → What we automate

Four ways in

One of these is usually the first job. It is rarely the one people expect.

Automation auditWhat your team still does by hand, ranked by hours saved against effort to fix. A costed order of work, not a list of ideas.Read this →Broken automationThe script that stopped, the scheduled job that quietly fails, the integration nobody owns. We name the cause before quoting the fix.Read this →AI agents that actAn agent that reads from and writes to your real systems, with a guardrail that stops it doing damage and a way to check its own work.Read this →Invoice approvalA small system that reads supplier bills, checks them against your limits and brings anything unusual to you. Measured as a demo, sold as a two-week pilot.Read this →

The order that does not waste money

Define it, remove it, fix it, automate it, check it.

The order is the method. Step four is the only one anybody can see, which is why projects start there and then fail.

01Agree the definition
02Remove what should not be there
03Fix the data
04Automate the repetition
05Put a check on it

Step four is only safe once one to three are done. Skip them and the automation is correct while the answer is still wrong.

Before anything gets automated

Three questions, in this order. If the answers are not numbers, the project is a guess.

The automation audit

Does this process need to exist? A surprising share of briefs ask us to speed something up that only happens because of a decision nobody revisits. Deleting it is cheaper than automating it.
Who owns the output? If nobody owns it, nobody notices when the automation produces a wrong one.
What does it cost now? In hours and in mistakes. If that is not a number, the project has no way to succeed.

The long answer

Business process automation, without a consultant is the article version: where to start, what to delete first, and how to tell whether you need help at all.

The measured version

Invoice approval is the built example: a two-week pilot, with the numbers from the demo beside it.

Tell us what is not working.

Describe the problem in a few lines. You get a plain answer about whether it is worth fixing and what it would take, before any money changes hands.

Start a conversation →

The work worth automating,and the work worth deleting.