Blog/Opinion
Pick the boring workflow first
The workflow that annoys the founder most is usually the worst first project. The one nobody talks about is usually the best.
1 min read · · Boring Agents
- 1Judgement-heavy work makes an impressive demo and a bad first deployment.
- 2A good first workflow repeats, has a right answer you can write down, and is countable.
- 3Saying no early is cheaper for everyone than a pilot that quietly dies.
The most common first request is to automate whatever annoys the founder most. Usually that thing needs judgement, context and three unwritten rules held by one person who has been there eleven years.
That's a terrible first project. Not because it can't be done, but because nobody can tell whether it's working. Without a right answer there's no grade, and without a grade the pilot drifts until someone stops paying for it. MIT's 2025 study of generative-AI pilots found most of them never reached production; the ones that did picked one narrow problem and did it well.
Good first workflows are boring. They repeat. They have a right answer you can write on a card. They happen often enough that a week of results means something. Chasing documents, answering the same twenty questions, keying a form into a system. Start there, get a grade, then earn the harder work.
Tell us the job your team hates. We'll tell you if AI can do it.
A 30-minute call with the people who would build it. No technical knowledge needed: just describe the job, how often it happens, and how you check it's done right.