Pick one workaround your team dreads. We'll show you what it looks like when AI takes it on, on your own data, with your people still making the decisions.
Most businesses have at least one of these running through a spreadsheet and an inbox today.
AI does: matches transactions against the books and lists the differences.
A person decides: what each difference is, and when the month is closed.
AI does: spots approved quotes that haven't been invoiced, and drafts the invoice.
A person decides: whether the invoice goes out.
AI does: routes the request, reminds the approver, and records the answer.
A person decides: yes or no.
No requirements document, no AI strategy needed.
The spreadsheet, the email thread, and who does what today. 4 questions to start.
The records, the rules and the handoffs, so AI has something real to work on.
A working prototype on your own data for qualifying businesses. Go further only if it earns it.
The spreadsheet, the email thread, and who does what today.
The records, the rules and the handoffs.
A working prototype on your own data for qualifying businesses.
Keep the AI you already use. When it fits, we build with AnyDB, which works with ChatGPT and Claude and limits each AI to what that person may access.
AI prepares and chases. A person approves.
Anything that doesn't fit the rules goes to a person, with the context attached.
Payments, credits and write-offs stay with the people accountable for them.
Every action AI takes is logged, so you can see what it did and undo it.
Sometimes the right first job is no AI at all. If so, we'll say so.
Our own prospect research, follow-ups and reporting are moving onto the same kind of system we build for clients. We'll write up each job: how we set it up, what worked, and what stayed human.
So the first job we suggest for you is one we've handed over ourselves.
Follow the series in Field Notes →2-minute notes from the people who'd actually do the work with you.