Approve once. Runs every week.
For go-to-market leaders · Design partners wanted
Say it once.The work does itself.
Every week, every team asks for the same things in the same meetings. Rulebook listens, learns the decision behind each ask, and builds the automation that does the work, in the tools you already use, with a record of who asked and why.
The problem
Leaders say the same things every week. Nothing remembers.
Update the CRM. Check the PO. Escalate that ticket. Every team has asks like these — repeated in every meeting, because nothing carries them out once the call ends.
Illustration — one ask, eight weeks
How it works
One spoken ask, from a call to a running automation.
- 01
It hears what you keep saying
In four weekly calls the CRO asked for the same thing: pause any campaign under 2x. Rulebook finds those moments and writes down what he saw, why it matters, and what he asked for.
- 02
It writes the rule. You can change it.
A plain sentence with numbers you can edit. Drag 2x to 3x and the backtest shows, instantly, what it would have done last quarter.
- 03
It reads your docs and checks your tools
It searches Notion, Confluence, Slack and Drive for your policies and how the work is done today. It uses an API or connected tool wherever one exists, and asks you only when there isn’t one.
- 04
It shows exactly how it will run
The trigger, the Claude skill, every tool, what it can touch and what it will never do. All before you approve anything.
- 05
Approve once. It sets itself up.
It writes the skill, connects the tools, schedules the trigger and wires the notifications. Nothing for you to configure.
- 06
Next time, you’re told.
When the rule fires, the owner gets a notification in the app and an email with a link to what it made. Every action says who asked, and why.
The real app, replaying a synthetic quarter · scroll to drive it
Any tool, any team
Claude works out how to do it, with whatever you already use.
It doesn’t pick from a menu of integrations. For each step it uses the most reliable way available, in this order, and it only asks you to show it something when there is no API or connected tool.
Claude asks you to show it once: a short screen recording or a Loom. It turns that into a browser script, tests it in a sandbox, and you can read it before it runs.
Every team that repeats itself
The same asks, every week. Now they run themselves.
Trust is earned
Ambitious work. Trust earned one step at a time.
On Aug 26, Mike Donovan, CRO said “Usage down thirty percent or no sponsor, inside ninety days, written plan inside a week, exec call booked.”
Every action has a receipt
Who asked, when, and why — attached to every reminder, task and update.
You’re always in control
Rules are versioned. Edit, expire, or switch any of them off in a click. Conflicts get flagged.
Not surveillance
We turn leaders’ own instructions into help at the right moment. Fewer nags, not more monitoring.
Questions
The ones leaders ask first.
How does it know how to do the work?
It reads how your team already does it: your Notion, Confluence, Slack and Drive, plus past examples like old redlines or recaps. It uses an API or connected tool wherever one exists. Only where there isn’t one does it ask you to show it, then it drives the browser or writes code, tested in a sandbox first.
What if my docs and my leaders disagree?
It flags the mismatch and asks. For example: your pricing policy says Deal Desk approval above 12%, but your CRO said 15%. You choose which to enforce, and it can propose an edit to the policy page.
Which teams is it for?
Anyone with a recurring call and repeated asks. We start with go-to-market: sales, marketing, customer success, support, deal desk and legal. Then product and engineering.
What can it touch? What will it never do?
Before you approve, every rule lists exactly which tools it can use and with what access, and what it will never do, such as sending anything to a customer, publishing a private offer, or approving a discount.
Will it do things I didn’t approve?
No. Nothing runs until the right person approves the rule, and risky work always waits for a human. New rules start by notifying or drafting only.
Do you record our calls?
No. We read transcripts from the call tools you already use. Your team decides which calls we see, and everyone on them is told.
Is this tracking my team?
No. It turns a leader’s own instructions into finished work. It’s built to cut nagging and rework, not to score people.
What if I change my mind?
Edit a rule, set an expiry, or switch it off. Every version and every action is logged.
Design partners
Be the team that sets the standard.
We’re working closely with a handful of go-to-market teams. You’ll shape what it does, and we’ll set it up with you by hand. Best fit: B2B software, 50–500 people, a recorded weekly call.
- Hands-on setup with the founder
- Your repeated asks, turned into finished work
- Direct say in what we build next