Product & engineering
Specs, release notes and code that follow the team’s conventions in every tool.
The decisions you keep explaining
Product and engineering teams use the most AI tools side by side: chat assistants for specs and notes, coding agents for code. Each one has to be told the team’s conventions separately.
- “Release notes describe what changed for the customer, not the implementation.”
- “Every spec has a section on what is out of scope.”
- “Commit subjects are imperative, under 72 characters.”
- “Use pnpm in this repository, never npm.”
With Decision Cartography
Said once, followed everywhere
A correction made to one assistant is saved as a decision, and every assistant that person uses follows it.
A rulebook for the team
The team’s writing and code conventions, in chat assistants and coding agents alike.
New people start ahead
A new engineer’s agents follow the repository’s conventions from the first commit.
Other teams
Sales
Proposals, follow-ups and outreach that sound like your team, whoever writes them and whichever AI drafts them.
Customer support
Replies that keep the team’s promises and tone, from every agent and every AI that helps them.
Marketing
One voice across every post, page and email, however many people and tools write them.
Operations
Processes the team agreed on, followed by every assistant that touches them.
Human resources
People communications that are consistent, careful and on policy, whoever drafts them.
Learning & development
Training material in the team’s format and level, from every author and every AI.