How a run works: briefing, then finished asset
A run has two steps. pmmAgent.ai writes the briefing, then your own AI writes the finished asset from it.
The two steps
- Briefing. You give a short brief and choose a tool. pmmAgent.ai builds a structured brief in the way a senior product marketer would: the facts it can rely on, the structure the asset should follow, the constraints, and the gaps it couldn't fill. It draws on your brief, your Brand Box and your project.
- Finished asset. Your AI, such as Claude or ChatGPT, takes the briefing and writes the battle card, the messaging, the email sequence or whatever you asked for.
pmmAgent.ai doesn't write the finished asset for you. That keeps the work in the AI you already use, with your files and your conversation around it.
What goes into a briefing
- Your brief. What you type, or what you ask your AI to send.
- Brand Box. Voice, messaging and guidelines, and visual identity. See The three parts.
- Project. The product or launch you're working on, including any PRD or launch brief you added.
Each run tells you which brand parts it used, so you can check it. See See which brand parts a run used.
On the web
You write the brief in the app and get the briefing back. Copy it, open it in Claude or ChatGPT with one click, or download it. Then ask your AI for the finished asset.
In your AI
With the connector, your AI sends the brief to pmmAgent.ai, receives the briefing and writes the finished asset in the same chat, in one step. The reply names the workspace the run used and ends with how many runs you have left. The connector is on Solo, Pro and Team. See What the connector does.
What counts as a run
Each time you run a tool, it uses one run. Pro tools use a Pro run. See Runs and Pro runs: how counting works.
Next - Web app or your own AI: which to use when - Write a brief that gets a strong result