Workflows
Agent Platform
Teach the agent your team's context
Save your scripts, notes and guides, so the agent works the way your team does.
The agent gets better at your robots as your team adds context. What one engineer works out helps the next person who asks.
When to use it
- Your team has analysis scripts it trusts.
- The agent keeps asking for the same thing, such as which topic holds the battery voltage.
- New staff need to work through the same steps your senior engineer would.
How to do it
- Save your scripts. Add analysis scripts and documents in the Scripts library tab of your organisation settings. Ask the agent to use one by name: "run our battery health script on this mission".
- Tell it what to remember. Say "remember that…" in the chat. Choose whether the note is for you or your whole organisation.
- Connect your guides. Connect Confluence, and the agent follows your triage guides and runbooks.
- Comment on missions and reports. Past comments and reports stay available to the agent.
- Let support cases build up. In a support role, each request becomes a case with its outcome. The next similar issue starts from what worked.
Example
Fictional example.
A validation team saves its MATLAB scripts for heading accuracy in the Scripts library. Their recurring report now runs those scripts on each new test, so the numbers match the ones the team has always signed off.
Tips
- Tell the agent how to read a result, not only which script to run: "a heading error over 0.5° on a straight run means a mounting problem".
- Add a short note to each script: what it checks and what a good result looks like.
- If the agent gets something wrong, correct it in the chat and ask it to remember the correction.
- Save terms your team uses, such as robot names, site names and topic names.