A better brief for your 2IC
Give your 2IC the business context it needs to propose useful work.
You do not need a technical specification to start. Give your 2IC enough business context to ask useful questions and prepare a plan you can confidently approve.
Start with the outcome
“Automate our customer service.” It does not say what arrives, what should happen or when a person should decide.
“Prepare replies to new sales enquiries using our service guide. Ask the sales manager to approve each reply before it is sent.”
Work through these questions
| Question | What to tell your 2IC |
|---|---|
| What starts the work? | A one-off request, a schedule or a supported business event |
| What should be achieved? | A specific result someone can check |
| Where is the information? | The business tools, folders, records and date range |
| Whose accounts should be used? | The authorised workspace member the work is for |
| Who does each part? | The available Agents and people responsible for decisions |
| What are the rules? | Required checks, business policies and actions that need approval |
| What if something is missing? | Who to ask and what must not be guessed |
| Where should the result go? | A document, folder, draft or supported business-system update |
Use this as a conversation starter
Help me plan a Sordit Job or Routine for this workspace. The business outcome is: [describe the result]. The work starts when: [request, schedule or event]. Use these sources: [tools and information], on behalf of: [person]. The result should be saved or delivered here: [destination]. These decisions need a person: [reviews or approvals]. If information is missing: [who to ask]. Check the available resources, ask me what is unclear and show the Task plan before creating anything.
Replace the bracketed parts with what you know. It is fine to say “I need help deciding” for the rest. The goal is a useful conversation, not a perfect first prompt.
Before you approve
- The proposal names the correct client workspace and account identity.
- Every Task has a clear responsibility and useful result.
- Sending, publishing and record changes happen only as intended.
- Human reviews go to the right people.
- Decision paths make sense, including requests for revision.
- The first run is small enough to review properly.