Practical Templates

Transform vague instructions like "review this contract" into clear, delegable specifications. Use this before delegating work to junior lawyers, AI, or external counsel.

TASK: [One-line description]

CONTEXT:
- Why this task exists: [Business reason]
- Who requested it: [Stakeholder]
- When it's needed: [Deadline and urgency]

OUTPUT REQUIRED:
- Format: [Email, memo, marked-up document]
- Length: [One page, detailed analysis]
- Audience: [Who will read this]

SCOPE:
In scope: [Specific issues to address]
Out of scope: [Issues to ignore]

SUCCESS LOOKS LIKE:
- [Specific, measurable outcome]

Document the assumptions underlying your legal advice. Use when giving advice that depends on facts, law, or commercial context.

ADVICE: [One-line summary]

FACTUAL ASSUMPTIONS:
- [What facts we're relying on]
- [What we haven't verified]

LEGAL ASSUMPTIONS:
- [Our interpretation of the law]
- [Areas of legal uncertainty]

COMMERCIAL ASSUMPTIONS:
- [Client's objectives]
- [Risk appetite]

IF ANY OF THESE CHANGE:
- [Specific triggers for review]

Document processes that currently live in people's heads. Use to capture knowledge and make your team more resilient.

PROCESS: [Name]

PURPOSE: [Why this exists]

TRIGGER: [What starts this process]

STEPS:
1. [Action] - [Who] - [Time]
2. [Decision point] - [Criteria]
   - If yes: [Next step]
   - If no: [Alternative]

OUTPUTS:
- [What gets produced]
- [Where it's saved]

COMMON EDGE CASES:
- [Scenario]: [How to handle]

Create structured feedback mechanisms to know if your work is actually useful. Stop operating blind.

FEEDBACK MECHANISM: [Name]

WHAT WE'RE MEASURING:
- [Specific aspect of quality]

HOW WE'LL MEASURE IT:
- Method: [Survey, interview, metric]
- Frequency: [When]
- Who provides feedback: [Stakeholder]

QUESTIONS TO ASK:
1. [Specific question about usefulness]
2. [Specific question about clarity]
3. [Open: what could be better?]

WHAT WE'LL DO WITH FEEDBACK:
- Review: [Who, how often]
- Action: [What triggers change]

Document decisions about standardising templates, clauses, or processes. Capture the rationale for future reference.

STANDARD: [Name]

DECISION DATE: [When]

PROBLEM WE'RE SOLVING:
- [What inconsistency existed]
- [Cost of not having standard]

ALTERNATIVES CONSIDERED:
1. [Option 1] - [Why rejected]
2. [Chosen option] - [Why chosen]

STANDARD APPROACH:
- [What the standard is]
- [When to use it]
- [When it's okay to deviate]

RATIONALE:
- [Why this approach]

Assess whether a task is ready for AI assistance. Use before piloting AI on any task.

□ TASK DEFINITION (3 checks)
  □ Clear instructions?
  □ Objective definition of "done"?
  □ Edge cases listed?

□ PROCESS (3 checks)
  □ Process documented?
  □ Process consistent?
  □ Current error rate known?

□ DATA (3 checks)
  □ Inputs standardised?
  □ Precedents consistent?
  □ Terminology standardised?

□ RULES (3 checks)
  □ Decision criteria explicit?
  □ Assumptions documented?
  □ Exceptions defined?

□ FEEDBACK (3 checks)
  □ Quality currently measured?
  □ "Good" defined?
  □ Test cases available?

□ JUDGEMENT (3 checks)
  □ Judgement separated from routine?
  □ Human review points clear?
  □ Escalation triggers defined?

SCORE: [X]/18
15-18: Ready for AI pilot
10-14: Fix gaps first
0-9: Not ready, focus on process