Free fillable and printable worksheet

Map the work before you choose the AI.

Turn “we should automate this” into a workflow your team can examine, practice, and improve. Start with the people, evidence, exceptions, and review points—not a product tour.

The goal: leave with one bounded workflow, one accountable human-review gate, and one small experiment whose usefulness can actually be judged.

Fill out the worksheet

No email gate. Nothing you type is uploaded, saved, or sent to TrainedToThink.

A practical sequence

Three questions make the conversation useful.

What actually happens?

Name the trigger, source material, people, handoffs, ordinary steps, exceptions, and finished result.

Where is judgment doing work?

Find the moments when someone interprets ambiguity, checks evidence, resolves an exception, or accepts responsibility.

What could we test safely?

Choose a small slice, representative non-sensitive examples, a reviewer, a success measure, and a stop condition.

The worksheet

Map one recognizable workflow.

Short, concrete answers are better than polished language. If two people disagree, record the disagreement—it is part of the workflow.

TrainedToThink workflow map

One workflow. One useful next step.

Use ordinary language. Do not enter confidential, regulated, or personally identifying information into a browser tool your organization has not approved.

1. Workflow name Use a name people on the team recognize—not a technology label.

2. Accountable owner Who is responsible for the finished result, even if several people or tools contribute?

3. Trigger What event, request, deadline, or condition starts the work?

4. Inputs and source material Which documents, messages, records, policies, systems, or observations are used?

5. People and handoffs Who touches the work, in what order, and where does it wait or change hands?

6. Ordinary path Describe what usually happens in four to seven plain-English steps.

7. Exceptions and judgment What cannot be handled by a simple rule? Where do experience and context matter?

8. Human-review gate Who checks what evidence before the result can affect people, money, policy, or customers?

9. Useful finished output What does “done” look like, and what must a trustworthy result contain?

10. Current friction Where are time, rework, delay, inconsistency, or avoidable mental effort being spent?

11. One safe first experiment What small piece could the team test with representative, non-sensitive examples?

12. Success measure and stop condition What will improve, and what result would tell you to pause, correct, or abandon the test?

13. Owner, date, and next useful step Finish with a commitment small enough to complete—not a transformation slogan.

Check whether this is a good first workflow

Quirky worked example

The Friday Afternoon Expense Archaeology Expedition

Every Friday, a manager reconstructs the story behind receipts, chat messages, calendar entries, mystery abbreviations, and one attachment named scan0004_final_FINAL.pdf. “Automate expenses” is too broad. This map makes the learning problem specific.

Trigger and inputs

Trigger: Friday at 2 p.m. Inputs include receipt images, card transactions, calendar events, travel policy, project codes, and explanations.

Judgment

Human work: decide whether missing context is harmless, an exception is reasonable, and which project benefited.

Review gate

Before submission: the employee confirms purpose and project; the manager checks policy exceptions and unsupported assumptions.

Safe experiment

Test: draft a missing-information checklist from ten sanitized examples. Measure review time; stop if details are invented.

A useful boundary

Do not confuse a map with permission.

Mapping a workflow does not authorize a team to put its data into an AI system. Live experiments involving sensitive information still require the organization’s approved tools, policies, access controls, and appropriate professional review.

After the map

Use the worksheet to make one decision.

Ready to practice

Collect three ordinary examples, one ambiguous example, and one edge case.

Needs framing

Clarify the owner, output, source material, or review point before evaluating a tool.

Too broad or risky

Choose one smaller handoff, summary, classification, or quality check.

Turn the map into practical training

Bring the awkward workflow—not a polished AI use case.

TrainedToThink workshops help teams examine the real work, practice on representative examples, name the human-review points, and leave with one useful next experiment.

Build your workshop blueprint Discuss a team workshop