The AI practice playbook

Eight practice problems worth solving with AI.

A field guide for turning recurring practice friction into bounded, reviewable AI workflows.

Field notes 01 to 08Fictional inputsHuman review required

Read before use

How to use this playbook

Start with the problem that costs your team the most attention. Read what the workflow does, inspect the guardrail, then use the starter in your preferred AI tool with fictional information. Keep the human review point visible as you adapt it.

The Safe AI Practice

What it does
Map a recurring practice problem into a bounded workflow with clear inputs, outputs, and ownership.
The guardrail
Choose tools and a deployment model only after review by your organization, compliance team, or qualified counsel.
Start here
Using fictional information only, map one recurring practice task into its inputs, steps, output, and named human reviewer.

The Operative Note That Refuses to Invent

What it does
Turn fictional dictation into a structured operative-note draft that identifies what was never said.
The guardrail
Flag unsupported facts instead of filling gaps, and preserve a qualified human review point.
Start here
Using fictional dictation only, draft an operative note from this template and mark every missing or unsupported fact as not provided.

The Letter Queue That Eats Your Evening

What it does
Turn messy source material into a concise, review-ready correspondence draft.
The guardrail
Use only supplied source material, identify conflicts, and keep a person responsible for the final message.
Start here
Using fictional source material only, draft a concise letter and list any missing or conflicting details before the draft.

The Prior-Authorization Packet

What it does
Organize a fictional request into a clear packet with requirements, supporting material, and unresolved gaps.
The guardrail
Do not infer criteria, history, or supporting facts that are absent from the supplied material.
Start here
Using fictional information only, organize this request into requirements, available support, missing items, and a review-ready cover draft.

The Denial Appeal and Peer-to-Peer Brief

What it does
Turn a fictional denial and its source record into a focused appeal outline and conversation brief.
The guardrail
Separate quoted policy language, supplied facts, and open questions so a reviewer can verify each one.
Start here
Using fictional information only, create an appeal outline and peer-to-peer brief that cites the supplied denial language and flags every unsupported point.

Know the Chart Before You Enter the Room

What it does
Condense a fictional chart into a focused pre-visit brief with unanswered questions made visible.
The guardrail
Treat the output as a draft for review, not a substitute for the source chart or clinical judgment.
Start here
Using a fictional chart only, create a pre-visit brief with relevant history, open questions, conflicts, and source references.

Give Your Staff a Practice Brain

What it does
Make approved practice knowledge searchable so staff can find an answer or receive an explicit not-found response.
The guardrail
Answer only from approved source material and say when the answer is not present.
Start here
Using fictional practice policies only, answer the question from the supplied sources or respond that the answer was not found.

Build Your Practice's 90-Day AI Plan

What it does
Sequence a small set of practice workflows into a practical 90-day build and review plan.
The guardrail
Start with bounded, reversible work and assign a human owner and review point to every workflow.
Start here
Using fictional practice details only, rank these workflow problems and turn the top three into a 90-day plan with owners, guardrails, and review dates.

Choose your next step

Bring one practice problem with you.

See a workflow built live, or join the free community for practical foundations, resources, and future invitations.

See the next live buildJoin the free community