1 / 23

Longhand AI competency course

The Decision-Making ProcessTurn an AI idea into a decision someone can challenge.

Make the consequential choice explicit, show the evidence and alternatives, define how it would work, and decide what happens when assumptions fail.

Version 1.075 minutesPractice evidence required

The finish line

A defensible DMP decision record

Your record contains the Big What, Why, What, How, What If, alternatives, unknowns, and the decision rule. It is practice evidence, not a certified decision.

Progress is saved in this browser.

Longhand AI competency course

The Decision-Making ProcessTurn an AI idea into a decision someone can challenge.

Make the consequential choice explicit, show the evidence and alternatives, define how it would work, and decide what happens when assumptions fail.

Version 1.075 minutesPractice evidence required

The finish line

A defensible DMP decision record

Your record contains the Big What, Why, What, How, What If, alternatives, unknowns, and the decision rule. It is practice evidence, not a certified decision.

Progress is saved in this browser.

Measurable learning objectives

By the end, you can produce evidence—not just repeat the framework.

Frame a consequential Big What decision instead of a vague AI aspiration.

Use Why, What, How, and What If to expose evidence, scope, execution, risk, and alternatives.

Compare build, defer, reject, and non-AI options without hiding uncertainty.

Produce an independent decision record another reviewer can challenge or reproduce.

You will see a worked example, diagnose a weak example, complete guided practice, and leave with an independent practice artifact. The artifact prepares you for a later human-scored performance assessment.

Course requirements

Know what you need—and what counts as evidence.

Prerequisites

A candidate recurring workflow or the incident-report practice case.
A draft TDAR Results Brief or equivalent result and baseline record.
Permission to analyze the workflow; this lesson does not authorize implementation.

Assessment connection

Milestone 1 — Opportunity Diagnosis · planned

A complete DMP decision record
At least one credible alternative, including the current/manual process where applicable
A build, defer, reject, or test decision with evidence and stop conditions

Lesson practice is formative. This course does not issue proficiency or certification; the listed performance assessment requires separate evidence and qualified human evaluation.

Authoritative sources and course version

Version 1.0 · estimated 75 minutes · sources reviewed 2026-08-10

Higher Education Rubric, Seventh EditionQuality Matters · Constructive alignment among objectives, assessment, materials, activities, and technology.OSCQR SUNY Online Course Quality Review RubricSUNY Online · Online-course review, authentic assessment, interaction, feedback, support, and accessibility.Artificial Intelligence Risk Management Framework (AI RMF 1.0)National Institute of Standards and Technology · Govern, Map, Measure, and Manage functions and trustworthy AI characteristics.

The guide

Explain it. See it. Challenge it. Apply it. Save the evidence.

Step 1 of 6 · 10 min

Make the choice and its logic visible before implementation.

Use the five-part decision record

A roadmap item can hide five different disagreements: the actual decision, purpose, scope, execution, and risk. The Longhand DMP turns those disagreements into reviewable fields rather than compressing them into a yes/no vote.

Do this

1Big What: write the consequential decision that must be made now—build, test, defer, reject, or choose another path.
2Why: state the result, evidence, urgency, and strategic reason that make the decision worth attention.
3What: define the workflow, users, inputs, outputs, boundaries, and finish line precisely enough to evaluate.
4How: describe the smallest credible operating model, including people, AI surface, data, deterministic checks, effort, and permissions.
5What If: expose uncertainty, failure modes, alternatives, stop conditions, and what evidence would change the decision.

Remember: DMP is a Longhand-developed decision record. It is not presented as an externally validated academic theory; its value must come from transparent evidence and better decisions.

What you leave with

Five-part DMP draft

Evidence-versus-assumption list

DMP review questions

Part	Question	Common failure
Big What	What decision must be made?	A topic or aspiration replaces the choice.
Why	What result and evidence justify attention?	Urgency or enthusiasm replaces evidence.
What	What exactly is in and out?	The proposal expands whenever challenged.
How	What is the smallest credible operating model?	The design ignores permissions, people, or review work.
What If	What could fail or change the choice?	Risk is reduced to a generic disclaimer.

Step 1 of 6 · Run it

Apply it before you move on

Do this

1Big What: write the consequential decision that must be made now—build, test, defer, reject, or choose another path.
2Why: state the result, evidence, urgency, and strategic reason that make the decision worth attention.
3What: define the workflow, users, inputs, outputs, boundaries, and finish line precisely enough to evaluate.
4How: describe the smallest credible operating model, including people, AI surface, data, deterministic checks, effort, and permissions.
5What If: expose uncertainty, failure modes, alternatives, stop conditions, and what evidence would change the decision.

Remember: DMP is a Longhand-developed decision record. It is not presented as an externally validated academic theory; its value must come from transparent evidence and better decisions.

Open the decision record

Help me open a DMP decision record for this AI opportunity. Ask one question at a time in this order: Big What, Why, What, How, What If. Separate observed evidence from assumptions. Always include the current/manual process as a possible alternative. Do not recommend a build until every section has enough evidence to review.

Opportunity: [describe]
TDAR result: [paste]

What you leave with

Five-part DMP draft

Evidence-versus-assumption list

Practice evidence target

You can distinguish the decision itself from the evidence, proposed scope, operating model, and uncertainty around it.

Step 1 checkpoint

Use the five-part decision record: practice checkpoint

Practice evidence target

You can distinguish the decision itself from the evidence, proposed scope, operating model, and uncertainty around it.

Recording this checkpoint saves self-reported lesson practice in this browser. It is not evidence of certified competency; certification requires a separately submitted artifact and qualified human evaluation.

Step 2 of 6 · 12 min

Turn a broad aspiration into a choice supported by a result and evidence.

Make the Big What and Why decision-grade

Teams often agree that AI is important while disagreeing about the actual decision. A decision-grade Big What names the reversible next commitment, and the Why connects it to evidence rather than pressure or novelty.

Do this

1Write the decision as one sentence with a verb, object, boundary, and timing: for example, 'Run a four-report pilot of source-grounded incident drafting; do not automate approval or publishing.'
2Connect the Why to the TDAR result and baseline. Identify which evidence is observed, estimated, or missing.
3State why the decision is needed now and what happens if no change is made.
4Define the smallest reversible commitment that can answer the important uncertainty before a larger build.

Remember: A reversible test is not automatically the right answer. It is useful when it resolves a decision-relevant uncertainty at acceptable risk and cost.

What you leave with

Decision sentence

Evidence table

Smallest reversible commitment

Decision strength

Weak	Stronger
Adopt AI for incident management	Pilot AI-assisted drafting for four weekly reports; keep significance and release decisions human.
We need to move faster	Median preparation is 150 minutes across eight reports; test whether drafting time can fall without increasing material corrections.
Build the full system	Test the source-grounded draft and correction rate before connecting any system of record.

Step 2 of 6 · Run it

Apply it before you move on

Do this

1Write the decision as one sentence with a verb, object, boundary, and timing: for example, 'Run a four-report pilot of source-grounded incident drafting; do not automate approval or publishing.'
2Connect the Why to the TDAR result and baseline. Identify which evidence is observed, estimated, or missing.
3State why the decision is needed now and what happens if no change is made.
4Define the smallest reversible commitment that can answer the important uncertainty before a larger build.

Remember: A reversible test is not automatically the right answer. It is useful when it resolves a decision-relevant uncertainty at acceptable risk and cost.

Stress-test Big What and Why

Review this Big What and Why. Identify vague verbs, hidden scope, unsupported urgency, missing baseline evidence, and irreversible commitments. Rewrite it as the smallest reversible decision that can test the most important uncertainty.

Big What: [paste]
Why: [paste]
TDAR: [paste]

What you leave with

Decision sentence

Evidence table

Smallest reversible commitment

Practice evidence target

A reviewer can identify the exact decision, why it matters, what evidence supports it, and which unknown the next step is meant to resolve.

Step 2 checkpoint

Make the Big What and Why decision-grade: practice checkpoint

Practice evidence target

A reviewer can identify the exact decision, why it matters, what evidence supports it, and which unknown the next step is meant to resolve.

Step 3 of 6 · 14 min

Scope the work and include the people, checks, and permissions it actually requires.

Define the bounded What and credible How

AI proposals often understate the effort outside the model: source preparation, access review, deterministic validation, human review, exception handling, measurement, and maintenance. The What and How make the full operating system visible.

Do this

1Define the user, trigger, authorized inputs, produced artifact, downstream consumer, and explicit non-goals.
2Choose the simplest AI surface that can access the required information and perform only the authorized actions.
3Add deterministic checks for schemas, totals, required fields, citations, or invariants where exactness is testable.
4Place human judgment and approvals at the relevant steps and include their real time in the effort estimate.
5Describe failure handling, evidence capture, and how the workflow returns to a known state.

Remember: An agent can reduce execution effort while increasing review, exception, or governance effort. Count the system, not only the generation step.

What you leave with

Bounded scope

Operating model

Full-effort estimate inputs

Operating model layers

Layer	Review question
Work	What recurring job and finish line are in scope?
Information	Which authorized sources are required and how current are they?
AI	Which bounded interpretation or generation work is appropriate?
Deterministic	Which exact rules, schemas, totals, or invariants must run?
Human	Who decides, approves, handles exceptions, and owns consequences?
Operations	How are logs, failures, changes, and measurement handled?

Step 3 of 6 · Run it

Apply it before you move on

Do this

1Define the user, trigger, authorized inputs, produced artifact, downstream consumer, and explicit non-goals.
2Choose the simplest AI surface that can access the required information and perform only the authorized actions.
3Add deterministic checks for schemas, totals, required fields, citations, or invariants where exactness is testable.
4Place human judgment and approvals at the relevant steps and include their real time in the effort estimate.
5Describe failure handling, evidence capture, and how the workflow returns to a known state.

Remember: An agent can reduce execution effort while increasing review, exception, or governance effort. Count the system, not only the generation step.

Build the operating model

Turn this proposed scope into a bounded operating model. Use sections for user and trigger, authorized inputs, AI work, deterministic validation, human judgment, external actions and approvals, failure handling, evidence retained, non-goals, and full effort. Flag any permission or ownership assumption.

Proposed What/How: [paste]

What you leave with

Bounded scope

Operating model

Full-effort estimate inputs

Practice evidence target

The operating model accounts for the full workflow, including data, people, permissions, validation, exceptions, evidence, and maintenance—not only the prompt or agent.

Step 3 checkpoint

Define the bounded What and credible How: practice checkpoint

Practice evidence target

The operating model accounts for the full workflow, including data, people, permissions, validation, exceptions, evidence, and maintenance—not only the prompt or agent.

Step 4 of 6 · 12 min

Convert risks and alternatives into tests, stop rules, and decision changes.

Use What If to make uncertainty actionable

A list of generic risks does not improve a decision. What If connects a failure or uncertainty to detection, response, accountable review, and the evidence that would change the choice.

Do this

1List the assumptions most capable of changing the decision: demand, data quality, permissions, reliability, review capacity, harm, cost, or adoption.
2For each, define an observable signal and the earliest safe point at which it can be detected.
3Specify continue, revise, pause, and stop responses with an accountable person for the decision.
4Compare at least one non-AI or lower-complexity alternative and the option to keep the current process.

Remember: A stop rule is useful only when the signal is observable and someone has authority to stop the work.

What you leave with

Decision-changing uncertainty table

Alternatives

Stop and escalation rules

What If example

Uncertainty	Signal	Decision rule
Citations do not reliably support the draft	More than one unsupported material claim in a report	Pause the pilot; fix source retrieval and rerun the test set.
Human review erases the time saved	Median review exceeds 45 minutes	Revise scope or compare with a template-only alternative.
Correction risk increases	Material correction rate exceeds baseline	Stop external use; investigate before another live case.

Step 4 of 6 · Run it

Apply it before you move on

Do this

1List the assumptions most capable of changing the decision: demand, data quality, permissions, reliability, review capacity, harm, cost, or adoption.
2For each, define an observable signal and the earliest safe point at which it can be detected.
3Specify continue, revise, pause, and stop responses with an accountable person for the decision.
4Compare at least one non-AI or lower-complexity alternative and the option to keep the current process.

Remember: A stop rule is useful only when the signal is observable and someone has authority to stop the work.

Turn risks into decision rules

Convert these What If concerns into a table with assumption, evidence, early signal, consequence, continue/revise/pause/stop rule, accountable human, and alternative. Add the current/manual process as a comparator. Avoid generic risks without detection or response.

Concerns: [paste]

What you leave with

Decision-changing uncertainty table

Alternatives

Stop and escalation rules

Practice evidence target

Your What If section contains decision-changing uncertainties, observable signals, response rules, owners, and credible alternatives.

Step 4 checkpoint

Use What If to make uncertainty actionable: practice checkpoint

Practice evidence target

Your What If section contains decision-changing uncertainties, observable signals, response rules, owners, and credible alternatives.

Step 5 of 6 · 12 min

Diagnose a polished recommendation that skipped evidence and alternatives.

Compare a decision record with a disguised preference

A decision memo can look complete while starting from a preferred solution, backfilling reasons, and ignoring the current process. Contrast makes those reasoning failures easier to recognize in your own work.

Do this

1Identify where the weak record assumes the answer, hides unknowns, or treats implementation detail as evidence of value.
2Trace the strong record from result and baseline through a reversible test and explicit decision rules.
3Check whether the current process and a simpler non-agent option received a fair comparison.
4Name the one piece of evidence most likely to reverse the recommendation.

Remember: The strong example makes a pilot defensible; it does not guarantee that the pilot will succeed or deserve a larger build.

What you leave with

Weak-record diagnosis

Alternative comparison

Decision-reversing evidence

Worked decision summary

DMP part	Reviewable version
Big What	Run a four-report, source-grounded drafting pilot; do not connect write actions or automate approval.
Why	Preparation baseline is 150 median minutes; test time reduction without increasing corrections.
What	Draft chronology and standard sections from authorized records with citations and missing-field flags.
How	Read-only workspace, deterministic required-field check, incident-lead review, versioned output and timing log.
What If	Pause on unsupported claims or worse correction rate; compare with a deterministic template if review time remains high.

Step 5 of 6 · Run it

Apply it before you move on

Do this

1Identify where the weak record assumes the answer, hides unknowns, or treats implementation detail as evidence of value.
2Trace the strong record from result and baseline through a reversible test and explicit decision rules.
3Check whether the current process and a simpler non-agent option received a fair comparison.
4Name the one piece of evidence most likely to reverse the recommendation.

Remember: The strong example makes a pilot defensible; it does not guarantee that the pilot will succeed or deserve a larger build.

Find the disguised preference

Weak record: 'We should build an autonomous incident agent because agentic AI is strategic. It will read our systems, write the report, and send it for approval. Risks will be managed with human in the loop.'

Diagnose the Big What, Why, What, How, and What If gaps. Compare it with a source-grounded draft pilot, a deterministic template, and the current process. State what evidence is needed before choosing.

What you leave with

Weak-record diagnosis

Alternative comparison

Decision-reversing evidence

Practice evidence target

You can identify circular justification, missing alternatives, hidden scope, and non-actionable risk language in a proposed decision.

Step 5 checkpoint

Compare a decision record with a disguised preference: practice checkpoint

Practice evidence target

You can identify circular justification, missing alternatives, hidden scope, and non-actionable risk language in a proposed decision.

Step 6 of 6 · 15 min

Make a build, test, defer, reject, or alternative decision on a new case.

Produce and defend your independent decision record

The final practice requires transfer: a bounded decision that another reviewer could challenge using the same evidence and decision rules.

Do this

1Choose a candidate workflow and define the immediate decision. Use a synthetic or de-identified case when authorization is limited.
2Complete Big What, Why, What, How, and What If with facts, assumptions, alternatives, and source references separated.
3Select build, bounded test, defer, reject, or non-AI alternative. Explain why the rejected options are weaker under the current evidence.
4Add the evidence that would reverse the decision and the date or event that triggers reconsideration.
5Save the DMP record beside the TDAR brief for the planned Opportunity Diagnosis assessment.

Remember: A decision record should make a future revision easier, not make the current decision look inevitable.

What you leave with

Independent DMP record

Alternative comparison

Decision-reversal evidence

Step 6 of 6 · Run it

Apply it before you move on

Do this

1Choose a candidate workflow and define the immediate decision. Use a synthetic or de-identified case when authorization is limited.
2Complete Big What, Why, What, How, and What If with facts, assumptions, alternatives, and source references separated.
3Select build, bounded test, defer, reject, or non-AI alternative. Explain why the rejected options are weaker under the current evidence.
4Add the evidence that would reverse the decision and the date or event that triggers reconsideration.
5Save the DMP record beside the TDAR brief for the planned Opportunity Diagnosis assessment.

Remember: A decision record should make a future revision easier, not make the current decision look inevitable.

Independent DMP template

# DMP Decision Record

## Big What
[Exact decision, boundary, and timing]

## Why
[TDAR result, baseline evidence, urgency, observed facts, assumptions]

## What
[User, trigger, workflow, inputs, outputs, finish line, non-goals]

## How
[People, AI surface, data, deterministic checks, permissions, review, effort, failure handling]

## What If
[Decision-changing uncertainties, signals, continue/revise/pause/stop rules, owners]

## Alternatives compared
[Current process, non-AI option, lower-complexity AI option, proposed option]

## Decision and rationale
[Build, bounded test, defer, reject, or alternative]

## Evidence that would reverse this decision
[Specific evidence]

## Reconsideration trigger
[Date or event]

What you leave with

Independent DMP record

Alternative comparison

Decision-reversal evidence

Practice evidence target

You saved a complete DMP record with a reversible decision, evidence, alternatives, uncertainty, stop rules, and reconsideration trigger.

Step 6 checkpoint

Produce and defend your independent decision record: practice checkpoint

Practice evidence target

You saved a complete DMP record with a reversible decision, evidence, alternatives, uncertainty, stop rules, and reconsideration trigger.

Formative practice complete

Your AI proposal is now a reviewable decision.

Keep the DMP record with the TDAR brief. The planned Opportunity Diagnosis assessment will require both, plus RICE comparison and human-accountability evidence.

Next course

RICE for AI Priorities

Compare multiple candidates using a consistent equation, visible evidence, and sensitivity analysis.

Open training

Review your gaps

Return to the AI Gear Audit

Update your formative self-rating and open the next direct lesson. Green remains self-reported, not certified.

Open training