The NAVI AI Enablement Journey · Guide 09

Adoption, Training and Value Measurement

A practical guide for helping people build useful habits, become increasingly self-sufficient and produce evidence that supports responsible improvement and expansion.

Starting smallTraining through workPractical supportValue evidenceExpansion decisions
Core idea

Meaningful adoption occurs when people repeatedly use trusted AI-assisted processes to improve real work, know how to review results and can obtain help or report concerns without depending on one expert. Each successful workflow should also leave people, examples, operating practices and decision evidence that make the next implementation easier.

Where this guide fits: Guide 08 establishes a reliable operating workflow. Guide 09 helps people adopt it, become increasingly self-sufficient, and produce evidence that supports responsible improvement, expansion, revision, or retirement.
What meaningful adoption looks like

Look for changed work, appropriate trust and recurring use. Platform activity shows where to investigate, but does not by itself prove usefulness or value.

1Useful work is completedA task, decision, deliverable or follow-through that matters to the business is finished.
2Use becomes recurringPeople return for real work, not only during training or demonstrations.
3Results are trusted appropriatelyUsers understand sources, limitations, review and pause conditions.
4Users become self-sufficientRoutine work no longer depends on one champion or advisor.
5Evidence changes the programFeedback, exceptions and outcomes improve training, workflows or priorities.
Practical test: can users explain which work they now complete differently, how they know the result is fit for use and what they do when it is uncertain?
Start small and train through real work

Give one person or a small group a visible purpose, one or two useful workflows and enough guided practice to complete actual work.

Choose the first users and work

Choose people because they participate in the selected work, not simply because they are enthusiastic about AI.

Real participantsInclude the owner, routine users and necessary reviewers or source owners.
Available during the testThey can run the work, review outputs and provide specific feedback.
Representative workUse recurring tasks and realistic conditions, not only the easiest example.
Observable resultThe team can compare before and after work and make a decision within a useful period.
1OrientExplain the work surfaces, approved sources, safe-use boundaries and support path.Outcome: the user knows where work belongs.
2Run one real task togetherFrame the job, select evidence, review the result and save the useful output.Outcome: a real deliverable or decision aid.
3Repeat with variationTry another normal case, an incomplete input and a likely exception.Outcome: practical confidence.
4Operate with light supportThe user runs the approved workflow and uses the stated help path when needed.Outcome: recurring use under real conditions.
5Teach back and improveThe user explains the method and contributes a useful example or correction.Outcome: capability spreads beyond one person.
Prepare people for
changing work

Adoption improves when people understand what is changing, why it is changing, what remains their responsibility and how the organization intends to use the capacity or capability created.

01 · Explain

Name the purpose and the work change

Describe the business reason, the tasks likely to be reduced or accelerated and the work that should improve as a result.

02 · Involve

Design with people closest to the work

Use their examples, exceptions and judgment to shape the workflow. Participation should influence the design, not merely validate a finished solution.

03 · Clarify

State retained and new responsibilities

Make review, decision, escalation, source ownership, maintenance and improvement responsibilities visible before launch.

04 · Support

Provide time to learn and adapt

Give people realistic practice, a safe way to raise concerns and enough support to build competence without pretending uncertainty is resistance.

Reducing effort is not a complete people plan. Leaders should state how saved capacity is intended to support customers, quality, deeper thinking, innovation, growth or another priority—and revisit that intention with evidence from real work.

Build practical AI fluency through real work

AI fluency is the ability to direct, evaluate and improve AI-enabled work responsibly. It grows through repeated practice in the organization’s own workflows, not through tool awareness alone.

Frame the workDefine the purpose, audience, constraints, completion rule and business consequence.
Select contextChoose approved knowledge, data, examples and tools that are relevant and sufficient.
Direct and delegateBreak the work into useful steps and specify what AI, automation and people should each do.
Verify and judgeCheck sources, logic, completeness, quality and fitness for the intended use.
Recognize exceptionsKnow when to pause, use a fallback, request help or escalate to an authorized decision maker.
Improve the methodPreserve corrections, examples, limitations and evidence so the workflow and guidance become better.
Build confidence and support

Make the necessary responsibilities clear without creating unnecessary structure. One person may handle several or all of them.

Direction and ownershipPurpose and accountability

Connect the work to a business priority and own its purpose, scope, participation, measures and continued use.

Routine useComplete the work

Supply the inputs, follow the approved method, check required items, save the result and recognize exceptions.

Review and approvalApply judgment

Check the required evidence or business judgment before the result is relied upon, published or used to take action.

Peer and technical supportHelp when useful

Share examples, answer routine questions and route unresolved source, permission, product or technical issues appropriately.

Client-operated

Routine work should become self-sufficient

Approved and understood workflowPurpose, trigger, sources, steps, output and review are known.
Case is within the tested boundaryInputs and expected result resemble practiced examples.
Required judgment is availableThe user or reviewer can decide whether the result is fit for use.
Issue is ordinaryExisting guidance, examples or support routes can resolve it.
Additional support

Ask for help when it is useful

You are stuck or the issue keeps returningThe cause is unclear, normal guidance is not resolving it or the work is consuming more effort than it should.
You want to move fasterImplementation support may help with design, setup, evaluation or expansion even when the team could eventually work through it independently.
Specialist authority is neededSecurity, privacy, legal, compliance or another qualified decision exceeds the user’s authority.
Self-sufficiency remains the goalNew workflows, sources and connections do not automatically require GrowthIQ involvement when the organization can handle them responsibly.
If people do not know when to use the workflow, find it slow or unreliable, or are unsure whether it is safe, reporting that experience is part of normal workflow maintenance and responsible adoption—not evidence that the person is resistant to AI.
“I do not know when to use it.”

The job or trigger is unclear

The person understands the platform but not the moment the workflow should begin.

Response: define the trigger, example input and completed output.
“It is slow or unreliable.”

The workflow may not fit the work

Setup, repair or repeated exceptions may outweigh the benefit.

Response: observe the work, remove steps, improve the method or stop.
“I am not sure this is safe or trustworthy.”

Evidence or authority may be weak

The concern may require clearer sources, review, access or policy—not encouragement.

Response: show the boundary or pause for an owner decision.

A simple feedback path

Use recognizable routes.

1
Ask a usage questionUse available guidance, an example, a knowledgeable colleague or focused coaching.
2
Suggest an improvementReport repeated confusion, missing examples or unnecessary workarounds.
3
Report a questionable resultPause reliance and preserve the task, inputs, output and expected result.
4
Escalate possible harmStop the affected use and follow the approved governance path.

What useful feedback includes

Specific evidence helps correct the right part of the system.

Task and workflowWhat the user was trying to complete and which process was used.
Inputs and sourcesFiles, records, dates and approved knowledge involved.
Observed and expected resultWhat happened, what should have happened and the supporting evidence.
Effect and action takenDelay, rework, external effect, pause, correction, send or system change.
Why measurement matters

Measurement shows whether the workflow is improving real work and provides evidence for what to continue, change or stop. Begin with the business result that matters, then add other measures only when useful.

Start here

Business value

Time, quality, capacity, customer, revenue, cost or risk outcomes show whether the workflow produces a result the organization cares about.

Use for: deciding whether the work is worth continuing.
Add when useful

Meaningful adoption

Repeat use of selected workflows, completed real work, reviewer acceptance and reduced workarounds help explain whether the operating method is taking hold.

Use for: improving training and preparing for broader use.
Optimize later

Platform activity

Active users, Chat activity, saved outputs and shared prompt use show where the platform is being used, but not whether the work creates value.

Use for: locating patterns and possible coaching needs.
1Business resultThe intended outcome: time, capacity, customer, revenue, cost, risk or another business effect.Include the baseline.
2Quality and reworkAcceptance, material errors, completeness, correction rounds or reviewer assessment.Define who judges quality.
3Recurring useWhether eligible users return when the actual workflow trigger occurs.Interpret volume in context.
4Exceptions and supportPauses, unsupported cases, correction burden, permission issues and manual fallback.Preserve negative evidence.
5Evidence qualityObserved examples, samples, system records, reviewer notes, estimates and known limitations.Do not overstate causation.
Begin with one intended business result and one quality, exception or failure measure. Add adoption and activity measures when they help explain results, improve the workflow or prepare for broader use. A simple note or table is enough.
BaselineHow the work performs before the change.
Learning question or targetWhat improvement or uncertainty the test addresses.
Observed resultWhat happened during representative real work.
Evidence and methodSource and whether observed, sampled or estimated.
Limitations and failuresUnsupported cases and what the evidence cannot prove.
Next decisionScale, revise, prepare or stop—with owner and checkpoint.
Use evidence to expand deliberately

Expand when people can repeat the work, support normal use and explain the value. Adding users without shared methods and ownership may increase activity without increasing value.

Evaluate each new workflow before expanding it

Define the intended business result, establish a reasonable baseline or completed example, test representative work and record quality, effort, exceptions and limitations. Then decide whether to scale, revise, prepare or stop.

1Guided individual useOne person completes useful work with direct coaching and records corrections.
2Supported team repetitionSeveral users practice the same workflow with shared examples, review and focused help.
3Independent team standardRoutine use is owned and supported internally; coaching shifts toward exceptions and improvement.
4Adjacent expansionA related group adapts the proven pattern after confirming its own work, sources, permissions and value.
ScaleUseful, sufficiently reliable, owned, repeatable and supported.
ReviseThe opportunity matters, but the method, scope or user experience must improve.
PrepareValue is visible, but knowledge, access, governance, integration, training or support is not ready.
StopValue is insufficient, burden is too high or another solution fits better.
Make future adoption easier

Concise operational guidance makes onboarding and expansion easier when people join, responsibilities change or a proven workflow moves to another team. Keep only what people will maintain and use.

Minimum setup

Use a focused system rather than a training academy or complex analytics stack.

Visible sponsor and small initial groupReal participants in one or two useful workflows.
Responsibility-based onboardingPractice normal work, exceptions, review and support routes without requiring a complex team structure.
Available help and support routeMake it clear where users can find guidance or ask for help when useful.
A small measure set and decision rhythmEnough evidence to support scale, revise, prepare or stop.

Maintained enablement assets

Preserve only what future users need.

Approved worked exampleRepresentative input, reviewed output and corrections.
Reusable instructions or workflow guideThe tested method, responsibilities, review and completion rule.
Known limits and exceptionsCases that pause, escalate or remain outside scope.
Short responsibility guidanceWhat someone must know and do when performing, reviewing or supporting the work.
Support and feedback routeWhere questions, improvements and concerns go.
Evidence and next decisionBaseline, results, limitations, owner and checkpoint.
Close the loop: Adoption and value evidence should change the implementation program. Feed the results back into the Guide 02 roadmap, Guide 05 governance decisions, Guide 06 connections, Guide 08 workflow design, and the broader foundation model in Guide 01. Use Guide 09A for day-to-day efficient AI use and token management.
This guide is a practical implementation framework, not a workforce policy, performance-management system, audited ROI methodology, legal opinion or substitute for qualified professional judgment. Label estimates, directional evidence and limitations accurately.