Human-Centered AI Adoption Roadmap: the AdoptionLab.AI 5E Model
The AdoptionLab.AI 5E Model. Select a step on the ladder to see its key themes.
Start at the bottom
Five phases for building lasting AI capability by putting people first. Select a step on the ladder, or pick one here.
- 5 re-Envision Redesign roles and workflows to focus people on higher-value work as AI handles the routine.
- 4 Embed Make proven practices part of the job. Update SOPs, onboarding, and performance expectations.
- 3 Extend Scale high-potential pilots, test metrics at scale, develop initial training and coaching plans.
- 2 Experiment Run small, low-risk pilots. Document what's already in flight. Learn fast. Share wins and lessons learned.
- 1 Explore Map your current state, set a clear vision, and create basic guardrails.
Phase 1 of 5
Explore
Map your current state, set a clear vision, and create basic guardrails.
Take stock before you act. Learn where AI already shows up, what your team hopes and fears, and which rules keep early use safe. Even if you decide not to adopt AI, this phase gives you the facts to make that call well.
Key themes
- Be the sense-maker Turn the AI headlines into a clear story about your team.
- Make a pledge you can keep Replace reassurance with a specific, written commitment.
- Run a baseline assessment Find out who uses AI today, how, and how they feel about it.
- Offer amnesty for hidden use Treat quiet AI use as early pilots, not policy violations.
- Write a two-part AI policy Pair clear red lines with a vision for good use.
- Sort your data with a stoplight Red, yellow, and green rules make safe pilots easy.
- Engage the middle layer Give middle managers a distinct role, early.
- Map what helps and what hinders A force field analysis shows what to remove, not just what to push.
- Lead with invitation, not obligation Keep AI optional through Phase 3 and let results draw people in.
- Give everyone basic training Build a shared language for what AI can and cannot do.
Phase 2 of 5
Experiment
Run small, low-risk pilots. Document what's already in flight. Learn fast. Share wins and lessons learned.
Learn by doing. Create a safe space for small bets inside your policy, and treat every pilot, successful or not, as a lesson worth sharing.
Key themes
- Lead from the front Go first on training and practice, and learn out loud.
- Appoint an AI Coordinator Name one person to track every pilot in flight.
- Document what's already in flight Catalog the pilots already happening, with curiosity, not blame.
- Prioritize data quality Garbage in, garbage out. Fix the inputs first.
- Set directional metrics Connect AI to measures you already track.
- Show the right numbers at the right time Report learning first. Returns come later, in order.
- Craft your AI story Every leader writes a short, honest "why."
- Celebrate wins and lessons The only failure is the one you don't learn from.
- Check in before moving on Re-run the survey and act on what you hear.
Phase 3 of 5
Extend
Scale high-potential pilots, test metrics at scale, develop initial training and coaching plans.
Move from a few promising pilots to the broader team. This is where most AI efforts quietly stall, so set go/no-go rules first, scale slowly, and decide in the open.
Key themes
- Review your pilot portfolio Close the pilots that won't scale, and keep the lessons.
- Define success before you scale Agree on what success looks like, who decides, and when to pause.
- Aim at bottlenecks Start from where work piles up, not from a list of AI use cases.
- Protect the core of the job Automate the admin. Protect the judgment and relationships.
- Keep it opt-in Let people join scaled pilots by choice, and track who is watching.
- Decide where the gains go Commit freed time and money before it arrives.
- Test your measures at scale Metrics that worked for two people may break for twenty.
- Pace growth with go/no-go checks Scale one or two pilots at a time, in short cycles.
- Run a mid-mortem when pilots stall Pause, look for patterns, and re-decide in the open.
- Build an AI community Book clubs, promptathons, and regular updates keep learning social.
Phase 4 of 5
Embed
Make proven practices part of the job. Update SOPs, onboarding, and performance expectations.
Turn a proven, scaled pilot into standard practice for the role. A process is embedded when it holds up even when no one is watching closely, and when it survives the original champion leaving.
Key themes
- Confirm it is ready A process that works is not the same as one you can stand behind.
- Formalize the new process Update SOPs, playbooks, and checklists with the AI-enabled steps.
- Build the quality loop Once use is routine, attention fades. Build checks that don't.
- Move from optional to required Require the standard, not the keystrokes.
- Update training and onboarding Make sure everyone in the role can use the new process with confidence.
- Partner with HR Update job descriptions and postings to reflect AI in the role.
- Celebrate the milestone Embedding is a major achievement. Make it visible.
- Sustain and monitor Embedding is the start of continuous improvement, not the end.
Phase 5 of 5
re-Envision
Redesign roles and workflows to focus people on higher-value work as AI handles the routine.
Shape the next version of your team. Use the capacity you have freed to deliver value that was out of reach before, and redesign roles with the people in them.
Key themes
- Treat freed capacity as an asset Why this phase is not a layoff plan.
- Reassess how work gets done Look past single processes to whole workflows.
- Redesign roles with the people in them Do it with people, never to them.
- Redesign the middle layer The promise from Phase 1 comes due here.
- Reimagine your value Shift from internal efficiency to outcomes for the people you serve.
- Lead with transparency Explain the reasons behind every role change.
- Share what you learned Help the rest of the organization catch up.
- Loop back for the next wave A loop, not a ladder.
