A Worked Engagement

How an AI-adoption engagement actually runs — start to finish, on real data. Four moves: understand the organisation, map every task to where AI belongs, build a personal learning plan for each person, then drive adoption with prompts and activities. Grounded in a synthetic ~9,000-person build of Canary Co and O*NET-coded occupations.

Method demo · 2026 Synthetic data O*NET 30.2 · 43 occupations
Read this first. Canary Co is a wholly fictional company and is not trademark-checked. Every employee shown is SYNTHETIC — names are generated, emails use the reserved .example domain, no real person is depicted. Occupation SOC codes are real O*NET 30.2 codes; the task statements and skill ratings in this demo are illustrative placeholders (carry an onet_status flag) and should be replaced with licensed O*NET data before any client-facing claim. This page demonstrates method, not findings about a real organisation.

The shape of the work

Most "AI adoption" stalls because it starts with the tool. This starts with the organisation — who does what, which tasks should stay human, and what each person needs to learn next. The tool comes last, once you know where it belongs.

Phase 01

Understand

Model the org from the ground up — divisions, units, roles, people, capability.

Phase 02

Map to AI

Decompose every role into tasks and route each to a three-tier AI model.

Phase 03

Plan learning

Target minus actual equals gap. Gaps become personal learning plans.

Phase 04

Drive adoption

Prompts, activities and a 90-day roadmap that meet people in their real work.

01

Discovery · Workforce intelligence

Understand the organisation

You cannot adopt AI into an org you can't see. The first move builds a complete, internally consistent picture: 6 operating divisions, 17 business units, 149 distinct job titles, 43 O*NET-coded occupations, and 8,980 synthetic people — each with a seniority band, tenure, employment type, region, and a role grounded in a real occupation code.

8,980
Synthetic staff
6
Divisions
17
Business units
149
Job titles
43
O*NET occupations
11
Regions

Headcount by division

Ports & LogisticsNorth Europe · Med · Gulf · APAC
1,980
Health & BeautyRetail · Pharmacy · E-commerce
1,900
InfrastructureNetworks · Asset mgmt · Field
1,740
ConnectTelecoms · Software · Customer ops
1,620
EnergyGeneration · Trading · HSE
1,140
Group (Corporate)Finance · HR · IT · Legal · Exec
600

Seniority pyramid

Lead
733
Senior
1,829
Mid
3,731
Junior
2,687

Employment mix

Full-time
7,340
Part-time
1,183
Contract
457

Capability is scaled by seniority — a Senior generally exceeds a Junior against the same role competency target. That single rule is what makes the gap analysis in Phase 03 meaningful rather than random.

How the picture is built

The synthetic half. A generator creates people, IDs, the reporting hierarchy and personas — seniority, tenure, employment type, region. Same random seed in, same org out, so a demo is fully reproducible.

employees.csvdivisions.csvbusiness_units.csv

The O*NET half. Each role carries a real SOC code (e.g. Pharmacist 29-1051.00, Customs Broker 13-1041.06). O*NET supplies the tasks, skills and competency profile that drive everything downstream.

roles.csvtasks.csvcompetency_capability.csv
Open the interactive Org Explorer → Search 8,980 staff · filter by division, role and AI tier · live in the browser
02

Task analysis · The three-tier model

Map skills & tasks to AI

A job is not one thing — it's a bundle of tasks, and they don't all belong in the same place. We decompose each occupation into its O*NET tasks and route every one to a three-tier model: stays human, human + AI, or AI-led with oversight. Across 43 roles that produced 176 mapped tasks.

Tier 1

Employee Only

None / human

Empathy, accountability, judgement, physical presence. The work that defines the role and carries the risk.

Test: would you want a person accountable if this went wrong?

Tier 2

Employee + AI

Microsoft 365 Copilot

Knowledge work helped by drafting, retrieval and analysis. The human stays in the loop and owns the output.

Test: does a draft or a summary save real time here?

Tier 3

AI Only

Copilot agent / autonomous agent

Structured, repeatable, low-judgement work — run by an agent with a human checkpoint, not a human keystroke.

Test: is this rules-based and high-volume enough to delegate?

How the 176 tasks split

Tier 1 · Employee only
63 · 36%
Tier 2 · Employee + AI
63 · 36%
Tier 3 · AI only
50 · 28%

Just over a third of tasks stay fully human. That is the headline that defuses the fear in the room: this is augmentation, and the most important work is precisely the part that doesn't move.

Automation lean by division (T3 share)

Ports & Logistics
13 / 43
Connect
8 / 25
Health & Beauty
8 / 24
Infrastructure
7 / 24
Group (Corporate)
8 / 40
Energy
6 / 20

Logistics and telecoms carry the most rules-based volume; corporate functions lean toward Tier 2 — assist, don't replace. The mix tells you where to pilot agents and where to pilot Copilot.

Worked example — one role, decomposed

Pick a role to see its O*NET tasks tiered and routed. Notice the pattern: the highest-importance tasks tend to stay human, while the highest-frequency, repeatable ones move to an agent.

T3 AI
Prepare and submit customs declarations for inbound & outbound freight
Copilot agent
imp 4.8
freq 4.9
T2 +AI
Classify goods against tariff schedules and determine duty owed
M365 Copilot
imp 4.7
freq 4.5
T3 AI
Reconcile shipping documents against declared contents, flag discrepancies
Copilot agent
imp 4.5
freq 4.6
T1 You
Advise clients on customs regulations and compliance obligations
Human
imp 4.3
freq 3.8
T1 You
Resolve disputed declarations with customs authorities
Human
imp 4.1
freq 3.2
T2 +AI
Respond to client shipment enquiries across channels
M365 Copilot
imp 4.1
freq 4.9
T2 +AI
Walk customers through troubleshooting
M365 Copilot
imp 4.0
freq 4.6
T3 AI
Log, categorise and route service tickets
Copilot agent
imp 3.6
freq 4.9
T3 AI
Provide status updates and tracking information
Copilot agent
imp 3.8
freq 4.7
T1 You
De-escalate dissatisfied customers / retain at-risk accounts
Human
imp 4.2
freq 3.4
T1 You
Review and dispense prescriptions
Human
imp 4.9
freq 4.8
T1 You
Counsel patients on medication use
Human
imp 4.7
freq 4.4
T2 +AI
Check for drug interactions
M365 Copilot
imp 4.8
freq 4.6
T3 AI
Maintain dispensing records
Copilot agent
imp 4.0
freq 4.5
T2 +AI
Design and build application features
M365 Copilot
imp 4.6
freq 4.7
T2 +AI
Write and review code
M365 Copilot
imp 4.5
freq 4.8
T3 AI
Generate boilerplate and unit tests
Copilot agent
imp 3.9
freq 4.3
T2 +AI
Triage and fix production defects
M365 Copilot
imp 4.3
freq 4.0
T1 You
Make architecture trade-off decisions
Human
imp 4.4
freq 3.0
T2 +AI
Prepare financial statements and reconciliations
M365 Copilot
imp 4.4
freq 4.5
T2 +AI
Analyse variances against budget
M365 Copilot
imp 4.2
freq 4.0
T3 AI
Process and post journal entries
Copilot agent
imp 3.9
freq 4.7
T1 You
Advise on financial controls
Human
imp 4.1
freq 3.0

The reassuring pattern

Across every role, the tasks that stay human are usually the most important ones — dispensing (4.9), counselling patients (4.7), advising clients (4.3), architecture calls (4.4). AI takes the repetitive volume. People keep the judgement. That is the sentence that gets a workforce on side.

03

Capability · Personal learning plans

Generate personal learning plans

Each role has a competency target from O*NET. Each person has a modelled actual. Target minus actual is the gap — and the largest gaps become a prioritised, individual learning plan with a concrete development action. Across the org that generated 14,031 plan lines for 5,908 people.

5,908
People with a plan
14,031
Development actions
4,934
High-priority gaps
9,097
Medium-priority gaps

Deep dive — one person, end to end

Judgment Active Listening Reading Comp. Critical Thinking Science
Role target Frank's actual

Frank Wood

Junior Pharmacist · Health & Beauty · UK & Ireland Retail · Part-time, 3 yrs · SOC 29-1051.00

The radar tells the story at a glance: Frank sits below the role target on every competency, with the widest gaps on Science (4.2 → 3.0) and Critical Thinking (4.5 → 3.4). As a junior on a clinical-facing role, those gaps are high-priority — they touch the Tier 1 tasks AI will never take from him.

His plan, generated from the gaps:

Science1.2 · HighDomain technical-refresher pathway.
Critical Thinking1.1 · HighCase-based decision exercises from real incident reviews.
Judgment & Decision0.9 · MedScenario simulations with structured debriefs.

The join: Frank's AI moment is the Tier 2 drug-interaction check and the Tier 3 dispensing-record agent — which free the time he reinvests in the Tier 1 clinical judgement his learning plan is built to strengthen. Phases 02 and 03 are the same coin.

A spread of people, a spread of plans

Kathryn Lewis
Lawyer · Legal & Procurement
MOD-E00001

Group HQ, London · Mid · 15 yrs · Full-time · 23-1011.00

Writing0.5 · MedBusiness-writing clinic for reports & correspondence.
Judgment & Dec.0.5 · MedScenario simulations with structured debriefs.
Angela Roberts
General & Operations Manager
MOD-E00005

Group HQ, London · Mid · 5 yrs · Full-time · 11-1021.00

Personnel Mgmt0.7 · MedFirst-line leadership programme.
Judgment & Dec.0.6 · MedScenario simulations with structured debriefs.
Coordination0.6 · MedCross-team coordination practice.
Dawn Cole
Junior Software Developer · Connect
MOD-E06275

Core Network & Engineering, UK · Junior · Full-time · 15-1252.00

Programming1.4 · HighHands-on engineering kata and pairing.
Critical Thinking1.3 · HighCase-based decision exercises.
Active Learning1.1 · HighSelf-directed pathway with checkpoints.
Ann Mason
Customer Service Rep · Connect
MOD-E06224

Customer Operations, UK · Mid · 1 yr · Full-time · 43-4051.00

Persuasion0.6 · MedConsultative influence module.

Ann's repetitive ticketing moves to a Tier 3 agent — so her one plan line sharpens the exact human skill (persuasion, retention) her role now hinges on.

Roll it up — the org-wide L&D priority

The same gaps, aggregated across all 14,031 plan lines, tell L&D where to invest first. The top of the list is unmistakable — and it is human skill, not tool training.

Critical Thinking
1,528
Service Orientation
1,241
Active Listening
1,210
Coordination
1,027
Speaking
958
Judgment & Decision Making
862
Complex Problem Solving
826
Systems Analysis
650
04

Adoption · Prompts · Activities · Roadmap

Drive adoption

Now the tool earns its place. Adoption lands when it meets people inside the work they already do — so every prompt below is tied to a real O*NET task and its tier, and every activity is drawn from the actual development actions in Phase 03. Prompts for the Tier 2 work, agents for the Tier 3 work, learning for the Tier 1 work that stays theirs.

Role-grounded prompt library

Tier 2Customs Broker · Classification

Tariff classification assistant

Classify the following goods for import into [country] and propose the most likely commodity code. For each line give: candidate HS code, duty rate, the rule that applies, and any documents I must check before I file: [packing list]

Human confirms the code and owns the declaration. AI does the lookup and the first-pass reasoning.

Tier 2Accountant · Reporting

Variance narrative writer

Analyse this budget-vs-actuals table and write a 2-paragraph narrative of the key variances. Flag anything needing management attention and quantify each driver: [table]

Month-end management accounts. Saves 1–2 hrs/cycle; the accountant still signs the commentary.

Tier 2Software Developer · Review

Code-review co-pilot

Review this diff for correctness, edge cases and security issues. List findings by severity with the line reference and a suggested fix. Do not rewrite — explain so I can decide: [diff]

Keeps the engineer making the architecture call (Tier 1) while AI accelerates the routine review.

Tier 2Pharmacist · Safety check

Interaction pre-check

For a patient already taking [list], flag potential interactions, contraindications and monitoring notes for newly prescribed [drug]. Cite the mechanism. This is decision support only.

Surfaces risks fast; the pharmacist's clinical judgement (Tier 1) makes the call and counsels the patient.

Tier 3Customer Service · Triage

Ticket categorise-and-route agent

For each inbound message: detect intent, set priority, tag the product area, draft a status reply, and route to the right queue. Escalate anything matching the at-risk or complaint patterns to a human.

Runs unattended with a human checkpoint on escalations — frees reps for de-escalation and retention.

Tier 1Ops Manager · Stays human

Decision-prep brief (not the decision)

Summarise these inputs into a one-page decision brief: options, trade-offs, risks, and the two questions I should ask before deciding. Do not recommend an option.

Even a Tier 1 task can use AI to prepare. The accountable judgement stays with the manager.

Adoption activities — pulled from the learning plans

For judgement & decision gaps

Scenario simulations with debriefs

Run real decisions in a safe room, then debrief structured against what good looks like. The single most-prescribed action across the org.

For critical-thinking gaps

Case-based decision clinics

Work real incident reviews as cases. Builds the exact skill AI can't supply and the org needs most (1,528 gaps).

For engineering & technical gaps

Kata & pairing labs

Hands-on, repeated practice with a more senior pair. How Dawn closes a 1.4 programming gap.

For customer-facing gaps

Consultative-influence module

Persuasion, service orientation and active listening — the human edge once agents take the ticketing.

For leaders

First-line leadership programme

Personnel management and coordination for new and stretched managers like Angela.

For everyone · tool fluency

Prompt clinics by role

30-minute, role-specific sessions using the library above on the person's own live work — not generic training.

The 90-day adoption roadmap

Days 0–30

Baseline & align

  • Stand up the org model and capability baseline (Phase 01).
  • Agree the three-tier model with division leads.
  • Publish the "what stays human" list — name it before the fear does.
  • Pick two pilot units: one Tier 2-heavy (corporate), one Tier 3-heavy (logistics).

Days 30–60

Pilot by tier

  • Roll the prompt library into the Tier 2 pilot; run weekly prompt clinics.
  • Stand up one Tier 3 agent (ticket triage or journal posting) with human checkpoints.
  • Launch personal learning plans for the pilot cohort.
  • Instrument it: time saved, quality held, adoption rate, opt-outs.

Days 60–90

Scale & govern

  • Extend to the next two divisions on the evidence, not the hype.
  • Wire learning plans into the LMS and manager dashboards.
  • Set the governance guardrails: attribution, data handling, accountability.
  • Review the tier map quarterly — roles drift, so should the map.

Governance is not a brake

Every Tier 3 agent ships with a human checkpoint. Every Tier 2 output has a named owner. The "what stays human" list is published, not implied. Adoption that skips this is the adoption that gets pulled after the first incident.


What this demonstrates

End to end. One coherent method from raw org structure to a 90-day adoption plan — not a deck of disconnected tactics.

Grounded in data. Every number on this page is computed from a real generated dataset of 8,980 people and O*NET occupation codes — reproducible from a seed.

Honest by design. Synthetic people, flagged placeholder content, explicit "what stays human." The method survives contact with a sceptical workforce.

Start with the organisation. The tool comes last, once you know where it belongs.