Executives, managers and professionals move through four phases in parallel, on a government-backed competency standard, tailored to your industry, your technology environment and your baseline. This page sets out what happens, in what order, and what the enterprise holds at the end of each step.
Executives, managers and professionals use AI for different purposes and carry different accountabilities. Each level therefore has its own diagnostic, its own sessions and its own deliverables, assessed on the same standard.
Set the direction, own the governance and sponsor the use cases that matter.
Board and executive committee
See the journeyTranslate the direction into how their teams work, and coach adoption day to day.
Middle management and team leads
See the journeyApply AI to their own work with skill, judgement and accountability.
Everyone who uses AI in their role
See the journeyThe engagement runs as a cycle. Advance measures against the Assess baseline and sets the scope of the next cycle. Each phase closes with a defined output.
A short survey, structured interviews and an inventory of the AI tools already in use establish where each level of the organisation stands.
You get A capability baseline per level and a recommended scope.
In detailLeadership agrees the direction and the rules before anyone is trained: what AI is for, which tools are sanctioned, who decides what, and how success will be measured.
You get An AI usage policy, agreed roles and targets for the first cycle.
In detailEach cycle activates knowledge, applies it to real work and anchors it as working practice. Each cycle builds on the last.
You get Knowledge scores before and after, assessed work products, and behaviour commitments followed up in the workplace.
In detailThe baseline measures are repeated, the change is reported, and the next step is agreed: a further cycle, another level, or hand-over to the organisation.
You get A results report for the board.
In detailActivate, Apply, Anchor. Each cycle develops knowledge, skills and behaviours together, and builds on the one before.
How AI systems work, where they fail and how output is verified. Concise inputs with a knowledge check before and after, so technique is built on an accurate mental model.
Practice on the organisation's own cases, on sanctioned tools or in the secure sandbox, assessed against a rubric that distinguishes work that is good from work that is merely done.
Commitments, workplace assignments and follow-up ratings in the weeks that follow, so new practice holds under the pressure of the day job.
You get Knowledge scores before and after, assessed work products as evidence of skill, and behaviour commitments that are followed up and rated rather than assumed.
Knowledge, skills and behaviours are assessed separately, with evidence from the workplace, and reported per level and per competency.
Multiple-choice items per session, taken before the session and again as a retention check. Answers are never disclosed, so the movement measured is real.
Scenario checks in session, debriefed in plenary, and a ten-point rubric applied to the deliverable each exercise produces. Assessed once, in the room, where it can be observed.
Self-ratings on day one, a scheduled assignment completed in the participant's own work, and a re-rating four weeks later at the close.
Assess closes with a baseline, Align with a policy and agreed roles, each delivery cycle with evidence of knowledge, skill and behaviour, and Advance with a result the board can act on.
No two organisations start from the same position, so no two cohorts run the same programme. Three inputs shape every engagement.
Cases are drawn from your sector and your own operations, so the judgement participants practise is the judgement their role demands.
Sessions run on the AI tools you have adopted or are evaluating. Verification practice is taught to hold whichever tool is in use.
The Assess results determine which sessions each level needs, how far each goes, and where the first cycle begins.
Practice happens on real work, so the environment it happens in matters. The organisation chooses the workspace; the standard of care is the same in both.
Delivery on the enterprise AI tools the organisation already governs, under its own access controls.
Where no sanctioned tool is yet in place: named data residency, no training on inputs, one login per participant, access closed at programme end.
Organisational data never enters a public tool.
Executive leadership, human resources, learning and development, and digital and technology each get a direct answer.
Board and executive committee. Set the direction, own the governance and sponsor the use cases that matter.
Short fortnightly briefings with ten-minute pre-reads. Each briefing produces a governance deliverable the organisation retains.
AI strategy and value
Briefing · 90 min
Governance, ethics and regulatory readiness
Briefing · 2 h
Investment, procurement and platform decisions
Briefing · 90 min
Workforce transformation and culture
Briefing · 2 h
Risk, assurance and incident readiness
Briefing · 90 min
External representation and public trust
Briefing · 1 h
An AI strategy on a page: priorities, guardrails and decision rights
A governance and regulatory readiness checklist, with criteria for evaluating vendors and platforms
Risk register entries, an incident escalation path, a workforce transformation plan outline, an external positioning statement and a stakeholder map
Middle management and team leads. Translate the direction into how their teams work, and coach adoption day to day.
Fortnightly workshops, each followed by an applied assignment completed in the manager's own team: compliance, opportunity cases and vendor oversight.
Leading by example: championing responsible AI
Workshop · 3 h
Coaching and first-line support for AI use
Workshop · 3 h
Policy, compliance and risk in the team
Workshop · 3 h · with workplace assignment
Spotting opportunities and feeding back to leadership
Workshop · 2.5 h · with workplace assignment
Managing the team through AI-driven change
Workshop · 3 h
Owning an AI initiative: alignment and third-party tools
Workshop · 4 h · with workplace assignment
A team-level AI risk and readiness assessment with a live compliance view
Opportunity cases, built from team insight, that leadership can act on
A vendor oversight assignment applying the organisation's assurance criteria to a third-party tool
Everyone who uses AI in their role. Apply AI to their own work with skill, judgement and accountability.
Two e-learning sessions, three workshops, a workplace application log and a personal development plan.
Recognising AI and your rights as a user
E-learning · 1 h
Personal data, privacy and trustworthy content
E-learning · 1.5 h
Adopting AI tools in everyday work
Workshop · 3 h
Professional judgement and responsible use of AI outputs
Workshop · 3 h
Adapting your role and growing with AI
Workshop · 2 h · with workplace assignment
A workplace application log per participant: the task, the AI output, the check applied and the decision taken
A personal development plan that responds to AI-driven change in the role
Knowledge scores before and after, and behaviour ratings at baseline and four weeks on, per competency
3 to 4 weeks
A short survey, structured interviews and an inventory of the AI tools already in use establish where each level of the organisation stands.
A survey of each level to be assessed, mapped to the competency standard
Structured interviews with executives, a sample of managers and a sample of professionals
An inventory of the AI tools in use, sanctioned and unsanctioned
A baseline of what people know, can do and actually do with AI, per level
You get A capability baseline per level and a recommended scope.
2 to 3 weeks
Leadership agrees the direction and the rules before anyone is trained: what AI is for, which tools are sanctioned, who decides what, and how success will be measured.
A leadership session that converts divergent expectations into one agreed direction
Sanctioned tools, decision rights and the limits of acceptable use
Roles and expectations for AI, defined per level
Targets for the first delivery cycle, set against the baseline
You get An AI usage policy, agreed roles and targets for the first cycle.
8 to 12 weeks per cycle
Each cycle activates knowledge, applies it to real work and anchors it as working practice. Each cycle builds on the last.
Activate: concise inputs on how AI works, where it fails and how output is verified, with a knowledge check before and after
Apply: practice on the organisation's own cases, on sanctioned tools or in the secure sandbox, assessed against a rubric
Anchor: commitments and workplace assignments, followed up and rated four weeks on
Each level on its own rhythm: briefings for executives, workshops for managers, applied sessions for professionals
You get Knowledge scores before and after, assessed work products, and behaviour commitments followed up in the workplace.
3 to 4 weeks, then quarterly
The baseline measures are repeated, the change is reported, and the next step is agreed: a further cycle, another level, or hand-over to the organisation.
The Assess measures, repeated
A results report per level and per competency, with the workplace evidence behind each score
The next step agreed: a further cycle, another level, or hand-over
A quarterly check to confirm that the change has held
You get A results report for the board.
Each pillar is assessed separately and with workplace evidence, so the movement recorded is real and the result can be defended.
Multiple-choice items per session, taken before the session and again as a retention check. Answers are never disclosed, so the movement measured is real.
Scenario checks in session, debriefed in plenary, and a ten-point rubric applied to the deliverable each exercise produces. Assessed once, in the room, where it can be observed.
Self-ratings on day one, a scheduled assignment completed in the participant's own work, and a re-rating four weeks later at the close.
Knowledge check and behaviour self-ratings.
Workshops, scenario checks and workplace assignments.
Ten-minute knowledge checks.
Behaviour re-ratings and the cohort report.
A training course
This solution
Starts with
A course catalogue
A measured baseline of your people, your tools and your risks
Who sets the rules
Nobody, or IT after the event
Your leadership, before anyone is trained
Where skill is built
In a classroom, on sample tasks
On the organisation's own cases, on sanctioned tools or in a secure sandbox
What is measured
Attendance and satisfaction
Knowledge, skill and behaviour, before and after, with workplace evidence
What remains
Slides
Policies, deliverables, behaviour commitments and a results report
How it ends
When the sessions stop
When the measures move, with a quarterly check afterwards
Earlier technologies rewarded the people who learned the tool. AI rewards the people who can judge its output. Four properties of the technology explain why, and twelve patterns of behaviour determine whether adoption succeeds.
AI systems are probabilistic. The same input can produce different outputs, so dependable use rests on verification and professional judgement rather than on operating the tool.
Generative models produce fluent, well-formed content that can be wrong. Human oversight is required before output informs a decision or reaches a client.
Productivity gains arrive with matching exposure: data leakage, bias, breaches of confidentiality and intellectual property risk grow at the same rate as the benefit.
Workforce behaviour divides between automation bias and algorithm aversion. Both are recognised failure modes of adoption, and conventional change management addresses neither fully.
None of these is a fault in the technology. Each is a predictable pattern of human behaviour, which is why the solution develops knowledge, skills and behaviours rather than tool operation. All twelve are tracked from Assess to Advance.
Fluency bias
Confident, well-formed output is read as correct, whether it is or not.
Automation bias
Output that looks complete is accepted, and verification decays into a formality.
System abandonment
One visible failure, and users conclude the tool is unreliable and stop using it.
False authority
Agreement is read as a second opinion, confidence as knowledge and tone as intent.
Skill decay
Skills delegated to AI atrophy until people can no longer judge the output for themselves.
Delegation failure
Thin instructions produce poor output; users blame the tool and disengage.
Peer conformity
People who would challenge AI alone defer to it in a group.
Unsanctioned use
Confidential data flows through personal accounts when sanctioned tools fall short.
Prohibition backfire
Banning AI removes visibility and moves use onto personal devices and accounts.
Risk creep
Use expands from low to high stakes without verification standards rising to match.
Unmeasured adoption
Scale without a baseline: value cannot be demonstrated and errors cannot be traced.
Voice uniformity
Similar tools and similar prompts erode the organisation's distinctive voice.
Three inputs, gathered before the first session, shape every cohort.
Before the first session we gather the situations your people actually face: the documents they draft, the decisions they check, the customers they answer. These become the cases, so no participant practises on an example from another sector.
Sessions run on the AI tools the organisation has adopted or is evaluating, or in the secure sandbox where no sanctioned tool is yet in place. Verification practice is taught to hold whichever tool is in front of the participant.
The Assess results set the scope: which sessions each level needs, how far each goes, where the first cycle begins and what the targets are. Each subsequent cycle is scoped against the results of the last.
Session length, cohort size and journey length are agreed in the Align phase.
Delivery on your sanctioned tools, or in an enterprise-grade secure sandbox. In either case, organisational data never enters a public tool.
Delivery on the enterprise AI tools the organisation already governs, under its own access controls.
Where no sanctioned tool is yet in place: named data residency, no training on inputs, one login per participant, access closed at programme end.
Data residency the organisation can name and audit.
Nothing a participant enters is used to train any model.
The same prompt can be run on two or three models. The verification practice is the same whichever tool the organisation adopts.
Issued with the first workshop invitation, sample files pre-loaded, access closed at programme end.
Safe practice is taught by example. The three questions participants learn to ask of the workspace become the three they ask of every AI tool afterwards.
Which country is my data stored in?
Is anything I enter used to train a model?
Who can see it?
Article 4 of the EU AI Act requires organisations that deploy AI to ensure a sufficient level of AI literacy among their staff. The NIST AI Risk Management Framework and ISO/IEC 42001 place workforce competence at the centre of responsible AI. The solution is designed to meet these expectations with evidence per competency rather than attendance records.