Four ways organisations build AI skill today, compared honestly: what each one actually changes, how long it takes, what it typically costs, and who it fits. Not a ranking of specific vendors — a comparison of the approach itself, so you can place whatever you are evaluating against it.
| Self-serve AI courses | Corporate LMS modules | Consultancy engagement | Capability programs (VERIFY) | |
|---|---|---|---|---|
| Primary focus | Individual tool skill: prompting, features, use cases | Awareness and policy compliance across the workforce | A strategy document or a pilot in one function | The organisation: workflows, roles, standards, measurement |
| Format | Self-paced video, a certificate at the end | Assigned modules inside the existing e-learning platform | Workshops and a deck, delivered by outside consultants | Live sessions plus coaching, run on your own real work |
| Typical time to complete | A few hours to a few weeks, self-paced | 1–2 hours per employee, once | 8–16 weeks for a strategy phase | 8–12 weeks, sequenced across the six VERIFY moves |
| Typical cost range | €0–50 per person | €10–40 per seat per year, bundled licensing | €50,000–500,000+ per engagement | €390 per person (founding price), guided program |
| What changes when it's done | People know the tool better. The work is unchanged. | A completion record. Rarely a behaviour change. | A roadmap. Often shelved once the consultants leave. | A redesigned workflow, a named owner, a measured baseline |
| Best fit | Individuals who want to get faster at a task themselves | Compliance and baseline awareness at scale | Organisations that need an outside strategic case made | Teams ready to change how the work itself gets done |
Cost ranges are typical market figures for each category, not quotes for any named provider. The capability-program figure is AI Capability Academy's own founding price.
Citizen-facing decisions need an explicit autonomy standard and an audit trail, not just tool familiarity. A capability program builds the governance and workflow redesign the EU AI Act increasingly requires; a course alone does not.
If the mandate is "build AI capability" rather than "deliver a course", the metric that matters is workflow change, not completion rates. A capability program gives you something to report that is not a training number.
You already have the mandate to redesign work. A capability program gives you the sequence and the evidence base; a consultancy engagement gives you a document. The difference shows up 90 days later.
AI training teaches individuals to use tools better — prompting, features, use cases. AI capability building redesigns the organisation's work itself: which workflows change, who drafts and who reviews, what standard a good AI-assisted output meets, and how progress is measured. Research from MIT NANDA (2025) finds 95% of enterprise AI pilots show no measurable bottom-line effect — the common pattern is skilled individuals inside unchanged work.
Self-serve courses run roughly €0–50 per person. Corporate LMS modules are usually bundled into existing licensing, around €10–40 per seat per year. Management consultancy engagements to build an AI strategy typically run €50,000–500,000 or more. Organisational capability programs, which work directly on redesigning workflows rather than producing a strategy document, are usually priced per person or per team rather than per engagement.
Independent research says rarely on their own. BCG's 2025 '10-20-70' study finds 70% of AI value comes from people and process, not the algorithm or the tool — meaning the training itself is a small share of the outcome. McKinsey's 2025 State of AI report finds workflow redesign, not tool literacy, is the single biggest predictor of bottom-line impact, and only 21% of organisations have done it.
Organisations following a sequenced approach (map the workflow, redesign it, put coaching and leadership behind it, measure output) typically see a first measurable change within 90 days on one workflow, with the full six-move sequence run over roughly 8–12 weeks per team.
It has some value for individual productivity, but the evidence base is consistent: skilled individuals inside an unchanged workflow is the exact pattern behind the 95% of enterprise AI pilots that show no measurable bottom-line effect (MIT NANDA, 2025). Training compounds only once the surrounding work is redesigned to use the new skill.
Your move
The free VERIFY capability scan scores you across the six moves in ten minutes — a baseline, not a sales pitch.