What AI capabilities do organisations need?
AI is now part of everyday work in many organisations. The harder question is what people need to be able to do with it, and how that capability can be developed, assessed and recognised.
3 min read

On this page
Most organisations no longer ask whether AI will affect their work. They ask what their people need to know, what they need to be able to do, and how anyone can tell when those capabilities are in place.
Beyond tool familiarity
Early conversations about AI skills often focused on access: which tools staff could use, and whether they had been shown how. Familiarity is a useful starting point, but it says little about whether someone can use AI well.
Capability shows up in the quality of the work. It is the difference between accepting an output and checking it, between automating a task and improving it, and between using a tool and knowing when not to.
Four layers of AI capability
A practical way to describe AI capability is in layers. Each builds on the one before, and each can be developed and assessed in its own way.
- 01
Foundational understanding
What AI systems can and cannot do, how outputs are produced and where errors tend to appear.
- 02
Practical application
Using AI on real tasks: framing requests, reviewing outputs and integrating results into existing work.
- 03
Judgement and responsible use
Knowing when AI is appropriate, protecting data, checking for bias and owning the final outcome.
- 04
Role-specific development
Applying AI to the decisions, processes and standards of a particular role or sector.
Why role context matters
A finance analyst, a care coordinator and a software engineer may use similar tools, but the decisions they make with them are very different. The standards they work to, the risks they manage and the evidence they need are specific to their role.
This is why generic training rarely closes capability gaps on its own. Development is most effective when it connects AI use to the real tasks, standards and responsibilities of a role.

“AI capability is not a single skill. It is the ability to apply sound judgement to new tools, in the context of real work.”
Developing and recognising capability
Organisations need a consistent way to describe the capability they expect, and to evidence it once it has been developed. Without that, investment in skills is hard to plan and harder to measure.
Qualifications and standards provide a shared language. They set out what competent practice looks like, how it can be assessed fairly and how it can be recognised by employers, training providers and sectors alike.
What organisations can do now
There is no single route, but a few steps help most organisations move from experimentation to confident, consistent practice:
- Map where AI is already used across teams, and where it could add value.
- Define the capability each role needs, not just the tools it uses.
- Build responsible use into everyday practice and guidance.
- Assess capability through practical tasks rather than recall alone.
- Use recognised standards to plan development and evidence progress.
Looking ahead
AI will keep changing, and so will the skills it demands. The organisations best placed to benefit will be those that treat capability as something to define, develop and recognise, rather than something to assume.
More insights
Related thinking on AI skills, qualifications and the standards behind them.
