Three levels of AI ambition
Not all AI investment is equal and the organisations experiencing the most meaningful outcomes are the ones that have been deliberate about which level they are operating at, rather than letting urgency dictate terms.
We see three levels at which organisations are applying AI today. Most are active at Level 1. A growing number are building at Level 2. The most forward-thinking leaders are asking Level 3 questions. These are not rungs on a ladder though; the most capable organisations are working across all three simultaneously. They are, however, a useful lens for an honest conversation about where your attention is, and where it isn't.
LEVEL 1:
Efficiency
Removing the constraints that cap how fast and how well the work gets done
This is where most organisations start, and it is often the right place to start. Automating repetitive processes, reducing manual effort, accelerating cycle times - these deliver measurable outcomes quickly and create the internal confidence and capacity to go further. The risk is treating efficiency as the destination rather than what it actually is: the foundation everything else is built on.
The clearest sign of a well-built Level 1 capability is not the cost it removes — it is what it makes possible for the people doing the work. Efficiency that frees a producer to spend more time with clients, or a teacher to spend more time teaching, is already crossing into Level 2. That is not a contradiction. It is the point.
Are we building efficiency as a genuine foundation or just reducing cost?
LEVEL 2:
Human Impact
Using capability to change what's possible for the people doing the work and the people it serves
When AI reaches this stage, the focus shifts from efficiency to tangible results that benefit employees, customers, students, and all individuals affected by the process. The technology may be similar to Level 1; what changes is the intention behind it and the care required to deploy it well.
This is where ethics stops being a compliance consideration and becomes a design requirement. It is also where the returns become harder to quantify but arguably more important to pursue.
Are we building AI around what people need and do we have the governance to know when it isn't working for them?
LEVEL 3:
Societal Impact
Applying capability to problems that couldn't be solved at scale without it
This is the level a lot of organisations have not yet asked about seriously, but the most purposeful leaders are beginning to. AI applied to societal challenges is not a philanthropic exercise. It is the recognition that some of the most durable outcomes, and the most defensible positions, come from solving problems that matter beyond the organisation itself. It also tends to demand the most rigorous thinking about responsibility, governance and the limits of what technology should be asked to do.
What problem exists within our reach and/or responsibility that AI could help address in a way that wasn't previously possible?