Unmanaged AI: role-based capability that boards can fund

Level 4: Role‑based AI

The fourth level is about focus: moving from scattered experiments to persona‑specific AI that maps clearly to KPIs and business outcomes.

Here is where Lexel’s workshop facilitation will:

• Define your primary personas (such as sales, finance, operations, service, marketing, technical).
• Identify 2–3 high‑impact AI use cases for each persona that can be delivered with available tools.
• Match licences, devices and training to those personas rather than spreading AI capabilities thinly and hoping value appears.
• Establish an ongoing training and onboarding rhythm so AI skills don’t disappear when people change roles.

By this stage, boards are no longer being asked to fund “AI” in the abstract. They are funding specific improvements for clearly‑defined roles, backed by usage data, security controls and a roadmap that has already delivered value at Levels 1–3.

What sits beyond: advanced automation and custom AI (Level 5)

There is a fifth level in the framework – advanced automation and custom AI solutions that integrate deeply with your data, workflows and line‑of‑business systems.

Our (Lexel’s) view is clear: most organisations should not start here.

Only when foundational governance, secure productivity, existing app AI and role‑based use cases are in place does it make sense to explore more complex automation and data‑driven AI. At that point, you have a stable runway, real usage data and proven value to justify deeper investment.

A future blog will touch on what this looks like for organisations that are ready – but for most NZ SMBs and mid‑market businesses, the main opportunity over the next 12–18 months lies firmly in Levels 1–4.

If you recognise your organisation in any of the challenges above, the quickest next step is simple: