Companies bought the AI licences. Their staff aren’t getting much out of them. In a Gartner survey of 574 Australian employees in the first quarter of 2026, 85% had been given enterprise AI tools, only half reported clear guidance, training or support for using AI, and 86% were also using personal AI tools. The only Australian data that measured actual use is still the government trial. It’s from the first half of 2024, so it’s dated, but in that trial (7,769 licences across nearly 60 agencies) a third of participants used Copilot daily. For reference, Microsoft’s own default bar for an “active user” is one use in 28 days.

That the gap is a people problem is what executives themselves report. BCG surveyed 1,000 CxOs in 2024. Their answer: 70% of AI implementation challenges are people and process, 20% technology, 10% algorithms. That 70/20/10 started life as a BCG partner’s rule of thumb in 2018 and got survey backing in 2024. Field experience first, measurement later. Australia has a sharper local number. In Hays’ survey (fielded February–March 2026; 7,000+ respondents across Australia and New Zealand, 5,223 in Australia; a recruiter’s own-database poll), 60% of respondents use AI at work. And 78% of those users have never received formal training from their employer.

Of the five functions AI jobs sort into, this installment covers Enablement.

Hiring for that gap has started

Three Australian postings confirmed in August 2026 (listings expire, so the substance stays here):

  • AI Enablement Lead: a Melbourne government agency (via recruiter Talent), AUD $190–206k. One person owning the AI roadmap, responsible-AI governance, and the organisational learning program.
  • AI Academy Lead: an ASX 150 company (name withheld), AUD $180–240k package, superannuation included. First assignment: get 300–500 Copilot premium users (mostly executives) actually using it. Prompt libraries, role-specific playbooks, internal how-to videos. And the duties say it outright: “Measure engagement, adoption and learning effectiveness and use insights and feedback to continuously improve the enablement approach.”
  • AI Adoption Specialist: ABC, the national broadcaster. Workshops, prompt libraries, capability clinics.

The backgrounds asked for aren’t engineering. They’re learning and development, change management, digital adoption, communications. One consultancy’s ad copy states this function’s reason to exist better than any brochure: “Most ‘do AI’ with a licence rollout, lunch and learn, and a completion dashboard. 6 months later, the work hasn’t changed much. We’re not doing that.” (Novigi, AI Adoption Coach posting)

The job isn’t teaching. It’s measuring.

Anyone can run a workshop. What this role gets paid for comes after. One finding from the Australian government trial is practically useful here. The more training people got, the more confident they were, and it showed up in numbers. Participants who received three or more forms of training were 28 percentage points more likely to say they felt confident than those who got one form (per the summary report; it’s a between-group comparison, so self-selection bias remains). The trial measured that far. What this role builds is the next thing: before-and-after data.

In BCG’s June 2026 survey (close to 12,000 frontline employees, managers, and leaders in more than a dozen markets), 42% of frontline employees who use AI regularly said they save eight hours a week, a full workday. Yet 66% receive limited or no guidance on what to do with that time, and more than half are not reinvesting it into more strategic work. That is why reported time savings alone don’t prove productivity. The UK Department for Business and Trade’s Copilot pilot (1,000 licences, Q4 2024) reached the same conclusion: 72% satisfaction and small observed time savings, but it “did not find robust evidence” of productivity improvement. This function’s receipts are the measurements that close that gap. Adoption before and after. Active use by team. Hours actually recovered.

Two things to watch

First, this is still a thin job category. Posting volumes are far below AI engineering; the titles are scattered (Enablement Lead, Adoption Specialist, Academy Lead, Transformation Lead); and the fact that two of the three postings above are recruiter-fronted or name-withheld is itself a sign of how early this is. Second, absorption is a real risk: HR and L&D are claiming this territory, and single postings often bundle several functions into one hire; the Melbourne role above does. Safer than going all-in on the title: add this skill to whatever you already do.

One counter-signal for honesty. On its January and July 2026 earnings calls, Microsoft said paid Copilot seats grew from 15 to 30 million, with average weekly engagement “on par with Outlook and Teams.” That’s the vendor talking, and the 15 million was only about 3% of the commercial M365 seat base. But if the shelfware era really is ending, the person who can prove “I got it used” only gets more valuable.

If you build, you’re not underqualified — you’re ahead

The backgrounds listed are education and change, but the postings themselves tip the scales toward people who build. The ASX 150 ad says it verbatim. They want someone who “genuinely lives and breathes AI, actively experiments with the technology themselves.” The role is “very much a ‘roll your sleeves up and do’ opportunity, rather than a role that purely sets strategy or oversees delivery.” Someone who has actually built things has the edge over an L&D generalist.

And the good thing about this function: the portfolio can be built before the job switch. Build one prompt library for your current team and measure how many people use it, before and after. Run one workshop and record what usage looks like two weeks later. A few before-and-after numbers like that are this function’s résumé.

Next up: Governance — the strange market where the rules got delayed and the hiring didn’t stop.