Accenture Is Tracking AI Log-Ins. Their Own Staff Call the Tools "Broken Slop Generators."

Accenture told senior staff this month that promotions to leadership roles now require "regular adoption" of AI tools. They've started tracking weekly log-ins. CEO Julie Sweet said the company would "exit" staff who can't adapt.

Two people familiar with the policy told the Financial Times that some of the tools were "broken slop generators." Another said they'd quit immediately if the policy applied to them.

This is not an Accenture problem. This is an enterprise AI adoption problem. And I'm seeing it everywhere.

The stick without the skill

Here's the pattern playing out across large organisations right now:

1. Company buys enterprise AI licences (usually Copilot)
2. Company rolls tools out to everyone
3. Company mandates usage
4. Employees open the tool, find it underwhelming, close it
5. Company tracks log-ins and ties them to performance reviews
6. Employees resent the tool, the mandate, and the company

The missing step? Actually teaching people what the tools do — for their specific role, with their specific workflows.

Most enterprise AI rollouts give employees access to mid-tier models through clunky interfaces and call it "digital transformation." Then they're surprised when adoption stalls.

The seniority inversion

The FT piece had a detail that didn't surprise me at all: three consulting executives admitted that getting senior staff to adopt AI is harder than juniors.

This makes complete sense. A 28-year-old analyst grew up with apps. They'll try anything. A managing director with 20 years of client relationships doesn't want to be told their methods are suddenly outdated — especially by a chatbot that hallucinates.

But here's the thing most companies get wrong: the senior people aren't resistant because they're stubborn. They're resistant because no one has shown them what AI does for their work. Not a generic demo. Not a company-wide webinar. A specific, practical session on their actual problems.

I recently trained a member of her firm's management committee. By the end of our first 90-minute session, she was already building her first chatbot. She told me the session changed her life. This isn't someone who lacked ambition or curiosity — she just needed someone to show her what was possible, with her, in the room.

Log-ins are not fluency

Accenture is measuring the wrong thing. Tracking whether someone logs in tells you nothing about whether they're getting value from the tool.

I could log into Copilot every morning and still get nothing useful from it. Does that count as adoption?

Real AI fluency means knowing which tool to use, when, and why. It means understanding that ChatGPT is better for brainstorming, Claude handles complex reasoning and long documents, and Perplexity actually searches the web properly. It means knowing how to write a prompt that gets a useful output on the first try — not the fifth.

You don't get that from a mandate. You get it from training.

What actually works

I run one-to-one AI training sessions. Ninety minutes, tailored to the person in the room.

Not a course. Not a webinar. Not a company-wide rollout with a Slack channel for questions nobody answers.

In a session, we work on your actual problems with the actual tools. A copywriter learns how to use AI without losing their voice. A lawyer learns legal research workflows. A senior leader learns which decisions AI can genuinely help with — and which ones it can't.

One programme director I coached found the tools incredibly frustrating. Outputs were poor and got worse the more he tried. In one session, I showed him how to manage context window lengths, choose the right model for the task, and structure his prompts properly. He said it was magical — the tool went from useless to useful. Same tool. Same person. Different training.

The difference between mandating AI and training people to use it is the difference between giving someone a piano and teaching them to play.

Accenture chose the piano. They're now penalising people who can't play.

If this sounds familiar

If your company has rolled out AI tools and usage is flat — the tools aren't the problem. The training is.

If you're a senior professional being told to "use AI more" but nobody has shown you what that looks like for your role — that's fixable. In 90 minutes.

Book a session here or message me if you're not sure whether it's right for you. I'll be straight with you.