A company buys ChatGPT Enterprise. Then:
- marketing uses it for copy
- one engineer automates half their workflow
- another refuses to touch it
- managers say “we should use AI more”
- nobody knows what “more” means
Three months later, leadership asks the inevitable question: “So… are we actually getting ROI from this?”. Silence. Not because AI failed.
Because adoption inside companies is mostly random. And random systems don't scale.
Most companies approach AI rollout like this:
- Buy tools
- Announce initiative
- Hope employees figure it out
That works for maybe two weeks. After that, you get what every company gets:
- inconsistent usage
- inconsistent output
- no shared standards
- no visibility into who’s actually effective with AI
Useful? Sometimes.
Measurable? Not really.
The uncomfortable part: most companies still evaluate people like AI doesn’t exist.
Performance reviews ask communication, ownership & collaboration.
Fine. But now we also need to ask:
- can this person delegate effectively to AI?
- can they verify AI output?
- do they know when not to use it?
- are they faster because of AI — or just noisier?
Because there’s a difference between “uses AI” and “works effectively with AI” is enormous.
And this is where most AI adoption projects quietly break. Not on infrastructure. Not on tooling. On management.
No competency model. No measurement system. No shared definition of “good AI usage” - just licenses and optimism.
The weird part is that companies already solved this problem once.
We don’t say: “Everyone has Excel now, good luck”. We train people, define expectations, measure proficiency.
AI will end up the same way. Except the impact is bigger.
At HRPulsar, we’ve been thinking about this a lot. Not as “AI replacing employees”. But as: how do you systematically measure and develop AI fluency inside teams?
And honestly, we don’t think the industry has good answers yet.
Especially for:
- role-specific AI competencies
- measuring real usage quality
- separating employee contribution from AI contribution One thing already seems obvious: buying AI tools is easy, building an organization that actually knows how to use them is the hard part.
Curious how other teams are handling this right now. Do you actually measure AI adoption in any meaningful way — or is it still mostly vibes?