Companies are failing at AI implementation in predictable ways. They either roll out tools without training, or rely on generic platforms like Coursera/Udemy, where people treat it like infotainment. The problem is that, once employees see something cool, they’re like “okay, that’s interesting” but often they’re shown countless examples that are not relevant to their job. No use showing Mr Smith from accounts how to deal with logistics using NotebookLM.
I’m seeing this right now with companies who “go all-in on AI” – they buy the training, they get people to join the call, and then 95% of people are camera off, just doing other things. Without a clear mandate from leadership saying “we’re investing in this, we expect you to give it the benefit of doubt, we are giving you the space to do so,” the training just doesn’t stick.
Here’s what makes this particularly challenging – companies often misinterpret these failures. When they see low adoption, instead of questioning their training approach, they look at it as “okay, that must mean the hype people were right, it’s actually just not worth it.” Right now their default is thinking that training to fail means the technology is not working.
ROI versus retention
It sounds like an absurd claim, but I’ve seen firsthand 300x ROI when companies get this right. Let me give you a concrete example from one of our AI coaches: A US National Park ranger had this process where every time anything needed replacing in a park – even just a broken window – you had to fill out a lot of paperwork. And, because it’s the government, it’s a lot of paperwork.
He told us it takes about two hours just to do paperwork for one broken window. In 45 minutes, thanks to the right type of training, he wrote a GPT that automated the whole process. But here’s where it gets interesting – other park rangers across the US started using the same GPT, and they’ve calculated it will save 20 years worth of work next year alone.
That’s 45 minutes of learning translating into decades of saved time.
On a macro level, this perfectly illustrates the US-Europe divide. The US is investing for ROI – they see it as critical to their business. Europe is still investing just for retaining employees, looking at this as an employee perk versus something critical to the business.
It’s literally just that – they see the opportunity and go for it in the US, while in Europe, people just see the risk and wait.
How training has to change
At Mindstone, we decided to significantly reduce the technical requirements – and the complexities – of what we teach. Why? Well, it’s better to focus only on tools that everyone in a company should have as part of their toolset. Not just what they could use – what they must use.
Here’s the thing – if it isn’t dramatically obvious after the training instantly that you’re getting value, then you definitely have not got the right training. Companies get stuck in this evaluation of tools, and then if this training doesn’t get them there, they think the problem is with the tool. It’s actually the other way around – your training wasn’t worth it.
Add to this the fact that a lot of companies think they have an internal expert or champion, whereas the truth is that these individuals are often nowhere near practically useful enough. They might be at the cutting edge of the technology, but they don’t know how to use this everywhere else.
So they end up designing this thing internally, but it falls flat. In tangible terms, the CTO is often the fall guy when HR might have a better shot of how to roll out AI across an organisation more practically.
What’s coming next
In 2-5 years, knowing how to use AI will be as expected as being able to send an email without cc’ing the entire company by mistake. The gap between those who invest in proper training and those who don’t is going to widen dramatically. Companies that understand this now and invest accordingly are the ones that will thrive.
Others are looking at this completely backwards. They’re getting stuck evaluating tools endlessly when the problem is literally right in front of them: their people don’t know how to use what they already have. It’s not about having fancy AI – it’s about actually using it.
The companies that get this now, that actually invest in proper training, are the ones who are going to be miles ahead. The rest will still be sitting there wondering why none of their expensive AI tools are doing anything useful.
