What 1,000 Businesses Taught Us About AI Adoption: Lessons from 12 Months of Training

We passed 1,000 businesses trained in AI last month. ProfileTree has been running these workshops for two years now and honestly, we’ve learned more about people than we have about technology.

Same patterns everywhere we go. Same mistakes. Same perceptions. And the same “wow” moments when something finally clicks.

Here’s what we have learnt along the way.

Everyone Wants the Advanced Toys

“What are the top 10 AI tools we should be using?”

We get asked this in every session. And I understand why people ask. They want the full picture. The comprehensive list. To be able to do everything from one day of training.

But most people asking this question haven’t even figured out ChatGPT yet. Not properly.

We’ll ask a group – who is using the memory feature? Who’s set up custom instructions? Who’s tried Projects where you upload your documents and it remembers everything?

Hardly anyone. Maybe one or two hands in a room of fifty.

People are using ChatGPT like a search engine and text generator and that is it. Getting generic answers. Then deciding they need a different tool.

There’s so much in ChatGPT that people just skip past. Personalisation. Projects. Team features. Different GPTs for different jobs. Most people are using maybe 5-10% of what is possible.

The phrase we hear constantly: “I thought I was using ChatGPT but I wasn’t even scratching the surface.”

Hence our push is to master one tool fully first. Then worry about what is next.

The Implementation Problem

This reminds me of our digital and website training that we still deliver. People come to our training. They’re engaged. Taking notes. Asking good questions. Leave the room buzzing. “This is going to change everything.”

Then we follow up four weeks later.

Hardly anything has been implemented. We hear the same reasons – as every business – “Been too busy.” “Things got in the way.” “Planning to start next month.”

Next month never comes. We all know this.

What most people do not understand. AI isn’t another thing to add to your list. It’s supposed to make the list shorter. You don’t find time for it – you use it to create time.

The businesses that actually change? They implement something the next day and then continue. One thing. Doesn’t matter how small.

Everyone else adds it to a to-do list and forgets about it within a week. We will see these people at another training course in 6 months. 

Stop Planning. Start Doing.

Bigger companies are the worst for this. “We need to map out our AI strategy across all departments before we do anything.”

They set up a working group. Have meetings about having meetings. Create documents. Discuss security. Commission reports.

Meanwhile some freelancer working from her kitchen is getting twice as much done as their entire marketing department.

We worked with a medium sized firm in Belfast. The owner just picked the most annoying part of his workload – writing tender responses – and started using AI for it straight away. No strategy document. No committee. Just ChatGPT and a great Project set up. 

Saving himself 2-3 days a month now. His competitors are probably still in the planning phase. We have seen in many of the larger companies HR or IT actually blocking a controlled AI rollout, while they are not considering that a lot of their staff are probably already using AI without approval. This Shadow AI use is common in many larger companies through to government departments. Organisations need to run faster to get ahead of the AI train. 

Small Businesses Are Winning with AI Implementation

This is the most interesting pattern we see. Freelancers, consultants, micro businesses – are all over AI. Fully embracing it.

Why? No permission needed. No IT department to convince. No procurement process. They just get on with it.

For a one-person business, AI is like hiring an assistant you could never afford before. Tasks that took all morning are done in an hour. More capacity without more hours of work.

We trained a copywriter in Dublin. Sceptical at the start and thought AI would replace her. By the end she was surprised and annoyed she had not fully reviewed AI sooner. It doubled her output and allowed her to do tasks she was never able to cover before such as keyword research and audit a website’s analytics. 

Large organisations will catch up eventually. But right now there’s a window where small and nimble operations are beating big and slow.

Security Questions

Every training session someone asks about data privacy. Its great to see that there are people thinking about data controls. 

No you should not dump confidential client info into the free version of ChatGPT without understanding what happens to it.

But the answer isn’t to ban everything and allow Shadow AI to take over an organisation. The answer is to understand what you’re using. This is where a great AI Policy comes into its own and supports a company that is trying to roll out AI carefully. 

Create easy to follow policies. Configure the approved AI tools properly. Make conscious choices with any personal data that could be seen as sensitive.

Getting security wrong today can have serious consequences and we have seen a handful of data leaks from leading AI platforms already. 

Still Plenty of Sceptics

Even now after moving from Chat GPT 3.5 to 5.1 – we have a number of people in each training group who do not see AI having a big change or impact on the way they work. 

We’ve stopped trying to convince everyone with slides and theory.

Now we try to do every session live within the AI platforms – i.e. no slides. We have someone gives us an actual problem they’re stuck on. We try and solve it in front of everyone. Real task, real solution, real time.

The “wow” moments don’t come from explaining what AI could do. They come when someone watches their own two-hour job get done in five minutes. These moments are special in every training session. 

Marketing First

About 80% of businesses, we train start with content. Blogs, social posts, emails.

This makes perfect sense. Marketing tasks are repetitive enough to benefit but low risk enough to experiment safely. Time savings show up almost immediately. What took a morning takes an hour. Then people get curious about other areas. The rollout tends to flow from here. 

The marketing teams are the easy converts – the harder work is with finance, operations and sales. But all is possible – its just finding the right problem or task to solve with the best AI solution. 

What We Tell Everyone Now

After a thousand of these conversations our advice has become. Learn one tool properly before chasing others. ChatGPT has more than enough to keep you busy and transform a lot of any job you do.

Start with something that annoys you. Not something impressive sounding. Something you actually hate doing.

Implement something within 48 hours of training. 

The gap between businesses using AI and businesses thinking about it is getting wider every month. We can see it clearly from running all these sessions. Some companies are achieving incredible efficiencies while others have not even looked at AI yet.

The tools are there. The learning curve isn’t steep. Results are real.

The key question is whether your clients will actually do something about it or just add it to the list.