Everyone has now watched a video about it and almost nobody has changed how they work. The gap closes when the training is attached to one real job rather than to the tool in general.
Quick answer
Courses teach the tool. People need to be taught the job, done this new way, with your material in front of them.
Train on one real task that somebody does every week, not on a demonstration.
Three things have to be learned: what to give it, what to check before it goes out, and when to stop and ask a person.
The review habit is the training. Remove the review and you have trained nobody.
A course gets bought, a few people start it, two finish, and the following month the work is being done exactly as it was before. This is not a failure of the people and it is usually not a failure of the course.
It is that the course taught a tool and the business needed somebody to do a specific job differently. Those are different lessons, and the second one cannot be delivered by a video that knows nothing about your customers, your prices or your approved answers.
Courses Teach the Tool, the Business Needs the Job
Watch what a general course covers. What the tool is, how to prompt it, a tour of features, a handful of examples from businesses unlike yours.
Now watch what your team member actually has to do on Tuesday. Take an enquiry that arrived badly written, understand what the customer is asking for, check it against what you offer, and reply in a way that sounds like your business and does not promise something you cannot deliver.
Nothing in the first list prepares anybody for the second. The missing piece is not knowledge of the tool. It is your material and the rules around it.
Train on One Real Task
Pick a task somebody on the team already does every week, where a mistake is visible and cheap to correct.
Drafting the reply to a common enquiry. Writing up a visit. Turning notes from a call into something the rest of the team can read. Preparing the first version of a quote from a standard list.
One task, one person, for two weeks. That is the whole programme. The reason it works is that the person is not learning a tool in the abstract, they are doing their job with something new in the middle of it, and every correction they make teaches them more than a chapter would.
Resist the urge to roll it out to everybody at once. The second person learns in days from the first person's corrections, which is a far better use of everybody's time than a group session.
Three Things People Actually Need to Learn
Across every task, the same three lessons decide whether this works.
What to give it. Most poor output is a context problem rather than a skill problem. People need to learn to supply the real material: the actual offer, the actual price list, the real previous example, the customer's own words.
What to check before it goes out. Not everything, which is slow, but the specific things that cause damage in your business. Price, timescale, anything promised, anything about qualifications or compliance.
When to stop and ask a person. The hardest one, and the one that protects you. A clear rule for the cases that never get handled automatically, and a named person to hand them to.
Teach those three, with your own examples in front of you, and almost everything else people worry about turns out not to matter.
The Review Habit Is the Training
There is a version of this where the training is a session and the work is unsupervised afterwards. It goes wrong quietly and takes a few months to surface.
What makes it stick is the review. For the first fortnight, the person doing the task shows the output to somebody who knows what good looks like, every time. Not to approve it, to correct it and say why. Those corrections are the actual curriculum, they are specific to your business, and they are worth more than any general material.
After two weeks the review usually drops to spot checks, because the corrections have stopped repeating. That transition is also your measure of whether the training worked.
Install It First, Then Train on What Is Installed
Training people on a general tool and hoping something sticks is the expensive order to do this in.
The sequence that works is the other way round. Decide the job, install it so the business knowledge it needs is already in place, then train the person on the thing that exists. That is what Strategic Marketer means by Train Your Team, and it comes after Build It, Fit It and Install It for a reason: there is no point teaching somebody to operate a system that has not been shaped around how the business runs. Once the Business Intelligence Agent holds the brand voice, the real offers, the customers and the approved facts, training stops being about prompting and becomes about reviewing, which anybody competent can learn in a fortnight.
If deciding the first job is the part that is stuck, that is what the assessment is for: https://strategicmarketer.com/#contact
This also answers the question most owners ask next, which is whether they personally need to learn all of this. They need to be able to judge the output and own the rules. They do not need to operate it, and a business where only the owner can run the new thing has recreated the bottleneck it was trying to remove. The groundwork this depends on is covered in turning a useful tool into an actual system.
Who Should Learn First
Not the most enthusiastic person, and not the most senior.
Pick the person who owns the task and whose judgement about the output you already trust. Enthusiasm without ownership produces experiments that nobody adopts. Seniority without the task produces a demonstration.
One more consideration. Choose somebody who will tell you when it is not working. The failure mode here is a team member who quietly stops using it and does not want to say so, and you find out two months later when nothing has changed.
How to Tell It Worked
Set the measure before you start, because afterwards everybody has an opinion and nobody has a number.
Count the task for one week beforehand. How many of them, how long each takes, how many come back for correction. Then count the same thing a fortnight in. If the time has not moved and the corrections are still repeating, the problem is the material the system was given rather than the person.
The second measure is quieter and more important. Does the work still happen the new way during a busy week? Anything that only survives a calm month was a habit, not training, and habits lose to busy months every time.
Start With One Task and One Person
Choose the weekly task, choose the person who owns it, put your real material in front of them, and review every output for a fortnight. That is a smaller commitment than any course and it is the version that changes how the work is done.
Schedule Your AI Business Assessment →
One conversation. A clear diagnosis of what is holding AI back in the business, the two or three places it would pay for itself first, and a prioritised roadmap you keep either way. No pressure and no obligation, and if it is not a fit, they say so.
Questions Owners Ask About AI Training
Do we need a formal course at all?
Rarely. A course is useful for somebody who will go deep and support everybody else. For the rest of a team, a real task with your own material and a fortnight of review teaches more and takes less time away from the work.
How long before people are competent?
Two weeks of daily use on one task is enough for most people to stop needing the review on every output. Competence across several tasks takes longer, and it comes from repeating this on the next task rather than from more teaching.
What about the people who do not want to use it?
Find out which objection it is. Some are protecting quality, which is worth listening to and often correct. Some are protecting their role, which needs an honest conversation rather than more training. Pushing a tool at either group without answering the objection produces polite compliance and no change.
Should the owner learn it first?
The owner should be able to judge the output and should own the rules about what gets checked. Operating it daily is a different job, and keeping that with the owner is how the bottleneck returns.
How do we stop people pasting in things they should not?
With a rule written down before anybody starts: what material can go in, what cannot, and who to ask when it is unclear. That rule is part of the training, not an afterthought once something has gone out.
Is it worth training people if we may change tools later?
Yes, because almost none of what matters is tool specific. Supplying the right material, checking the risky claims and knowing when to stop and ask a person transfer to whatever you use next.