Why a company needs AI training, and which kind to choose

A company needs AI training when using artificial intelligence should stop being a private habit of a few people. After the training, the team should have a shared way of working with the chat, along with recorded decisions and materials that everyone can come back to. Buying licenses and giving employees access to the tool will not make that happen by itself. The kind of training is worth matching to the problem the team actually faces. A team that has not yet started using artificial intelligence needs something different from a company that wants to train many people in various uses of these tools, such as writing texts, analyzing data in spreadsheets and preparing presentations. And a team that uses the chat every day, yet a month later has neither shared decisions nor materials it could come back to, needs something different again.

Why the tool alone is not enough

The teams I work with usually already use artificial intelligence. The problem rarely lies in someone not knowing how to write a prompt for the chat. Most often it looks like this:

• everyone works with the chat in their own way, so nobody benefits from the solutions others have worked out;
• decisions and materials prepared with the chat get lost in the conversation history, and a month later nobody can reconstruct what a decision was based on;
• the team knows the tools but does not use them regularly, so people reach for them only when someone happens to remember.

None of these symptoms comes from the team knowing too little about the tool. All three come from the team having no shared way of working, and a company does not get one along with the license. The team has to learn it, and that is exactly why a company orders training.

Three kinds of training for three different problems

Very different services are sold today under the label “AI training”. The simplest way to tell them apart is by the problem they solve and by what stays with the team after the sessions. Off-the-shelf courses from catalogs are sold by large training companies, while I run a hands-on workshop, so the table compares my service with what other companies offer.

A capability demo makes sense when the team is only starting to work with artificial intelligence. A well-run demo also answers the question of which company data must never be pasted into the chat, and in a company that answer is sometimes more important than the list of capabilities itself. A team that already works with the chat will gain little from yet another demo, because its members already know what the tools can do, and what they lack is a shared way of working.

An off-the-shelf course from a catalog is a course that a training company sells to all its clients with the same program. I do not run courses like this, but I consider them a good solution for a company that wants to train many people in various uses of artificial intelligence in a predictable format. The modules have a fixed length and a fixed program, and once one module is finished, it is time for the next one in the skills path. If the company can only buy training with co-funding, it is worth knowing that in Poland some funding sources require the service to be listed in the Development Services Database (Baza Usług Rozwojowych). Before choosing a course, check whether it is listed there.

A hands-on workshop solves the problem of a team that can use the tools but gets no lasting benefit from them. I start it by checking how well each person knows the tools, because teams are rarely even in this respect: some people work with the chat every day, while others have opened it maybe twice. Throughout the workshop, each person works on their own computer and builds their own way of working, so after the sessions they have a working solution rather than notes from someone else’s demo. I also tailor the sessions to what the team does day to day. A team implementing an ERP system or another IT system can use the workshop to analyze its own implementation documentation with the help of artificial intelligence, set up the seed of an implementation knowledge base from it, and learn to manage the implementation on the basis of that knowledge base. After the sessions, the person leading the implementation keeps the knowledge base going. I run this workshop myself, so I am writing about it as the person who sells it.

Four questions to ask before ordering training

Does the team already use artificial intelligence? If not, start with a capability demo. If it does, a demo will cost the team a day that could have been spent working on its own materials.

Can the company only buy training with co-funding? If so, the choice narrows to services that co-funding can pay for, and it is better to settle that condition at the start than after a conversation about the program. My workshop is not listed in the Development Services Database, so when that condition applies, I refer people straight to courses that are.

What should stay with the team the day after the sessions? The form of training depends on the answer. A demo leaves knowledge, an off-the-shelf course leaves a completed module, and a hands-on workshop leaves whatever each person has built on their own computer. If the company wants to change how the team works day to day, knowledge alone is usually not enough.

Who runs the sessions, and what happens afterwards? Training after which the collaboration ends is a different service from working with someone who, after the sessions, can bring the new way of working into the tools the team uses every day. It is worth asking about further support before placing the order, because after the sessions it may turn out that the team has the materials but nobody to help apply them in everyday work.

What training will not do

One day of training will not produce a shared company knowledge base that the whole team keeps adding to. I teach each person to run their own knowledge system, because a system like that works best when one person looks after it. Notes that many people write to one place at the same time quickly stop forming a coherent whole. For the same reason, a workshop for a team implementing a system produces only the seed of an implementation knowledge base, which is then maintained by one person, the one leading the implementation. Building a shared knowledge base for the whole company is a separate task and a subject for a separate conversation.

One day also only allows a certain amount of work. After the last edition of my workshop, three of the five people who filled out the survey rated the pace as too fast. Everyone built a website during the workshop, but only one person published it the same day. I drew a conclusion from the surveys: today I offer the same program over two days, or with a narrower scope in one day. So when ordering training, it is worth asking not only about the program, but also about how much of it the team will realistically manage to do in the time planned.

What comes next in this series

In the coming articles in this series I will describe what a workshop day at the computers looks like, what further collaboration looks like when the team wants to develop the way of working it built in the sessions, and how to tell whether the training has had an effect. The course of the last edition of my workshop is described in the case study of the “second brain with AI” workshop. If you would like to talk about running it for your team, get in touch.

All my training and workshops are in Polish. For websites and search visibility, I also work in English.