What it contains, how it is produced, and when it is not worth the money

An AI strategy is a decision-making document that answers which work tasks in your business should be changed with the help of AI, in what order, what that requires, and who owns them afterwards.
It starts from how you actually work, not from a list of available tools. A useful AI strategy is short, prioritised and possible to start working to the same week it is delivered.
An AI strategy easily becomes too long and too general. It describes the possibilities of the technology in twenty pages and end with a recommendation to "explore pilot projects". Nobody can start working to that.
A strategy effort that is worth its money does three things in practice: it finds out where your time actually goes, it strikes out most of what you thought of, and it points to who is responsible for what remains. The rest is justification.
Below is what I believe an AI strategy should contain and how it is produced. It is the same way of working I use in client assignments, and it is deliberately concrete enough that you could do parts of it yourselves if you would rather.
Content
What is done in the business during an ordinary week, by whom, how often, and how long does it take? It sounds trivial, but this is the part that most often goes missing, and without it prioritisation becomes guesswork. This part requires talking to the people who do the work, not only to the people who describe it.
A list where each item has an estimated time saving and an estimated effort, both rough. The point is not precision but order: what gives a lot and costs little should be done first, even if it is dull. In my experience the most profitable items are rarely the most impressive ones.
The most important and most often left out part. A strategy without exclusions is a wish list. Here you write down which areas you will leave untouched for the next year and why, so that the next enthusiastic suggestion can be measured against a decision rather than against the mood of the day.
What may be sent to which services, what GDPR requires, and what the EU AI Act requires of your particular areas of use? For most small and medium-sized businesses this is less dramatic than it sounds, but it needs to be written down, otherwise your staff will guess. And they guess generously.
AI competence is not a course, it is a distribution. Who needs to use the tools every day, who needs to be able to judge a proposal, and who only needs to know what applies? Training everyone in everything is expensive and ineffective.
Not a three-year plan — that would be wrong anyway. Six months, three to five items, and one named person per item. An item without an owner is not a plan, it is a hope.
Process
Half an hour free of charge to establish whether a strategy is the right thing for you right now. Sometimes it is not, and then I will say so.
Interviews with the people who do the work and, where it can be arranged, the chance to watch it being done. This is where the material is.
A working session where the list is gone through together and pared down. You decide, I argue. What survives is what you have chosen yourselves.
A short document and a walkthrough. The material is yours, whether you continue working with me or with someone else.
The whole effort usually takes two to four weeks for a small or medium-sized business. What determines the time is how many parts of the business need mapping and how easy it is to get hold of people — not how long the document takes to write.
You receive a fixed quote after a first free meeting, before the work begins.
Pitfalls
Management knows what ought to be done in the business. The people who carry it out know what is actually done. The gap between those two pictures is usually where the best areas of use are hiding.
A list of fourteen initiatives is not a prioritisation. If you cannot point to what should be done first, and what is deliberately not going to be done, the work is not finished.
The most common way an AI initiative dies is not that it fails — it is that nobody had it on their plate when everyday life returned. Names on paper, not roles.
The longer the document, the less likely anyone is to work to it. If it does not fit on a few pages, the prioritisation has not been finished.
When it is not needed
If you are three people and already know exactly which problem you want to solve, do not buy a strategy. Build the thing. Strategy work is justified first when there are more ideas than time, when several people need to change how they work at the same time, or when earlier attempts have fizzled out without anyone really knowing why.
It is also the wrong moment if nobody in management has the time to make the decisions the strategy requires. A document that nobody acts on is an expensive summary of what you already knew.
I would rather say no to an assignment than deliver a document that ends up in a folder. That is not generosity — it is that I want to be able to point to the work afterwards.
Experience
Glänne & Söner is a one-person business that runs a publishing house, coaching, talks and development in parallel. It only works because I have built an AI team that handles parts of the work — and because I have been forced to prioritise ruthlessly, since I am the only one who can carry out what is left.
Among the things I run in live operation are My AI Team, which builds tailor-made AI colleagues for small and medium-sized businesses, and Nyheter AI, where a newsroom of AI agents monitors 27 sources around the clock. These are systems that have to work on Mondays, not demonstrations.
That experience is what I bring into strategy work. I know roughly what things cost in time to maintain after three months, which is the figure that is easily forgotten in a budget.
Hiring a consultant
Anyone looking for an AI strategy consultant will find consulting firms, freelance consultants and advisers through trade associations and business networks. The difference between them is not visible on the website but in how they work. Four things can be checked before you sign:
1. Do they start from how you work? Serious work begins with interviews and observation of what actually takes time in your business, not with a tour of the available tools.
2. What is the deliverable? A short, prioritised document with named owners and a plan for the first months. Ask to see an anonymised example of an earlier one.
3. Are they independent? Ask whether the consultant earns money from the tools or licences that are recommended. It need not disqualify them, but you should know.
4. Is the document yours? You should be able to continue with it without the consultant, and get a fixed quote before the work starts.
I take on strategy assignments across Sweden, and how a remote assignment works is described on a page of its own. If you would like to start by seeing what I do in general, there is AI consulting for businesses.
Frequently asked questions
For a small or medium-sized business, usually two to four weeks from the first call to a finished document. Most of the time goes on interviews and observing how the work is actually done, not on writing the document.
The scope decides. It depends on how many parts of the business need to be mapped and how many people need to be interviewed. You receive a fixed quote after a first free meeting, before the work begins.
The document belongs to you whether or not you then continue working with me. If you want benchmarks, they are in the article on what an AI consultant costs.
Not always. If you are a few people and already know exactly which problem you want to solve, it is usually better to just build that thing.
A strategy pays off when several people need to change how they work, when there are more ideas than time, or when earlier attempts have fizzled out.
Yes, and this page is written so that you can. The six parts and the process are not secret. What you pay for when you hire someone is mainly that an outsider asks the questions, and that the prioritisation is not steered by internal considerations.
With an AI consultant who builds the strategy from how you actually work and who is independent of tools and licences. There are consulting firms, freelance consultants and advisers through trade associations and business networks. I take on strategy assignments across Sweden, mostly remotely with one or two visits. Four things to check before you choose are under the heading “How to choose an AI strategy expert”.
Yes. I am based in Lindesberg in Örebro County, and most of the strategy work — interviews, prioritisation and the document itself — works just as well remotely. What works better in a room, such as the kick-off workshop, I try to gather into one or two visits.
How such an assignment works is described on the page about AI consulting across Sweden.
Either you carry on working to the plan yourselves, or we continue in a limited build assignment or a few hours a week. I have no lock-in and no licence costs that would make one option favour me over the other.
A free half-hour call. You get an honest answer on whether an AI strategy is the right thing for you right now — even when the answer is no.
Book a first meeting