The AI Ansoff Matrix Strategy skill for Claude

What a skill is exactly, and why files like this work, you can read in what are Claude skills.

You will probably recognise the problem the skill solves straight away. Everyone wants to grow, but the options run into each other in every conversation. One person wants to launch a new service, another finally wants to enter that foreign market, and meanwhile the cheapest route, getting more out of the customers you already have, gets skipped because it does not feel exciting. Every option has a different risk, a different investment and a different pace, and as long as you do not put them on the same measure, the option with the best story wins instead of the best ratio.

Igor Ansoff's matrix has existed since 1957 to force exactly that comparison, and this skill turns filling it in into a conversation instead of a strategy day.

The role description in the file sets the tone straight away: a strategy sparring partner that gives no woolly advice, but an honest look at what every growth direction costs, what it delivers and how big the risk is. This file is written for business owners and SMEs who have to make a real choice: between several growth options, for next year's strategic planning, on an investment or pivot decision, or when competition is pushing and the position needs improving. Anyone who says to Claude how do I grow, which direction do I choose or I want to scale up while the skill is loaded gets not a general story but the matrix.

There is one thing you should know up front: the skill weighs, it does not research. It works with the situation, the goal and the risk room you supply, and where figures are missing it works with ranges explicitly labelled as an estimate. The quality of the choice therefore stands or falls on the honesty of your input. Anyone who overstates their own risk room gets advice that looks right on paper and pinches in practice.

The model's most important insight sits in the diagonal: the risk profile rises from top left to bottom right. Market penetration is the lowest-risk route, because you know the product and you know the customer. Diversification is the highest-risk route, and Ansoff himself called that quadrant the suicide square when it is not thought through. That quote is stated literally in the SKILL.md, and it is not there for decoration: the file instructs Claude to always judge Diversification with extra strictness, never boldly optimistic.

Is a model from 1957 still usable? The file itself takes a position on that, with a nuance that many fill-in-the-blanks forms miss. It cites the later additions of Peter Drucker and Henry Mintzberg, and draws an honest conclusion from them: the matrix gives no prescriptive answer but a way of thinking. The value lies in systematically weighing four options that otherwise run into each other. The four routes themselves do not age, because there is no fifth: everything a business can do to grow falls into one of these boxes. What does age is the habit of only looking at one or two of them.

The difference between talking about growth and filling in the matrix shows fastest in the language. This is what the skill does with the sentences that come up in every growth conversation:

Then comes the scope, and it is stricter than you would expect from a growth model. The file lists seven situations in which you apply the skill, from a choice between several growth options to an investment or pivot decision, and sets one hard exception against them: for operational problems, such as a production issue, an HR matter or a customer service complaint, you do not apply it. A matrix chooses a direction, it does not put out today's fire.

The intake consists of six questions, short and businesslike: what is your current product or service, who are your customers or which market are you in, what is your goal, what time horizon applies, how much risk can you carry, and which options had already crossed your mind. What matters is the exception that comes straight after: has the user already given enough context, then the questions are skipped and the analysis starts at once. The question list is a safety net, not a form.

The core of the file is the output structure, and it is uncompromising about it: always deliver this exact structure, no variations. Six parts. A situation sketch in three sentences, neutral and without judgement. The four quadrants filled in, each on the same seven points: a concrete action for this user instead of general advice, a risk score from 1 to 5, the required investment as a level with a rough range, the speed to result, the expected effect, when this quadrant is the choice and when it is too limited.

Then the verdict: in one sentence which quadrant deserves the most energy now, with three arguments. After that three concrete steps for the next 14 days, a skip list with one or two quadrants that are not a priority now, and finally a re-evaluation moment: a concrete moment or a metric at which you check the choice again.

The tone is prescribed too, and it fits a sparring partner rather than a report. Direct, no hedging paragraphs. No it depends as the main answer. Acknowledge risk, but leave the ball with the user. The you-form, no English jargon where a Dutch word will do (the file is written in Dutch and answers in Dutch), and no list of ifs and buts but a real recommendation. The file closes with four hard rules and a source list of five titles, and both get their own section further on.

Those last two titles say something about the file's honesty. Mintzberg wrote one of the best-known critiques of strategic planning, and Wensley questioned blind trust in strategic box models. That a skill which fills in a matrix carries such sources along is not decoration: the file draws from them the conclusion that the matrix is a way of thinking, not a prescriptive answer. The value lies in systematically weighing four options that otherwise run into each other, not in the illusion that a 2x2 predicts your future.

Why the model still works lies in what it forces. Almost every growth conversation derails the same way: the options are discussed one by one, each with its own enthusiastic advocate, and never side by side on the same measure. The seven fixed points per quadrant make that impossible. Anyone who has to honestly write down for every route what it costs, how fast it delivers something and when it is too limited, sees for themselves which route fits their own cash position and risk room. And the skip list forces the closing piece that is missing from most strategy sessions: saying out loud what you are not going to do.

What the matrix does not do is tell you where you stand right now. Other frameworks exist for that, and they are separate skills in the same library. The SWOT analysis skill first maps your strengths, weaknesses, opportunities and threats, and names the Ansoff Matrix itself as the logical next step. The PESTEL Analyst skill analyses the outside forces that decide whether a new market is attractive at all. And anyone with several products who first wants to know which of them should fund growth puts their portfolio on the BCG Matrix Strategy skill: that matrix divides attention over what you already sell, this one chooses where the new growth comes from.

After that, using it is simple: describe your situation, your goal and your doubt, and ask for the matrix. The more complete your story, the less the skill has to ask, because the six intake questions are skipped as soon as the context is already there. A fuller guide for every environment is in installing Claude skills, and anyone who wants to learn to write their own SKILL.md after using skills finds the full recipe in writing SKILL.md.

There are also situations where you are better off leaving it alone, and the file names them itself. Operational problems are one: a production issue, an HR matter or a service complaint is not a growth question, and a quadrant will not solve it. An unclear customer picture is the second: if you do not yet know who your customer is, the skill first sends you to a Lean Canvas or a JTBD analysis, because without a customer picture every growth direction is a gamble. And for small SMEs the honesty the file explicitly prescribes applies: with fewer than five employees, Market penetration is almost always the highest ROI in year one, however much you would like to build something new.

The last honest boundary is the one every framework has: the matrix makes your weighing systematic, not your future predictable. The figures in the analysis are ranges based on your input, the risk scores are estimates, and the re-evaluation moment is there precisely because a choice can turn out wrong. What the skill does do is make sure that at that checkpoint you can look back on an explicit weighing with reasons attached, instead of on a gut feeling nobody can trace any more.

The cheapest growth route from the matrix, Market penetration, almost always starts in the same place: being found by the market you already serve. Before you consider a new market or a new product, it therefore pays to know how much is still sitting untapped in your current position. The free SEO scan shows in a few seconds where your site stands. In the guide which strategy model when you see how this model relates to the others. And if you want to talk through what else AI can do for your business, from loose skills to full automation, we simply do that in a conversation.