The SWOT analysis skill for Claude

You'll recognise the problem this skill solves if you've ever built a SWOT. You open an empty template with four boxes, dutifully fill in some points, and after an hour you're staring at four lists you basically already knew. Strengths like good service, opportunities like social media. No decision follows, so the document disappears into a folder and the choice you made it for gets taken on gut feeling anyway. The tool isn't the problem: SWOT has been one of the most used analysis frameworks for decades. The problem is that almost nobody does the second half, crossing the quadrants, and it's exactly that second half this skill is built on.

It's written for anyone who has to choose a direction: business owners weighing up a new service, marketers building the case for a yearly plan, teams preparing a launch, and also professionals who want to review their own career, because a personal SWOT is listed as a separate edge case in the file. It's part of the skill library we make available for free from our AI and automation service, with no account and no sales email afterwards.

One thing you should know beforehand: the skill analyses, it doesn't decide. It sets out the options sharply, crosses them and says where the risk and the opportunity lie, but the choice stays yours. And the quality of the analysis stands or falls with the honesty of your input. Whoever hides their weaknesses still gets them pointed out, because calling things out is stated literally in the instructions, but whoever doesn't know their market gets assumptions back instead of certainties.

The points are vague. We deliver quality isn't a strength, it's an opinion. For every point it demands a short justification: why is this a strength, what shows it, and why does it matter for the goal of this analysis. A point that can't carry a why doesn't make it into the document.

Internal and external get mixed up. Competition isn't a weakness and customer satisfaction isn't an opportunity. Strengths and weaknesses are things you have influence over: team, product, money, knowledge, reputation. Opportunities and threats come from outside: market, customer behaviour, technology, regulation, competitors. The skill keeps that split strict, because the confrontation matrix only works if the axes are clean.

The analysis stops at the list. This is the biggest one. Four filled quadrants feel like a finished product, but nothing follows from them. The confrontation matrix, crossing internal with external, is the step that turns a list into a strategy, and it's exactly the step that gets skipped in practice. In this skill it's compulsory: step four of five, not optional.

Everything has to sound positive. The moment the owner fills it in themselves, a SWOT tends to turn into a marketing leaflet. The skill's tone instruction is explicitly written against this: clear and honest, naming weaknesses without discouraging, and calling it out when everything looks like a strength.

In diff form: what goes out and what comes back in its place.

After that, the file sets out the division of roles. The user provides context about the situation, the goal and the market, and sharpens Claude's points with their own knowledge. If there's too little context, Claude asks a few targeted questions, fills in the four quadrants with justification, builds the confrontation matrix and delivers the action plan. The input checklist has three questions: what the business or person does and for whom, what the analysis is for (a decision, a launch, a yearly plan or a career step), and what the biggest market development or competitor is. If the user has little to offer, the skill starts with clear assumptions and flags them as an assumption, so the user can confirm or adjust them.

The core is the five-step method. First the goal in one sentence, because without a goal it becomes a list with no direction. Then the internal side: strengths and weaknesses, five to seven points per quadrant, each with a short justification. Then the external side: opportunities and threats, with the same requirement. Then the confrontation matrix, where internal is crossed with external into four directions: grow, defend, improve and survive. And finally the action plan: the two most important confrontations translated into concrete actions, with a first step for this week.

What stands out in the file is how much space the edge cases get: five of them. Little information: work with flagged assumptions. Too many points: pick the strongest five to seven. A personal SWOT: strengths and weaknesses as a professional, opportunities and threats in the field. A competitor with no internal data: work with publicly known information and flag the uncertainty. And the best edge case of all: if everything looks like a strength, Claude calls it out and names the weaknesses the user is skipping. The fixed output format closes with the SWOT numbered and justified, the confrontation matrix in four directions, and an action plan with a top priority and two follow-up actions. If you want to learn to write an instruction file like this yourself, the method is in writing a SKILL.md NL.

That makes the framework older than almost everything currently surrounding it in strategy models, and it has survived those decades for a simple reason: the question of what can I do, what can't I do, what's coming at me, and what helps me, is the shortest possible summary of any strategic situation.

What got added over time, especially in Dutch marketing-plan practice, is the confrontation matrix: crossing the internal axis with the external one. That isn't academic decoration. A list of strengths says nothing until you know which opportunity or threat that strength stands against. Two engineers wanting to train is a nice fact; two engineers wanting to train while customers are already asking about heat pumps is a strategy. It's exactly that crossing the skill forces, because it's the step that's almost always missing from self-made analyses.

SWOT is also rarely the only framework you need. The external row of the matrix, opportunities and threats, gets stronger the more systematically you work through the outside world, and there are separate skills for that: the PESTEL Analyst works through politics, economy, society, technology, environment and law one by one, and the Porter Five Forces Analyst NL analyses the forces in your industry, from new entrants to buyer power. Anyone who has to pick a growth direction after the SWOT can lay the four routes of the Ansoff Matrix Strategist alongside it.

And there's a fair scientific caveat worth mentioning: SWOT has faced criticism in the strategy literature for years, precisely because it so often ends as four lists with no consequence. That criticism is fair, and it's exactly why this skill makes the confrontation matrix and the action plan compulsory. The framework isn't the problem, using it half-way is the problem. Anyone who wants to test the outcome for risk afterwards can hand the analysis to the Pre-Mortem Analyst, which has you write down in advance why the plan might have failed.

Even without installing the skill, you can improve your own analysis with these principles: separate internal from external, give every point a why, and don't stop before you've crossed the quadrants.

After that, using it is simple: say what your business does, what you need the analysis for and what's happening in your market, and ask for a SWOT. It asks about anything missing itself, so you don't need to fit anything into a format beforehand. If you're new to skills and want to understand how this kind of file works first, the knowledge base has the broader explanation of working with AI.

There are also situations where you're better off leaving it alone. A SWOT is a snapshot for a decision, not a dashboard: doing it fresh every week is administrating, not choosing. For ongoing tracking of your execution, a goals framework makes more sense, such as the OKR Coach, which turns the chosen direction into measurable goals. And a SWOT isn't a revenue-model analysis: whether the heat pumps from the example actually pay off belongs in a Business Model Canvas, not in a quadrant.

The most important limit is the input. It can ask follow-up questions, flag assumptions and call things out, but it can't know what you don't tell it. For a market you don't know yourself, the outcome stays a pile of flagged assumptions, and that's exactly as it should be: the alternative is false certainty. So use the outcome the way it's meant to be used, as preparation for a decision you make, not as a replacement for it.

There's a good chance your confrontation matrix contains an opportunity that has something to do with being found online, because that's where it sits for most businesses. If you want to know where your site stands now before you put that opportunity into a plan, the free SEO scan shows it in a few seconds. The guide which strategy model, when NL shows how this model relates to the others. And if you want to talk further about what else AI can do for your business, from individual skills to full automation, we're happy to do that in a conversation.