The scene is always the same. Meeting room, flip chart or PowerPoint slides, four quadrants drawn. The entrepreneur or the external consultant asks the ritual question — "okay, let's start with the strengths" — and the team begins to fill the boxes with what it knows, what it senses, what it hopes. Two hours later there's a 2x2 grid full of entries. It looks like serious work has been done. It's even been printed.
That grid, in 90% of cases, will never be opened again. It will become an attachment at the back of the business plan, a generic reference in some sales slide, an archive item. And in the next six months, when the company actually has to decide whether to enter Poland or Romania, whether to open a sales office in Dubai or work remotely, whether to look for a distributor or sell directly — that grid won't guide any of those decisions. The decisions will be made with other criteria, often instinctive, sometimes good, sometimes terrible.
It isn't the SWOT that's a weak tool. It's the way it's used that makes it a reassurance ritual rather than an exercise in judgment. And in the specific context of export, where the variables are far more numerous and the margin for error costly, this weakness becomes dangerous.
Why generic SWOT fails for export
SWOT was born as a tool of internal strategic thinking, for a company that looks at itself in the mirror in its home market. In that context it works decently, because whoever fills it in has direct competence on almost all the entries: they know their own strengths, recognize their own limits, have a real feel for the competition, have reliable intuition about the opportunities.
Abroad, no. Abroad almost none of the four cells is populated with real information. The "strengths" are transferred from the home market by analogy, without verifying whether they're still strengths in that context. The "weaknesses" tend to be the ones you already knew, not the new ones that emerge only in a foreign market — language, yes, but also less obvious things like the fact that your standard payment cycle is impractical for an American buyer used to net 60. The "opportunities" are fished out of generic trend reports. The "threats" stop at "local competition" and "tariffs," when the real risk is almost always cultural, regulation-specific, or reputational.
The result is a SWOT filled in with a panoramic view, not a ground-level view. And panoramic views give the illusion of having seen, while hiding exactly what you need in order not to make mistakes.
There's then a second problem, more subtle. The classic SWOT treats the four categories as if they were independent — strengths here, opportunities there — when in reality the strategic decision arises from the intersection. A strength becomes relevant only if it intersects an opportunity. A weakness is critical only if it intersects a threat. Filling in the four quadrants without then putting them into a system is like having four piles of cards on the table and never playing anything with them.
What makes an export SWOT genuinely useful
There are SWOTs that decide something, even for export. They have three characteristics, and all three require real work — you don't reach the result with a two-hour workshop.
First characteristic: every entry is a verifiable fact, not an opinion. "Our product has premium quality" isn't a strength, it's self-congratulation. It becomes a fact when it turns into: "Our product is X certified, won Y in 2024, and is cited as a benchmark by Z." Same logic for all the other cells. "There's growing demand for organic in Europe" isn't an opportunity — it's a newspaper headline. It becomes an opportunity when you write: "In Germany the organic-premium segment in our category is growing 7% a year since 2022, with market share gone from X to Y, and three main players that today import 40% of the requirement." It changes everything: from hearsay to plannable.
Second characteristic: it includes the target market in the same grid, not separately. The classic SWOT limits itself to the company. The SWOT that decides something for export integrates the specifics of the target market into every cell. Not "we're weak on international logistics" in general, but "we're weak on logistics to Southeast Asia in particular, where our delivery times are double the local average." Not "opportunities in emerging markets," but "opportunity in Indonesia where our category is 60% imported and none of the three main European players yet has a local distributor." A SWOT that doesn't name the market specifically is a SWOT that talks about nothing.
Third characteristic: it produces testable decisions, not descriptions. The useful output of a SWOT isn't the grid. It's the list of three or four concrete moves that derive from the grid, with criteria for verification over time. Example: "We enter the Polish market through an exclusive distributor, with a maximum investment of 40k€ in the first six months, and a decision review at twelve months based on reaching these three specific metrics." If the SWOT doesn't produce something of this kind, it hasn't done its job — it has only given the illusion of having done it.
AI has changed the game. But not the way you think.
Until about two years ago, doing a serious SWOT on foreign markets required weeks of research, access to costly databases, often external consultants. It was expensive enough to be done a few times, hopefully well. Today the same SWOT, populated with verifiable, market-specific facts, you generate in half a day with Perplexity, Claude, or ChatGPT — asking for sources, updated data, regulatory references, profiles of local players.
It seems like good news. In part it is — the cost of access to the tool has collapsed, and this democratizes export planning for small companies that previously couldn't afford that level of analysis.
But there's a mirror risk, bigger than the one companies are recognizing. The AI-generated SWOT is plausible, well-structured, and almost always superficial. AI is extremely good at filling the four quadrants with entries that sound sensible. It's much less good at distinguishing which of those entries are actually true for your specific company in the specific market you want to enter. It generates, with great fluency, the illusion of analysis.
The SWOT generated in half an hour with AI risks being worse than no SWOT, because it gives the same sense of reassurance as one generated with two weeks of real work — and the entrepreneur doesn't distinguish the two. The pattern already observable in companies that have integrated AI without a critical filter is clear: more SWOTs produced, faster decisions, identical or worse decision quality.
The right approach is to use AI for the first pass — generating a reasonable analysis, identifying the entries that the human eye missed, building a starting hypothesis — and then use human judgment for the second pass: verifying every entry, correcting the approximations, eliminating the plausible-but-false, adding what only someone who knows their business can know.
Total time: a day. Result: a SWOT that really decides something, built at the cost of one that ten years ago cost twenty times as much.
The complementary tools, but only if you actually use them
Every export blog in the world says to "integrate SWOT with PESTEL, Porter, benchmarking." It's true in theory. In the practice of companies that don't have a dedicated strategy function, accumulating frameworks becomes an accumulation of rituals that get filled in and not used.
If you have to keep just one alongside the SWOT, keep PESTEL, but used asymmetrically — not as a grid to fill in, but as a risk checklist for the "opportunities" and "threats" cells of the SWOT itself. The question to ask PESTEL is just one: "What could change in the next 24 months in my target market among the six variables — political, economic, social, technological, environmental, and legal — in such a way as to flip my opportunities into threats or vice versa?" Answer in half a page. More than that is theory, and theory is paid for in time.
Porter's Five Forces for export makes sense only for structured companies entering a market with an investment above a certain threshold — say above 100k€ of commitment. Below that, it's an academic exercise that produces slides and not choices.
Competitive benchmarking, on the other hand, is underestimated, and it's where AI today really changes the cards. Building in two hours a competitive profile of five competing players in a foreign market — positioning, public prices, partnerships, communication, physical presence — is today within reach of anyone who knows how to query a generative system well. That profile, intersected with the SWOT, is worth ten times what the SWOT is worth on its own.
When the SWOT shouldn't be done
There's a case, and it should be said because no one says it. The SWOT shouldn't be done — or should be done in a minimal version, two cells at most — when the company hasn't yet decided whether it really wants to export. When it's still in the "we're thinking about whether it makes sense" phase, a complete SWOT becomes an alibi for not deciding. Weeks of analysis are produced, sheets are filled in, meetings are held — and the real decision is postponed, which is a decision of risk and ambition, not of analysis.
The first tool of export isn't the SWOT. It's an honest conversation between those who own and those who lead the company on this question: are we willing to lose calmly for two or three years on a new market, knowing that learning is the first product and revenue the second? If the answer is yes, then the SWOT serves to choose well which market to lose on. If the answer is no — or it's a lukewarm yes conditioned by "let's see how it goes" — no analysis tool will save the project.
The companies that really export have made peace with that first question before picking up any framework. The others confuse the analysis with the decision, and end up doing both badly.