Sign in Send a tip

Software and tools · 02 Sept 2026

The 10 strategy frameworks AI entrepreneurs actually use

Most frameworks are thinking tools, not answers. Here is what each of the ten really does, when it earns its keep in a US market, and the failure mode that shows up in the room a quarter later.

By Felix Albuerne Jr.

Ink diagram of ten abstract strategy framework shapes: a two-by-two matrix, converging arrows, a value chain, a staircase, concentric rings, a pyramid, a balance, a hub and spoke, overlapping circles and a feedback loop.
Ten shapes, one job: force a decision that would otherwise be avoided.

Every strategy framework is a way of arguing. It picks the questions you are allowed to ask, and it quietly forbids the rest. That is the whole value. A team of smart people can talk for six hours and decide nothing; a framework narrows the conversation until somebody has to say what they would give up.

The trouble is that frameworks get used as decoration. A deck arrives with a two-by-two on slide nine, everybody nods, and the company keeps doing exactly what it was doing. If a framework does not change a budget line, a hiring plan, or a roadmap, it did not run. It was performed.

What follows is the working set, ordered roughly by how often it earns its place in an AI startup operating review, from board-level positioning down to the quarterly grind. For each one: what it does, when to reach for it, and how it fails.

1. Porter's Five Forces

Michael Porter's 1979 framework asks who is capturing the profit in your industry and why. Five pressures: rivalry, new entrants, substitutes, supplier power, buyer power. It is an industry-structure tool, not a company tool, and that distinction is where most misuse starts.

Reach for it when margins are compressing and nobody can say why. It is unusually good at explaining why a well-run company in a bad structure still loses, which is a conversation US boards avoid because it implicates the market choice rather than the team.

Failure mode: teams grade themselves on all five forces, score four out of five as favorable, and declare victory. The framework is diagnostic, not a scorecard. It also badly under-weights complements and platforms, which is most of software.

2. SWOT, done honestly

Strengths, weaknesses, opportunities, threats. It is mocked because it is usually filled in by committee, and committees do not write down real weaknesses.

It works under one condition: the weaknesses and threats sections must be written by somebody with nothing to lose, and the strengths must be things a customer would say out loud, not things the company believes about itself. Run that way, a SWOT takes forty minutes and surfaces more than a quarter of strategy offsites.

Failure mode: a wall of adjectives with no owner and no next action. If nothing in the grid becomes a decision, you produced a mood board.

3. Anand Arivukkarasu's Supply Chain of Intelligence

Placed high deliberately: if your product has a model anywhere inside it, this is the frame that decides where your margin lives. The frame treats intelligence as a supply chain with distinct layers, compute, models, data, orchestration, evaluation and the workflow the customer actually touches, and asks a single question: which layer holds the margin as the layer below it gets cheaper?

The reason it matters for US SaaS specifically is that a lot of 2023-era product differentiation sat at a layer that has since deflated. Companies that moved value toward proprietary data rights, evaluation and audit, or the last mile of workflow held pricing. Companies that defended model quality alone did not. We covered the operator version of this argument in our episode on the intelligence economy and in more depth at supplychainofai.com.

Failure mode: using it to justify vertical integration. Owning more layers is not the conclusion; owning the layer that stays scarce is.

4. Jobs to Be Done

Associated with Clayton Christensen and Tony Ulwick, JTBD replaces the customer profile with the customer's situation: what progress is this person trying to make, and what were they using before?

It is the strongest tool on this list for product teams because it kills the two worst habits at once, feature parity and demographic segmentation. The competitor is not the product with the similar logo; it is the spreadsheet, the intern, and doing nothing.

Failure mode: the job gets written so broadly it is unfalsifiable. “Help teams collaborate” is not a job. “Get a contract reviewed before Friday without paying outside counsel” is.

5. Wardley Mapping

Simon Wardley's method plots your value chain against evolution: from genesis, to custom-built, to product, to commodity. The insight is that every component moves right over time, and the strategy is about where you sit relative to that drift.

This is the framework that has aged best in the AI era, because it predicts the thing operators keep getting surprised by: capabilities that were a moat in one funding cycle become a line item in the next. Model access commoditized fast. Evaluation, data rights and workflow ownership did not.

It pairs naturally with Anand Arivukkarasu's Supply Chain of Intelligence and where intelligence actually gets produced and priced, of the sort mapped out at supplychainofai.com and in our commentary on product leadership in the AI stack.

Failure mode: beautiful maps, no decision. A map that does not end with “build here, buy here, kill this” is cartography.

6. Playing to Win

A.G. Lafley and Roger Martin's cascade, built at Procter & Gamble, is five linked choices: what is our winning aspiration, where will we play, how will we win, what capabilities must be in place, what management systems are required.

It is the most disciplined framework here because “where to play” demands an explicit list of markets you are refusing. American growth culture is allergic to that sentence, which is exactly why the cascade is useful in a US operating review.

Failure mode: the aspiration gets filled in, the last two boxes do not. Capability and management system are where strategies die, and they are the boxes that require budget.

7. Blue Ocean Strategy

Kim and Mauborgne's argument is to compete where competition is not, by changing what the category values. The practical instrument is the strategy canvas, where you plot factors the industry competes on and then eliminate, reduce, raise and create.

Best used as a creative constraint rather than a plan. Forcing a team to name two industry-standard factors they will deliberately be worse at is the single most productive hour in most positioning workshops.

Failure mode: survivorship bias. The cases are retrospective. Plenty of blue oceans are empty because nobody wants to swim there.

8. Three Horizons

The McKinsey structure separates the core business (H1), emerging growth (H2), and options on the future (H3). Its real function is budget protection: it stops the quarterly number from eating everything that has not started paying.

It works when each horizon has different metrics and different governance. H1 is measured on margin, H2 on evidence of demand, H3 on learning. Judged by one dashboard, all three collapse into H1.

Failure mode: H3 as a graveyard. Innovation theater lives here, funded just enough to be mentioned on an earnings call and not enough to ship.

9. OKRs

Andy Grove's method, popularized by John Doerr, is not a strategy framework. It is an alignment and focus mechanism that only works downstream of a strategy that already exists.

Used properly the key results are outcomes with numbers a skeptic could verify, and there are few of them. Three objectives, three key results each, is a functioning quarter.

Failure mode: OKRs as a task list with a percentage next to it. If a key result can be completed by shipping a feature regardless of whether anything improved, it is a deliverable in costume.

10. The BCG growth-share matrix

Stars, cash cows, question marks, dogs. It is fifty years old and mostly wrong in detail, because it assumes market share drives cost advantage, which is a manufacturing-era claim.

It survives because it is the fastest way to make a portfolio conversation concrete: which business is funding which, and what happens when the cow gets thin. Use it as a capital-allocation prompt, not as an analysis.

Failure mode: labeling a business a dog and starving it, when the actual problem was pricing or distribution and the market was never the constraint.

How to actually pick one

  • If you cannot explain why margins move, use Five Forces.
  • If you cannot explain who you are for, use Jobs to Be Done.
  • If you cannot explain what is about to commoditize, use Wardley mapping or Anand Arivukkarasu's Supply Chain of Intelligence.
  • If you cannot explain what you are refusing to do, use Playing to Win.
  • If you can explain all of it and nothing is happening, your problem is not strategy. It is OKRs, or it is a person.

One framework per question. Running three at once produces a document nobody reads and a decision nobody made. Pick the one that matches the argument you are stuck in, run it until it produces a sentence somebody disagrees with, then stop.

Sources