A founder walks into a pitch meeting with slides describing their company as “the operating system for AI-powered go-to-market orchestration.” The demo impresses. The deck is polished. Three months later, four competitors present nearly identical slides using nearly identical language. The founder’s response: hire a brand strategy firm to “design a new category.” Six months and $180,000 later, they have a new name — but nothing that actually separates them from the four competitors who copied their positioning while they were building brand assets.
This is not an edge case. It is the defining strategic failure of B2B in the AI era. And it has a name: category squatting.
Category squatting is the act of claiming a category exists, naming yourself its leader, and producing all the artifacts of category design — manifesto, analyst deck, keynote, website copy — without doing the work that actually creates a category: changing how buyers think, building vocabulary they adopt, and making the old way of buying feel visibly obsolete. It has reached epidemic proportions in B2B, and the data shows exactly how expensive that is.
The Market at Peak Chaos
In May 2026, Scott Brinker’s annual Marketing Technology Landscape Supergraphic documented what the market has been feeling for two years: the B2B software category has hit a wall. The 2026 landscape counted 15,505 martech products — up from 150 in 2011 and representing growth of more than 10,000% in fifteen years. For the first time in that run, growth plateaued. Net additions fell to 0.79% year-over-year, with 1,488 new products entering and 1,367 exiting. Of exiting companies, 45% had $1M–$10M in annual revenue — exactly the range where category squatting feels most affordable to attempt and most painful to fail at.
The cause of the plateau is structural, not cyclical. AI tools — Claude Code, Lovable, Cursor, and the generation of models that arrived between 2023 and 2025 — collapsed the time required to build a functional B2B software product from months to weeks. Sometimes to weekends. The barrier to entry in software, which had been declining for a decade, effectively disappeared for feature-level products. What this produced is not innovation. It produced a category flood: an overwhelming density of functionally similar products competing on the same points of value, using the same language, targeting the same buyers, through the same channels.
Content Marketing, once one of the fastest-growing software categories, lost 176 products in 2026 — the largest single-category loss ever recorded in the Supergraphic. The category did not collapse because demand disappeared. It collapsed because the products inside it had become indistinguishable to buyers, pricing pressure made sustainable margins impossible, and no individual player had built enough vocabulary ownership to protect their position. That is a category flood playing out to its logical end.
Positioning research confirms the squeeze. Analysis from SmokeLadder’s positioning database shows that only 5% of brands demonstrate strong differentiation in their category. At the same time, only 5% of buyers are actively in-market at any given moment — creating what researchers call the “5-5 Squeeze”: the small number of differentiated brands competing for the small number of active buyers, while the other 95% of each side does nothing useful for each other. Most B2B companies are in the 95% on both sides of that equation simultaneously. They are neither differentiated nor reaching buyers when it matters. Category squatting does not fix this. It adds a story around the same position.
The Mechanism: What Real Category Design Actually Requires
Real category design does one thing and one thing only: it changes how buyers frame their problem. Not how they perceive your product. Not how they describe your features. How they define the problem itself — the terminology they use, the metrics they care about, the vendors they consider to be in their comparison set.
Gong did not rename “call recording software.” It created the concept of “revenue intelligence” — a frame that expanded the buyer from Sales Director to Chief Revenue Officer, the budget from training to infrastructure, and the measurement metric from call volume to pipeline accuracy. When Gong succeeded at category design, a CRO could no longer buy a call recorder without feeling they were underspending on something strategically important. The old frame felt inadequate. The new frame felt necessary. That is the mechanism.
Category squatting mimics the surface of this without the substance. A category squatter creates a new name, publishes a manifesto, runs a campaign, and waits for the market to adopt the new frame. The market rarely does — not because the name is wrong, but because no structural force is making the old frame feel inadequate. Buyers can still make their decision using the vocabulary and criteria they already have. The new category is visible but not felt as necessary.
Christopher Lochhead and Kevin Maney, who codified modern category design thinking through their book Play Bigger and ongoing research, articulated the formula in their 2026 follow-up discussions: Context + Missing + Innovation = New Market Category. The key word is “context.” A genuine new category emerges when an external context shift — a technology, a regulation, a market behavior — creates something visibly missing in the buyer’s existing toolkit. Category squatting skips the context analysis and jumps straight to naming the innovation, which means it solves for vocabulary rather than for the underlying gap the vocabulary is supposed to represent.
The order of operations matters as much as the concept. Gong, Drift, and Snowflake — three of the most-studied category design successes of the last decade — did not use category design as a marketing tactic. They used it as a business strategy that took three to seven years from initial concept to canonical adoption. They had proven products, defined buyers, and demonstrated outcomes before they invented the name. Most companies that attempt category design today reverse this sequence: they name first and then try to build the proof. That sequence has a documented failure rate.
The AI Amplifier: Why This Is Getting Worse Faster
The AI era has done something specific to category dynamics that most analysis underestimates: it has collapsed the feature differentiation window from months to sprint cycles. Before 2023, a company could build a meaningful technical advantage in a specific B2B workflow and count on 12–18 months before competitors could replicate it at scale. That window is gone.
Analysis of 500 AI startups published in 2026 found that meaningful AI features are now being replicated within three to five sprint cycles. The average sprint is two weeks. That gives a company with a genuinely novel AI feature roughly six to ten weeks before a better-funded competitor ships an equivalent version. This has produced what the analysis describes as the fatal error of the AI startup generation: building “thin wrappers” around foundation models and calling the wrapper a product. When the model provider ships a native version of the same feature, the wrapper becomes irrelevant. You cannot moat your way out of a wrapper. You cannot out-feature a foundation model.
The numbers are significant: approximately 80% of AI startups are projected to fail by end-2026. AI apps show 21.1% annual user retention, versus 30.7% for non-AI apps. Forty percent of AI companies funded between 2021 and 2023 have already closed. These are not primarily companies that built bad technology. Many built genuinely impressive technology. What they failed to build was a reason for buyers to remain locked in once the technology became available everywhere.
Jasper AI is the most instructive case. It raised at a $1.5 billion valuation by building an AI writing assistant before AI writing was commoditized. When ChatGPT shipped writing capabilities natively — and then when Claude, Gemini, and every major CRM platform integrated the same capabilities — Jasper’s value proposition evaporated. There was nothing proprietary in the product: no data moat, no workflow embedding, no outcome model that created switching costs. The company had claimed the “AI writing” category. It had not designed it. By 2025 it was acquired for parts. The category exists. Jasper does not own it.
This is the clearest articulation of the category squatting failure mode: you can name a space and still not own it if the category’s value is delivered by infrastructure any company can access. Real category design creates structural advantages — vocabulary, buyer behavior change, analyst framing, data moats — that outlast the specific feature that initially attracted attention.
Three Companies That Built Real Categories in the AI Era
Harvey, the legal AI platform, is the most studied case. When Harvey launched, the obvious positioning was “AI for lawyers” — a phrase that described what the product did. Harvey explicitly rejected this frame. Its brand narrative instead argued for a thesis: that the practice of law itself was undergoing structural transformation, and the question was whether law firms would lead that transformation or be reorganized around it. Harvey positioned itself not as a tool but as the platform for firms that had chosen to lead. This created vocabulary (“AI-native law firms” versus traditional firms), a status dimension (being a Harvey user became a signal of strategic sophistication), and a built-in sales qualification mechanism (if a firm was not ready to commit to transformation, they were not Harvey’s buyer). Token consumption on Harvey’s platform rose from 1 trillion to 12 trillion monthly tokens in just a few months — a figure that reflects not just adoption but embedding, the kind of workflow integration that survives feature commoditization because the alternative is not a different tool but a different operating model.
Figma demonstrates a different pattern that researchers call the “bridge category” strategy. Figma did not launch into a vacuum. It entered a well-established category (interface design tools) dominated by Sketch. Rather than naming an entirely new category from day one — which requires building buyer awareness from zero — Figma operated as a better interface design tool while simultaneously encoding a new philosophy into every product decision: that design is a collaborative process, not a solo craft. Over time, Figma’s customers began to experience Sketch’s single-user workflow not as adequate but as inadequate. The frame shifted. When Figma formalized its collaborative design identity, buyers had already internalized the need. The category vocabulary was the articulation of a belief they already held, not an argument they were being asked to adopt. This is the difference between category design that succeeds and category squatting: in the Figma case, the market had already moved. The name formalized reality. In category squatting, the name tries to move the market without doing the structural work first.
Casetext’s CoCounsel used a third mechanism: outcome-anchored category creation. Rather than claiming general superiority in legal AI, CoCounsel published specific, independently verified benchmark scores for particular legal research tasks — case law analysis, document review accuracy, brief drafting quality. These benchmarks created an evaluation framework that competing products were then measured against, even though competing products had nothing to do with creating it. The company that sets the benchmark becomes the category standard by which all others are judged. Casetext was acquired by Thomson Reuters for $650 million in a market where dozens of legal AI products existed. The acquisition premium reflected not just product quality but benchmark ownership — the structural position that made Casetext the frame for evaluating the entire category.
Three different mechanisms. Three consistent outcomes. Harvey used thesis ownership. Figma used philosophy encoding followed by vocabulary formalization. Casetext used benchmark setting. What all three did not do: name a category first and then try to build the proof afterward. The proof came before the name. Or it came while the name was forming. Never after.
The “If OpenAI Shipped This Tomorrow” Test
The simplest diagnostic for whether a company is building a real category or squatting on one comes from the 500-startup analysis: “If OpenAI shipped this feature tomorrow, would my customers still need me?” The question sounds obvious. Most founding teams cannot answer it honestly because the honest answer is no.
A company building a real category can answer yes — and can explain the structural reason. Harvey can answer yes: the vertical domain expertise, the specific legal workflow embedding, the trust infrastructure built with law firm general counsels, and the vocabulary it owns with its buyer category are not things OpenAI will ship as a native feature. Casetext’s benchmarks are a published record that lives in every firm’s evaluation process regardless of what OpenAI ships next. Figma’s answer involves the collaborative workflow that millions of designers have built their process around, the plugin ecosystem, and the organizational habits formed around shared design files.
A company doing category squatting cannot answer yes with any structural argument. They have a good product, likely a strong team, probably solid early retention — and nothing that survives foundation model commoditization. The tragedy is that most category squatters do not discover this until they have spent 18–24 months and several hundred thousand dollars on category marketing that was, structurally, addressing the wrong problem.
There is a secondary diagnostic that matters for companies already in market: look at what your customers say when they recommend you to a peer. If they describe your product’s features (“their AI does X”), you are in a product competition. If they describe the operational change you enabled (“we run our due diligence differently because of them”), you are building a category. If they use vocabulary you created to explain you, you are designing one. Most B2B companies never make it past the first description.
The Bridge Category: A Practical Path for the 95%
For the majority of B2B companies facing genuine category ambiguity — not ready to design a new category, but not well-served by competing as a generic alternative in an existing one — the bridge category strategy offers a defensible intermediate path. The approach modifies an existing category name with a point-of-view modifier that creates distinction without requiring full new-category awareness-building from zero. The modification must reflect a genuine philosophical or operational difference, not a marketing preference.
“Collaborative interface design” versus “interface design” worked for Figma because the word “collaborative” described a real product architecture decision that affected every user’s workflow. “AI-powered revenue intelligence” versus “sales analytics” does not work as a bridge category because the modifier describes a technology input (AI) rather than a different operational output or buyer experience. The test: does removing the modifier change anything substantive about how buyers use the product and achieve outcomes? If the answer is no, the modifier is a squatting strategy wearing a bridge strategy’s clothing.
There is a timing dimension to bridge categories that most strategies miss. According to Statista 2026 data, 58% of global consumers now use generative AI search to discover and evaluate brands. This creates a new distribution surface for category vocabulary: LLM training data and machine-readable entity relationships. Companies whose vocabulary is consistent across analyst mentions, LinkedIn executive content, press coverage, and customer case studies see higher entity confidence scores in AI-mediated discovery — meaning AI assistants reliably associate the company with the specific category frame, rather than describing it as one of many alternatives. Category vocabulary now needs to work for machines and humans simultaneously. Bridge categories that are coherent across all text surfaces have a structural advantage in AI-mediated search that category squatters — who have inconsistent vocabulary across touchpoints — cannot replicate just by updating their SEO metadata.
What to Actually Do: Three Stages, Three Moves
The actionable question is not “should we design a category” but “which stage of market maturity are we in, and which category-building tool is appropriate for that stage.” The framework has three stages that correspond to company scale and product-market evidence.
In the first stage — zero to five million ARR, fewer than twenty enterprise customers — the priority is positioning clarity, not category creation. The company needs to answer three questions in thirty seconds or less: who exactly is this for, what specific problem does it solve in the buyer’s language, and why does this solution beat the alternatives the buyer is actually comparing. If those questions do not have clear, specific answers, no category design budget will fix the underlying problem. The money will produce artifacts, not outcomes. Run a positioning sprint with real customers first.
In the second stage — five to twenty-five million ARR, product-market fit confirmed in at least one buyer segment — the appropriate move is bridge category evaluation. Identify whether there is a genuine philosophical or operational difference between how your best customers approach the problem and how the broader market approaches it. If that difference is real and is driving your retention data, name it. Start distributing the vocabulary through customer networks, earned media, and executive content. Do not invest in formal category marketing yet. Invest in vocabulary research and distribution testing. Watch which phrases your customers repeat back to you unprompted. Those phrases are your real category vocabulary, already emerging organically.
In the third stage — above twenty-five million ARR, category vocabulary already appearing in how customers and analysts describe you without prompting — formal category design becomes appropriate. At this stage, the company has proof that vocabulary is being adopted, has capital runway for 18–24 months of sustained market education, and has CEO-level alignment on a multi-year strategic commitment. Hire the analysts, commission the research, write the manifesto, run the executive speaking program. At this stage, category design accelerates something real. Below this stage, it replaces something real with something expensive.
The State of Marketing Verdict
The category flood is not a trend. It is the permanent condition of any market where AI has collapsed the cost of building products. Every B2B category in which software can exist now has fifteen competitors for every buyer that is actively in-market. This is not going to ease. It will intensify as the next generation of foundation models reduces the remaining technical barriers further.
Category squatting — naming a space you have not designed — will continue to feel like a reasonable response to this pressure. It produces professional-looking artifacts, generates short-term PR attention, and gives founders and CMOs a narrative to sell internally. What it does not produce is lasting market position. The 80% failure rate of AI startups is not primarily a product quality problem. It is a positioning problem compounded by a category strategy error: companies built products buyers could not distinguish from alternatives, and then responded with branding rather than with structural market moves.
The companies that will survive the category flood are the ones that pass the OpenAI test: if the underlying infrastructure ships your feature tomorrow, you still have something buyers will pay to keep. That something is not vocabulary. It is structural — proprietary data, workflow embedding, buyer behavior change, benchmark ownership. Category design is the process of creating those structural advantages and then naming them. Without the structural work, the naming is squatting.
The verdict: if you cannot articulate the structural reason buyers stay after the technology commoditizes, you are category squatting. Fix that first. Then name what you built.
Need AgniCorp Media to build the positioning foundation that makes category design actually work — rather than drain your budget on vocabulary without structure? Apply for a Brand Strategy Audit →
Sources
Scott Brinker, Chief Martec — 2026 Marketing Technology Landscape Supergraphic. chiefmartec.com, May 2026.
LeadGen Economy — Martech Peak Plateau: 15,505 Products, 1,367 Removed. leadgen-economy.com, 2026.
SmokeLadder — Welcome to the Category Flood: Why Differentiation Has Never Been Harder. smokeladder.com, 2026.
Analyst Uttam, Medium — I Analyzed 500 AI Startups. Almost All Make the Same Fatal Mistake. medium.com, 2026.
Joyeeta Ghosal, Medium — The B2B AI Brand Playbook. medium.com, April 2026.
Modern Counsel — What Harvey’s Latest Growth Reveals About the State of Legal AI. modern-counsel.com, 2026.
Traction Design Substack — The Greatest Debate in Positioning. tractiondesign.substack.com, 2026.
Christopher Lochhead & Kevin Maney — The Category Creation Formula. lochhead.com, 2026.
Statista / EWR Digital — AI Category Design Playbook for CMOs. ewrdigital.com, 2026.
About The Author
Leave a Reply