The Intuit Line
Intuit fired 3,000 people and signed AI deals the same day. The question is which side of the cut you are on.
Raj Patel had not planned to spend his Tuesday evening performing an autopsy on his own career. He is a product manager at a mid-size vertical SaaS company in Denver that sells scheduling software to dental practices, the kind of company with 400 employees and a Slack channel called #wins where people post screenshots of closed deals. He had opened his laptop after dinner to catch up on email and instead found himself reading the Intuit headline for the third time: 3,000 people, 17 percent of the workforce, gone in a single announcement.
The number was large enough to be abstract. What made it personal was the paragraph underneath. Intuit had signed multi-year partnerships with Anthropic and OpenAI the same day, embedding their AI models across TurboTax, QuickBooks, Credit Karma, and Mailchimp. The company’s CEO, Sasan Goodarzi, told CNBC the cuts had nothing to do with AI. They were about reducing complexity, flattening management layers, eliminating coordination-heavy roles. Raj read that sentence twice. He managed a team of four. His title contained the word “coordination” in its job description. His coffee was getting cold on the desk beside him and the apartment had the particular quiet of a place where the other person has already gone to bed.
He did something that would have seemed strange a year ago and feels almost reflexive now. He opened Claude, pasted the Intuit press release, and typed: “I am a product manager at a 400-person vertical SaaS company. My job involves writing product requirements, prioritising features based on customer feedback, coordinating between engineering and sales, and maintaining our product roadmap. Based on what Intuit just did, how vulnerable is my role to being replaced or restructured? Be specific and honest.” What came back was not reassuring. It was not alarming either. It was a question disguised as an answer, and it is the question this article is about.
The AI told Raj that his role contained two distinct types of work. The first type was intelligence: synthesising customer feedback into feature priorities, interpreting usage data, writing specifications that translated business needs into engineering requirements. This was pattern recognition and domain knowledge, the kind of work that frontier AI models are increasingly capable of performing. The second type was interface: the relationships with the dental practice owners who called him by name, the Thursday standups where he read the room to know which engineer was frustrated and which feature was about to slip, the regulatory nuances of healthcare scheduling that lived in his head and nowhere else. The AI could not do that work. Not yet.
Raj stared at the screen. He had never thought about his job in those terms. He had always assumed the valuable part was the intelligence, the ability to look at a usage dashboard and know which metric mattered. The AI was suggesting the opposite. The valuable part, the part that would survive, was the stuff he thought of as overhead: the meetings, the relationships, the hallway conversations where someone mentioned a compliance requirement that was not in any document.
This is the threshold that Intuit crossed on May 20, and it has a shape that every knowledge worker can measure against their own role. Call it the Intuit Line. Below it, your work is primarily intelligence: pattern recognition, domain expertise, content generation, analytical reasoning. Above it, your work is primarily interface: relationships, regulatory navigation, institutional memory, cross-functional coordination that requires reading human signals no API can parse. The 3,000 people Intuit cut were, by the company’s own restructuring logic, below the line. The people who survived are above it. The question the rest of the economy will spend the next twelve months answering is where, exactly, the line falls for everyone else.
Intuit’s restructuring revealed something that most corporate layoff announcements carefully obscure. The company did not simply reduce headcount. It replaced an entire layer of its architecture. For forty years, TurboTax’s competitive advantage was its tax engine, the accumulated logic of thousands of tax code interpretations, edge cases, and state-by-state filing rules maintained by teams of tax knowledge engineers. That engine was the intelligence layer. It was the reason TurboTax could charge $203 for something that is technically free to do by hand. By contracting with Anthropic and OpenAI, Intuit declared that frontier AI models can replicate that intelligence more cheaply than the humans who built it. The restructuring charge alone, $300 to $340 million, suggests the company expects the savings to dwarf the cost within two years.
Goodarzi insists the cuts target coordination roles, project managers and business operations staff, not the AI-adjacent technical functions. But the timing tells a different story. On the same earnings call where the layoffs were announced, Intuit disclosed that TurboTax Live revenue grew 36 percent to $2.8 billion, now representing 53 percent of total TurboTax revenue. The product growing fastest is the one that connects a human tax expert directly to a customer through a screen. The interface layer is expanding. The intelligence layer is being outsourced. The coordination layer, the connective tissue that held the old architecture together, is the first casualty of the new one.
The three-layer model that Intuit has quietly built looks like this. At the bottom sits intelligence: the analytical engine, the domain expertise, the pattern recognition that used to require thousands of specialists. This layer is now rented from AI labs. Anthropic’s Claude and OpenAI’s models are embedded across the platform through multi-year contracts. In the middle sits the interface: TurboTax’s user experience, the live expert connections, the customer relationships, the sales teams who understand which dental practice owner in Scottsdale needs a phone call and which one prefers email. This layer is owned by the company and is where the surviving employees sit. At the top sits the moat: IRS e-filing partnerships, state-by-state regulatory agreements, SOC 2 certifications, the compliance infrastructure that took decades to build and that no AI lab can replicate by training a model. This layer is defended, and it is the reason Intuit can rent intelligence from Anthropic without worrying that Anthropic will become a competitor. Claude can interpret tax code. Claude cannot file your taxes with the IRS.
Raj did not know any of this when he sat down on Tuesday night. What he knew was that his company, like Intuit, was a vertical SaaS business with a product built on domain expertise. The dental scheduling software his team maintained was not as complex as TurboTax, but its value proposition was structurally identical: we understand your industry’s rules better than you do, and our software encodes that understanding so you do not have to. If Intuit could outsource forty years of tax knowledge to an AI lab, what was stopping his company’s board from asking the same question about dental scheduling regulations?
He ran the diagnostic. It took twenty-three minutes, not thirty, because the AI was faster than he expected. Step one: he listed every output he had produced in the past month. Product requirements documents. Feature prioritisation matrices. A competitive analysis. Two customer feedback summaries. A quarterly roadmap presentation. Step two: he fed five of those documents into Claude and asked it to produce equivalents using only the raw data inputs he would normally start from. The AI produced a competitive analysis that was, he admitted to himself, about 80 percent as good as his. It missed the nuance that one competitor had quietly acquired a compliance consulting firm, something he had learned from a conversation at a trade show that existed in no public document. But the structure, the data synthesis, the recommendations were competent. Step three: he listed his interface work. The Thursday standups. The quarterly business reviews where he presented to dental practice owners who had been customers for six years and called him when they had a problem. The compliance calls with the state dental board in Colorado. The hallway conversations with the CTO about technical debt that shaped architecture decisions no document captured.
Step four was the calculation. Raj estimated that 55 percent of his weekly hours went to intelligence work: the documents, the analysis, the synthesis. Forty-five percent went to interface work: the meetings, the relationships, the calls. He was close to the line. Not safely above it. Not clearly below it. Sitting on the threshold where the answer depends on whether your company’s board reads the Intuit earnings call this quarter or next.
The turn in this story is not that the AI could do Raj’s intelligence work. He expected that. The turn is that when he looked at the 45 percent of his role that was interface, the part he thought was safe, he realized how much of it was already eroding. The Thursday standup had moved to Slack two months ago. Three of his six key customer contacts had started using the product’s built-in support chatbot instead of calling him. The compliance calls with the state dental board had become quarterly instead of monthly after the board published its updated guidelines as a searchable database. The interface layer that was supposed to protect him was thinning, not because anyone decided to thin it, but because every tool that made his life easier also made his presence less necessary.
This is the part of the Intuit story that the comfortable reading misses. The comfortable reading says the intelligence layer is vulnerable and the interface layer is safe. The uncomfortable truth is that the interface layer is safe today the way a second-floor apartment is safe in a flood: the water has not reached you yet, but it is rising. Intuit’s TurboTax Live product is instructive. It connects a human expert to a customer through a screen. The human is the interface. But Intuit’s Anthropic partnership now allows Claude users to access TurboTax’s capabilities directly inside the chatbot, getting real-time tax estimates and refund projections without ever touching the TurboTax interface or speaking to a human expert. The product that protects the interface layer is also the product that will eventually bypass it.
The sectors most likely to follow Intuit’s template in the next twelve months share three characteristics. They sell software built on encoded domain expertise. They face margin pressure from AI-native competitors. And they have regulatory moats deep enough to prevent the AI labs from going direct to consumer. Legal technology companies like those maintaining case law databases sit on decades of legal reasoning that frontier models can now replicate at inference cost. Healthcare IT companies running clinical decision support systems encode medical knowledge that the same models are trained on. Financial advisory platforms built on analytical intelligence face the same arithmetic Intuit faced: why maintain a thousand analysts when the model knows the same formulas?
The speed of the cascade depends on one number that Wall Street will publish for Intuit in July: the gross margin improvement from the restructuring. If Intuit’s operating margin expands by three or more percentage points in the quarter ending July 31, every vertical SaaS CFO in America will have a slide in their next board presentation titled “The Intuit Playbook.” If the margin does not improve, if AI vendor costs eat the headcount savings, the template stalls.
What happens next depends on which of three forces prevails. The most probable outcome, and Scenarica puts it at four in ten, is that the template works but the economics shift. You are a product manager at a company that follows Intuit’s lead. Your intelligence layer is outsourced to an AI lab. Your costs drop. For eighteen months, maybe two years, the savings are real. Then the AI lab raises prices. Not dramatically, not all at once, but steadily, the way any vendor does once the switching cost is too high. Your company discovers it has traded a $27 billion payroll line for an AI vendor bill that grows 15 percent a year. You still have a job, but your company is now dependent on two or three AI providers the way it used to be dependent on its own engineers. A different dependency, not independence.
Three in ten odds on the second path: the template cascades fast and wide. Intuit’s margin improvement is dramatic enough that the playbook spreads across legal tech, healthcare IT, and fintech within a year. If you work in knowledge engineering, content development, or analytical roles at a vertical SaaS company, the next twelve months look like Intuit’s May 20. Fifty thousand or more knowledge workers across the sector face the same restructuring. AI lab revenues surge as enterprise contracts multiply. The Intuit Line becomes not a diagnostic but a verdict.
The remaining three in ten go to the scenario nobody at Intuit is modelling publicly. AI labs cannot replicate domain-specific accuracy at production quality. Frontier models make too many tax errors, miss too many regulatory edge cases, produce confident answers that are subtly wrong in ways that create liability. Intuit discovers this six months into the partnership, quietly rehires specialists under different titles, and the restructuring becomes a cautionary tale rather than a template. If you are below the Intuit Line and this scenario materialises, your expertise becomes more valuable, not less, because the market has just proven that the machine cannot do what you do.
What shifts the probabilities is measurable. Watch Intuit’s Q4 earnings in late August for the restructuring’s first margin impact. If operating margin expands more than 200 basis points, the cascade scenario gains probability. Watch AI error rates in regulated domains. The IRS processes roughly 164 million individual returns per year. If AI-assisted TurboTax filings generate a statistically significant increase in amended returns or audit flags by January 2027, the reversal scenario gains probability. And watch AI lab pricing. Anthropic and OpenAI are competing aggressively on enterprise pricing today. If either raises prices for enterprise API access by more than 20 percent before the end of 2026, the vendor dependency scenario becomes the headline.
Watch the job postings at Intuit itself over the next ninety days. If the company begins hiring for roles with titles like “AI output reviewer” or “model quality analyst,” it means the intelligence layer is not as cleanly outsourceable as the restructuring assumed, and the Intuit Line is less of a bright boundary than a blurry zone.
Watch whether competitors follow. Sage, Xero, and Block all report earnings in the next sixty days. If any of them announces a similar restructuring with AI lab partnerships, the template is no longer one company’s experiment. It is an industry’s new operating model.
Raj closed his laptop at 11:47 PM. He had not written the email he sat down to write. He had, instead, done something that might matter more: measured himself against a threshold he had not known existed three hours earlier. Fifty-five percent intelligence, forty-five percent interface. He was close enough to the line that the answer depended on timing, on whether his company’s board read the Intuit playbook this quarter or next, on whether the AI lab pricing held or shifted, on whether the relationships he had built with dental practice owners in Colorado were truly irreplaceable or merely expensive. He opened a new browser tab, navigated to his company’s job postings page, and checked whether any new roles had appeared in the last forty-eight hours. None had. That was not reassurance. That was a clock that had not started ticking yet.
ANNEX: WHERE DO YOU FALL WHEN THE INTUIT LINE REACHES YOUR INDUSTRY?
Raj’s Tuesday night diagnostic is a dry run for a calculation that 113,000 tech workers laid off in 2026 never got to make in advance. Three scenarios define the next twelve months, and they sum to 100 percent of the probable outcomes. Your position relative to the Intuit Line determines which scenario matters most to your career.
Template works, vendor costs rise: 40%
You keep your job, or one like it, but the company you work for has traded one dependency for another. The intelligence layer is outsourced to AI labs. Costs drop initially. Then AI lab pricing rises as the market consolidates from five credible frontier providers to three. Your employer’s AI vendor bill grows 12 to 18 percent annually, eating the margin improvement that justified the restructuring. If you are in the interface layer, your role survives but your budget tightens. If you are in the intelligence layer, you were already cut. The new hires your company makes are not knowledge engineers. They are vendor managers and AI output reviewers, people whose job is to check the machine’s work, a role that did not exist eighteen months ago.
The variable to watch is AI lab enterprise pricing. Anthropic’s Claude Enterprise seat pricing starts at roughly $20 per user per month with API usage billed separately, keeping headline costs competitive. If Anthropic or OpenAI raises enterprise API pricing by more than 20 percent in any single quarter, the vendor dependency scenario becomes dominant. Probability of a 20-percent-plus price increase: 15 percent within 3 months, 35 percent within 6 months, 55 percent within 12 months.
Template cascades across vertical SaaS: 30%
You are a knowledge engineer, a content strategist, a domain analyst at a vertical SaaS company. Your company’s board has seen Intuit’s Q4 margin numbers and they are impressive. The CFO has a slide. The restructuring is announced in your all-hands meeting on a Tuesday, the way these things always are. Fifty thousand or more knowledge workers across legal tech, healthcare IT, fintech, and HR tech face the same cut within twelve months of Intuit’s announcement. AI lab revenues from enterprise contracts double. The Intuit Line becomes an industry standard that consultancies package into frameworks and sell to boards as transformation roadmaps. If you are below the line, the window to move above it is twelve months, maybe less.
The variable to watch is Intuit’s Q4 operating margin, reported in late August 2026. If operating margin expands by 300 basis points or more relative to Q3, the cascade accelerates. Probability of 300-plus basis point expansion: 20 percent in the Q4 report (August 2026), 40 percent by Q2 FY2027 (February 2027), 55 percent by end of calendar 2027.
AI accuracy fails in regulated domains: 30%
The intelligence layer turns out to be harder to outsource than Intuit assumed. Frontier models produce tax interpretations that are 95 percent accurate, which sounds impressive until you calculate what 5 percent error means across millions of filings. Amended returns spike. Audit flags increase. Intuit quietly rehires domain specialists under new titles and reduces its dependency on external AI. The template does not spread. If you are below the Intuit Line and this scenario materialises, your expertise appreciates. The market has just demonstrated that the machine cannot do what you do at production quality, and your salary reflects the proof.
The variable to watch is IRS error and audit flag rates for AI-assisted filings in the 2026 tax season, trackable through IRS filing season statistics published weekly. Amended returns represent a small fraction of roughly 164 million individual filings annually. If AI-assisted TurboTax filings generate a statistically significant increase in amended returns or IRS correspondence audits relative to human-prepared returns, the accuracy failure scenario gains probability fast. Probability of a measurable accuracy gap becoming public: 10 percent within 3 months, 25 percent within 6 months, 45 percent within 12 months.
Sources:
TechCrunch, “Intuit to lay off over 3,000 employees to refocus on AI,” May 20, 2026.
CNBC, “Intuit (INTU) Q3 earnings report 2026: Company cutting 17% of staff,” May 20, 2026.
CNBC, “Intuit CEO says company’s 17% workforce cut had ‘nothing to do with AI,’” May 20, 2026.
Intuit Investor Relations, “Intuit and Anthropic Partner to Bring Trusted Financial Intelligence and Custom AI Agents to Consumers and Businesses,” February 24, 2026.
Intuit Investor Relations, “Intuit Reports Strong Third-Quarter Results and Raises Full-Year Revenue Guidance,” May 20, 2026.
TechJournal, “113K Tech Layoffs in 2026 While AI Spending Hits $725B,” May 2026.
Intuit, “Intuit AI Principles: Strategic Partnerships with OpenAI and Anthropic,” 2026.
Disclaimer: This report is published by Scenarica Intelligence for informational purposes only. It does not constitute investment advice, a solicitation to buy or sell any financial instrument, or a recommendation regarding any particular investment strategy. Scenarica Intelligence is not a registered investment adviser or broker-dealer. All scenario probabilities and assessments represent the analytical judgment of Scenarica Intelligence and are subject to change without notice. Past performance of any asset or strategy discussed does not guarantee future results. Readers should conduct their own due diligence and consult with qualified financial advisers before making investment decisions.
Scenarica Premium: The full Scenarica suite includes Geopolitics, Economy, Bitcoin, AI, and Sunday Edition.
Scenarica Intelligence
We don’t predict the future. We price it.