In March, Sequoia Capital partner Julien Bek published a thesis called “Services: The New Software.” The argument: the next generation of big software companies will be ones masquerading as services firms.
The company he has in mind doesn’t sell tools to accountants. It closes the books. It doesn’t sell tools to lawyers. It reviews the contracts.
The economic logic is simple. For every dollar companies spend on software, they spend six dollars on services. The venture money has been chasing the software dollar. Bek’s argument is that the real prize is the six dollars.
Copilots versus autopilots
Bek divides AI companies into two categories: copilots and autopilots.
Copilots sell tools to professionals. Harvey helps lawyers draft contracts faster. Rogo helps investment bankers run analyses. The professional stays in the loop, and the tool makes them more productive.
Autopilots sell the work directly to the end customer. Crosby drafts NDAs for companies without involving a law firm. WithCoverage sells insurance to CFOs without a broker. Rillet closes books for companies with no accounting firm in the loop.
The competitive dynamics are asymmetric. If you sell a tool, every improvement in the underlying AI model is a competitive threat. Someone builds a better copilot, yours is obsolete. If you sell the work, every improvement in the model makes your service faster and cheaper to deliver. The autopilot gets stronger with every model upgrade. The copilot gets more replaceable.
Sequoia is telling founders to skip your firm and sell the work straight to your clients. It’s funding them to do it.
The map of your market
Bek published an opportunity map, and it reads like a target list: each vertical priced, and the startups already going after it named underneath. Find yours.
Management consulting: $300 to $400 billion. The largest line on the map, and the one Bek is least sure about. His note reads “Best candidates TBD.” Consulting is mostly judgment work, nobody has cracked it, and Sequoia is telling the room the prize is there anyway.
Recruitment and staffing: $200 billion-plus. Sourcing, screening, and matching are what Bek calls “pure intelligence work”: pattern-driven and high-volume, exactly what these companies take first. Mercor and Juicebox are funded to place candidates without a staffing firm in the transaction.
Insurance brokerage: $140 to $200 billion. Commercial lines are standardized, distribution is fragmented, and no incumbent controls the market. WithCoverage and Harper are selling directly to CFOs today.
Accounting and audit: $50 to $80 billion of outsourced work in the US alone. Rillet is building an AI-native ERP that closes the books without an accounting firm involved.
Legal transactional work: $20 to $25 billion. Contract drafting, NDAs, and regulatory filings all produce output a buyer can check, which is what makes skipping the firm feel safe. Crosby and Lawhive are funded and operating.
Add the lines up and you get most of a trillion dollars. Every one of those verticals is a firm type I talk to regularly.
The playbook they’re following
Bek’s strategy for these startups is specific. Start where outsourcing already exists. If a company already sends work to an outside firm, three things are true: they’ve accepted external delivery, a budget line exists that can be substituted, and the buyer is already purchasing an outcome rather than a tool.
The startups don’t need to convince companies to try something new. They need to offer the same outcome at a lower price with faster delivery. The fact that AI gets cheaper every quarter means their margins improve while yours compress.
Then the progression: start with the routine, pattern-driven work. Accumulate proprietary data about what good outcomes look like. Use that data to move into the judgment-heavy work that currently requires experienced professionals.
Today’s routine work is the beachhead. Tomorrow’s advisory work is the destination.
Incumbent advantages have a shelf life
These are funded companies with named founders and identified markets. WithCoverage raised capital to sell insurance directly to CFOs. Crosby raised capital to handle legal work without law firms. Mercor and Juicebox are placing candidates without staffing firms. Rillet is closing books without accounting firms.
None of them have the client relationships, the domain knowledge, or the trust an incumbent firm has built over years. But they don’t need those things yet. They’re starting with the commodity work: the standardized, high-volume tasks that make up a large share of what most firms bill for. By the time they move into the judgment work, they’ll have accumulated proprietary data about what good outcomes look like. Incumbents don’t have that data, because they never had to codify it.
The incumbent advantages are real, and they have a shelf life.
Building new offerings around domain expertise and client trust is what a firm can do while it still has both. Keep billing hourly for work an autopilot delivers at a fraction of the cost and you are the market these startups were funded to take.
The map is public
Sequoia told its portfolio companies that professional services revenue is the next big opportunity. The firm that backed Apple, Google, Stripe, and Nvidia put that in writing as a funding thesis, and the founders it funds are reading it as instructions.
Rillet is closing books today. Crosby is drafting NDAs today. They learn what good looks like from every engagement they run, and the incumbent’s version of that knowledge sits where it has always sat, in the heads of a few senior people who never had to write it down.