AI services startups and Sequoia's case for selling work

On March 5, 2026, Sequoia Capital's Julien Bek published "Services: The New Software." His opening line sets the thesis: "The next $1T company will be a software company masquerading as a services firm." The essay gives founders a framework for choosing what to sell and where to start.

Sell the work, not the tool

Bek starts with a fear most AI founders share: the next model release turns their product into a feature. His answer is to change what you sell. A company that sells a tool competes with the model. A company that sells the work gets better as the model improves. His example: a business may pay $10K a year for accounting software and $120K for an accountant to close the books. The opportunity is to close the books.

Intelligence versus judgement

Bek splits professional work into intelligence and judgement. Intelligence is rule-following, even when the rules are complex. Judgement relies on experience and taste. He argues that software engineering moved first because most of it is intelligence work. He writes that software engineering accounts for over half of AI tool usage across professions, with every other category still in single digits.

Copilots and autopilots

A copilot sells a tool to a professional, who stays responsible for the output. Bek cites Harvey selling to law firms and Rogo selling to investment banks. An autopilot sells the outcome directly to the end customer. His examples are Crosby, which sells NDA drafting to the company that needs one, and WithCoverage, which sells insurance to the CFO rather than to the broker.

Bek's rule: the higher the share of intelligence work in a field, the sooner autopilots win.

Outsourced work as the wedge

Bek writes that for every dollar spent on software, six are spent on services. He recommends starting where work is already outsourced, for three reasons:

• The buyer has accepted that the work can be done externally.
• A budget line already exists and can be swapped cleanly.
• The buyer is already purchasing an outcome.

He summarizes the difference: replacing an outsourcing contract is a vendor swap, while replacing headcount is a reorganization. The playbook is to start with outsourced, intelligence-heavy tasks and expand toward in-house, judgement-heavy work as the system improves.

Sequoia's opportunity map

Bek plots services categories and gives labour market estimates. Examples from the essay:

• Insurance brokerage: $140B to $200B.
• Accounting and audit: $50B to $80B outsourced in the US.
• Healthcare revenue cycle: $50B to $80B outsourced in the US.
• IT managed services: over $100B.
• Transactional legal work: $20B to $25B.
• Management consulting: $300B to $400B, which Bek describes as mostly judgement work.

On accounting, Bek cites a US shortage of roughly 340,000 accountants over five years and notes that 75% of CPAs are nearing retirement. He presents the map as illustrative.

Founder takeaways

• List the outsourced spend in your target customer. Find the invoices they already pay to agencies, bookkeepers or law firms.
• Score each task on intelligence versus judgement. Start with the most rule-based task that has a clear quality check.
• Price per outcome. Match the unit the customer already buys, such as a closed month or a drafted contract.
• Plan the expansion path. Show investors how each completed task builds data for the next, harder one.

Bek also notes the innovator's dilemma facing copilot companies: selling the work cuts out their own customers. That gap is where he expects pure-play autopilots to win.

Sources

1. Services: The New Software
   Sequoia Capital (Julien Bek) · Mar 5, 2026