Sequoia Capital: The next trillion-dollar company won't sell software, but will directly sell outcomes

Author: Julien Bek

Translation: Shen Chao TechFlow

Shen Chao Introduction: Sequoia Capital partner Julien Bek wrote a clear framework article with the core argument: the next trillion-dollar company won’t sell software tools, but will directly sell work outcomes. For every $1 spent on software, companies spend $6 on services. When AI makes “doing work” nearly free, the real opportunity isn’t in Copilot (assistive tools), but in Autopilot (automating tasks).

He dissects automation opportunities in insurance, accounting, healthcare, legal, IT, procurement, recruiting, consulting, and more, accompanied by an opportunity matrix chart plotted along the axes of “Intelligence vs Judgment” and “Outsourcing vs In-house.” This provides valuable insights for AI entrepreneurs and investors.

Full text below:

The next trillion-dollar company will be a software company disguised as a service company.

Every founder building AI tools is asking the same question: what if the next version of Claude turns my product into a feature? This concern is valid. If you’re selling tools, you’re racing against the model. But if you’re selling the work itself, every model improvement makes your service faster, cheaper, and harder for competitors to replicate. A company might spend $10,000 annually on QuickBooks and $120,000 on accountants to close books. The next legendary company will do the bookkeeping for you directly.

Intelligence vs Judgment

Writing code is primarily “intelligence.” Knowing what to do next is “judgment.”

Translating a requirements document into code, testing, debugging: the rules are complex but ultimately are rules. Judgment is different. It requires experience and taste, intuition developed through years of practice. Deciding what feature to build next, whether to incur technical debt, or when to release before being fully prepared.

A year ago, most Cursor users used AI for autocomplete. Today, more tasks are initiated by agents than by humans. Software engineering accounts for over half of AI tool usage across professions, with all other categories still in the single digits. The reason is that software engineering is mainly intellectual work. AI has crossed that line—it’s capable of autonomously completing most intellectual tasks, leaving judgment to humans. Software engineering was the first to reach this point, but it will spread to every profession.

Caption: AI tool usage share across professions, with software engineering far surpassing others

Copilot and Autopilot

Copilot sells tools. Autopilot sells work.

Until recently, AI models were still developing in both intelligence and judgment, so the right approach was to start with Copilot: putting AI into the hands of professionals, letting them decide how to use it. Harvey sells to law firms, Rogo to investment banks. Professionals are the customers; tools make them more efficient, and they are responsible for the output.

Today, models are smart enough that in some categories, the best starting point is to go directly to Autopilot. Crosby sells to companies needing NDA drafting, rather than to external legal advisors. WithCoverage sells to CFOs needing insurance, rather than to brokers. Customers are buying results directly. In any profession, work budgets far exceed tool budgets, and Autopilot can capture the work budget from day one.

The higher the proportion of intelligence in a field, the faster Autopilot wins.

Integration

Today’s judgment will become tomorrow’s intelligence. As AI systems accumulate proprietary data on “what good judgment looks like” within their domains, the frontier shifts. Copilot and Autopilot will converge. The transition from Copilot to Autopilot has already begun in several categories. But the starting point matters because it determines where Autopilot can win customers now and begin accumulating the data that will eventually enable it to handle judgment tasks.

Autopilot Approach: Outsourcing as an Entry Point

For every $1 spent on software, $6 is spent on services.

The TAM (Total Addressable Market) for Autopilot is all labor expenditure in a category, including both in-house and outsourced work. But the right starting point is where outsourcing already exists.

If a task is already outsourced, it tells you three things: First, the company accepts that this work can be done externally. Second, there is a ready budget item that can be cleanly replaced. Third, the buyer is purchasing results. Replacing an outsourcing contract with an AI-native service provider is a vendor switch. Replacing internal staff is organizational restructuring.

The strategy is to start with outsourced, knowledge-intensive tasks. Manage distribution. As AI accumulates data, expand into internal, judgment-intensive work. Outsourced tasks are the wedge; internal work is the long-term TAM.

Crosby starts with NDA drafting: a well-defined, primarily intellectual task that most companies already outsource to external lawyers. Budget is clear, scope is defined, ROI is immediate, and replacement is frictionless.

Opportunity Map

Plot each service vertical along the “Intelligence to Judgment” spectrum and the “Outsourcing to In-house” ratio to create a priority map, with labor TAM in parentheses. The following is not exhaustive.

Caption: Autopilot opportunity matrix across service verticals (distribution of intelligence/judgment ratio and outsourcing/in-house ratio)

Insurance Brokerage ($140-200 billion)

The largest market on this list. Standard commercial insurance is highly standardized: the broker’s value is essentially in comparing quotes and filling forms—purely intellectual work. Distribution is highly fragmented, with thousands of small brokers running the same processes, none controlling the customer relationship. WithCoverage and Harper are interesting new entrants.

Accounting and Auditing (outsource-only in the US: $50-80 billion)

The US has lost about 340,000 accountants over the past five years, even as demand grows. 75% of CPAs are nearing retirement; licensing is lengthy, and starting salaries lag behind tech and finance. This structural shortage is pushing accounting firms to adopt AI faster than almost any other profession. Rillet is building AI-native ERP for direct bookkeeping. Basis starts with AI-driven CPA copilot.

Healthcare Revenue Cycle Management (US outsource: $50-80 billion)

While “healthcare” sounds judgment-intensive, billing is almost purely intellectual work. Medical coding involves translating clinical notes into about 70,000 standardized ICD-10 codes. Rules are complex but ultimately rule-based. Outsourcing is mature and billed by results. Autopilot only needs to do the same at lower cost. Anterior has gone the furthest.

Claims Adjustment (including TPA, $50-80 billion)

On the other side of insurance policies, claims adjustment is another independent Autopilot scenario. Standard claims are adjudicated based on policy language and damage lists, with reserves set using actuarial tables. The claims adjuster workforce is aging, with no new recruits. Much of the market is outsourced to independent adjusters and TPA firms like Crawford and Sedgwick. One industry, at least two Autopilot opportunities: Pace is working on claims processing Autopilot, Strala on AI-native TPA.

Tax Advisory ($30-35 billion)

CPA licensing creates a regulatory moat, but 80-90% of underlying work is intellectual. Tax Autopilot becomes more valuable as it covers more jurisdictions, deepening the data moat. Multi-jurisdiction complexity is precisely why SMEs outsource—no internal accountant can cover everything. TaxGPT is an early entrant; Europe has Skalar and Ravical.

Legal Work (contracts, NDAs, regulatory filings: $20-25 billion)

High intellectual content, often outsourced routinely. Work output is standardized enough for quality to be verifiable, allowing buyers to trust AI outputs without deep legal expertise. Harvey is a rising leader, rapidly shifting toward Autopilot; Crosby and Lawhive are new entrants native to Autopilot.

IT Managed Services ($100+ billion)

Every small and medium business outsources IT. Patching, monitoring, user configuration, alert routing: repetitive intellectual work across thousands of identical environments. Existing software layers (ConnectWise, Datto) sell tools to MSPs. No one yet sells “your IT is up and running” as a result. Edra automates IT processes; Serval automates IT support.

Supply Chain and Procurement ($200+ billion)

Most companies only negotiate seriously with the top 20% of suppliers. Long-tail suppliers are unmanaged because it’s not cost-effective. Contract leaks account for 2-5% of procurement spend. Entry points are the work that’s abandoned: no budget items to argue over, no incumbents to replace, just free money. Magentic develops direct procurement AI; AskLio handles indirect procurement. Tacto builds record systems and Copilot for mid-market simultaneously.

Recruiting and Staffing ($200+ billion)

The largest service market on this list. The top of the recruiting funnel (screening, matching, outreach) is purely intellectual, but closing and cultural fit assessment rely on pattern recognition accumulated over years. Autopilot’s entry points are high-volume, low-judgment roles where matching is standardized. Juicebox, Mercor, Jack & Jill are emerging leaders building across the spectrum.

Management Consulting ($300-400 billion)

A huge market, but work is mainly judgment-based. An interesting question is whether AI can decompose consulting into intelligence components (data collection, benchmarking) and judgment components (strategy advice), automating the intelligence layer while leaving judgment to humans. The best candidates are yet to be determined.

The fastest-growing AI companies in 2025 will be Copilot. By 2026, many will attempt to become Autopilot. They have product and customer awareness. But they also face an innovator’s dilemma: selling work means pushing clients out of their current workflows. This is the window of opportunity for pure Autopilot companies.