Author: Julien Bek
Compiled by: Deep Tide TechFlow
Deep Tide Intro: Sequoia Capital partner Julien Bek has written a clearly structured article, with the core thesis being: the next trillion-dollar company will not sell software tools, but will directly sell work outcomes. For every $1 spent on buying software, businesses spend $6 on services. When AI drives the cost of "doing things" close to zero, the real opportunity lies not in Copilots (assistive tools), but in Autopilots (automating the work).
He breaks down the automation opportunities in service industries like insurance, accounting, healthcare, law, IT, procurement, recruitment, and consulting one by one, including a matrix chart plotting opportunities along the dimensions of "Intelligence vs. Judgment" and "Outsourced vs. Internal." It offers reference value for both 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 an AI tool is asking the same question: what happens when the next version of Claude turns my product into a feature? This fear is justified. If you sell tools, you are racing against the models. But if you sell the work itself, every improvement in the model makes your service faster, cheaper, and harder to compete with. A company might spend $10,000 a year on QuickBooks, and then $120,000 on an accountant to close the books. The next legendary company will just close the books for you.
Intelligence vs. Judgment
Writing code is mostly "intelligence." Knowing what to do next is "judgment."
Translating a requirements document into code, testing, debugging: the rules are complex, but they are ultimately rules. Judgment is different. It requires experience and taste, an intuition built up through years of practice. Deciding what feature to build next, whether to take on technical debt, when to ship before it's perfect.
A year ago, most Cursor users used AI as autocomplete. Today, agent-initiated tasks outnumber human-initiated ones. Software engineering accounts for over half of all AI tool usage across professions, while all other categories are still in the single digits. The reason is that software engineering is primarily intelligence work. AI has crossed that line—it can autonomously perform most of the intelligence work, leaving judgment to humans. Software engineering got there first, but it will spread to every profession.
Caption: AI tool usage by profession, software engineering far exceeds other categories
Copilot and Autopilot
Copilot sells tools. Autopilot sells work.
Until recently, AI models were still developing in both intelligence and judgment, so the right path was to start with Copilot: put AI in the hands of professionals and let them decide how to use it. Harvey sells to law firms, Rogo sells to investment banks. The professionals are the customers; the 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 straight to Autopilot. Crosby sells to companies that need to draft NDAs, not to external legal counsel. WithCoverage sells to CFOs who need insurance, not to insurance brokers. The customer is buying the outcome directly. In any profession, the work budget is much larger than the tool budget, and Autopilot can capture the work budget from day one.
The higher the proportion of intelligence in a field, the faster Autopilot wins.
Convergence
Today's judgment becomes tomorrow's intelligence. As AI systems accumulate proprietary data on "what good judgment looks like" in their respective fields, the frontier moves. Copilot and Autopilot will converge. The transition from Copilot to Autopilot has already begun in several categories. But the starting position matters because it determines where Autopilot can win customers now and start accumulating the data that will eventually allow it to handle judgment tasks as well.
The Autopilot Play: Outsourcing is the Entry Point
For every $1 spent on software, $6 is spent on services.
The TAM for Autopilot is all labor expenditure in a category, both internal and outsourced. But the right starting point is where outsourcing already exists.
If a task is already outsourced, it tells you three things. First, the company has already accepted that this work can be done externally. Second, there is an existing budget line item that can be cleanly replaced. Third, the buyer is already buying an outcome. Replacing an outsourcing contract with an AI-native service provider is a supplier swap. Replacing an internal employee is an organizational restructuring.
The play is: start with outsourced, intelligence-intensive tasks. Nail distribution. As the AI accumulates data, expand into internal, judgment-intensive work. Outsourced tasks are the wedge; internal work is the long-term TAM.
Crosby started with NDAs: a well-defined task, mostly intelligence work, that most companies already outsource to external counsel. The budget is ready, the scope is clear, ROI is immediate, replacement is frictionless.
Opportunity Map
Plotting each service vertical on a spectrum from "Intelligence to Judgment" and the ratio of "Outsourced to Insourced" yields a prioritization map, with the labor TAM in parentheses. The list below is not exhaustive.
Caption: Autopilot opportunity matrix for various service verticals (distributed by Intelligence/Judgment ratio and Outsourced/Insourced ratio)
Insurance Brokerage ($140-200 Billion).
The largest market on this list. Standard commercial insurance is highly standardized: the broker's added value is essentially comparing prices and filling out forms between different underwriters, pure intelligence work. The distribution layer is extremely fragmented, with thousands of small brokers each running the same process, none controlling the customer relationship. WithCoverage and Harper are interesting new entrants.
Accounting & Auditing ($50-80 Billion outsourced in US alone).
The US has lost about 340,000 accountants in the last five years, while demand has grown. 75% of CPAs are nearing retirement, the licensure path is long, and starting salaries lag behind tech and finance. This structural shortage is driving accounting firms to adopt AI faster than almost any other profession. Rillet is building an AI-native ERP to close the books directly. Basis started as a Copilot for accountants.
Healthcare Revenue Cycle Management ($50-80 Billion outsourced in US).
Hearing "healthcare" makes one think judgment-intensive, but the billing layer is almost pure intelligence work. Medical coding is translating clinical notes into about 70,000 standardized ICD-10 codes. The rules are complex but ultimately rules. Outsourcing is already mature and billed on outcomes. Autopilot just needs to do the same thing at a lower cost. Anterior is the farthest along.
Claims Adjusting ($50-80 Billion including TPAs).
On the other side of the insurance policy, claims adjusting is another distinct Autopilot scenario. Adjusting claims for standard lines involves adjudicating against policy language and a list of damages, setting reserves with actuarial tables. The adjuster workforce is aging, with no one replacing them. The market is heavily outsourced to independent adjusters and TPAs like Crawford and Sedgwick. One industry, at least two different Autopilot opportunities. Pace is doing Autopilot for claims processing, Strala is building an AI-native TPA.
Tax Advisory ($30-35 Billion).
The CPA licensure system creates a regulatory moat, but 80%-90% of the underlying work is intelligence work. A Tax Autopilot's data moat deepens with each additional jurisdiction it covers. The complexity of multiple jurisdictions is precisely why SMBs outsource it, as no internal accountant can cover it all. TaxGPT is an early mover; in Europe, there's Skalar and Ravical.
Legal Transactional Work ($20-25 Billion).
Contract drafting, NDAs, regulatory filings: high intelligence share, routinely outsourced. The work product is standardized enough that quality is verifiable, so buyers can trust AI output without deep legal expertise. Harvey is the rising leader, quickly moving towards Autopilot; Crosby and Lawhive are Autopilot-native new entrants.
IT Managed Services ($100+ Billion).
Every SMB outsources IT. Patching, monitoring, user provisioning, alert triage: intelligence work run repeatedly across thousands of identical environments. Existing software layers (ConnectWise, Datto) sell tools to MSPs. No one yet sells "your IT is running" as an outcome directly to companies. Edra is automating IT processes, Serval is automating IT support.
Supply Chain & Procurement ($200+ Billion).
Most companies only seriously negotiate with their top 20% of suppliers. The long tail is completely unmanaged because it's not worth a person's time. Contract leakage accounts for 2%-5% of total procurement spend. The entry point is the abandoned work: no budget line to justify, no incumbent to displace, just found money. Magentic is doing AI for direct procurement, AskLio for indirect procurement. Tacto is building both the system of record and a Copilot for the mid-market.
Recruiting & Staffing ($200+ Billion).
The largest service market on this list. The top of the recruiting funnel (screening, matching, outreach) is pure intelligence work, but closing and assessing culture fit is judgment built on years of pattern recognition. The Autopilot entry point is 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).
Huge market, but the work is mostly judgment. The interesting question is whether AI can unbundle consulting into intelligence components (data gathering, benchmarking) and judgment components (strategic advice), automating the intelligence layer and leaving the judgment to humans. Best candidates TBD.
The fastest-growing AI companies in 2025 were Copilots. In 2026, many will try to become Autopilots. They have product and customer awareness. But they also face the innovator's dilemma: selling work means kicking their own customers out of their jobs. This is the window of opportunity for pure Autopilot companies.
