Source: Sequoia - Services: The New Software
What It Covers
Julien Bek’s Sequoia essay argues that the next large AI company may look less like a software vendor and more like a services firm rebuilt around AI. The core distinction is between selling a tool that professionals use and selling the completed work directly.
Key Claims
- Selling AI tools puts companies in a race against frontier model capabilities; selling the work means model improvements make the service faster, cheaper, and harder to compete with.
- The essay separates intelligence from judgment: software engineering crossed the threshold first because much of the work is rule-governed intelligence, while judgment remains human-led.
- A copilot sells a tool to a professional; an autopilot sells the outcome to the buyer who needs the work done.
- Autopilots can capture labor or services budgets, which are much larger than software budgets in many categories.
- The best initial wedge is outsourced, intelligence-heavy work because the buyer already accepts external delivery, has an existing budget line, and is purchasing an outcome.
- The long-term expansion path is from outsourced intelligence-heavy work toward insourced, judgment-heavy work as the AI system compounds domain data and judgment examples.
- Sequoia’s opportunity map highlights insurance brokerage, accounting and audit, healthcare revenue cycle, claims adjusting, tax advisory, transactional legal work, IT managed services, supply chain/procurement, recruitment/staffing, and management consulting.
- In IT managed services, the essay frames SMB IT work such as patching, monitoring, provisioning, and alert triage as repeated intelligence work across many similar environments.
- The essay names Serval as automating IT support and Edra as automating IT processes.
Strategic Interpretation
- This source strengthens ai-autopilot-services as a broader market frame for service-led-ai-itsm-delivery, vertical-ai-for-services-economy, and outcome-automation-vs-step-automation. ^[inferred]
- For initlabs, the sharpest question is not only “which software category?” but “which outsourced or under-served work budget can be substituted with a better AI-native outcome?” ^[inferred]
- The essay supports treating Treeline and service-led managed IT as structurally important, not merely adjacent, because they sell the work rather than only the ITSM tool. ^[inferred]
Limitations
- The opportunity map is illustrative and investor-authored, not a market-size audit.
- Some named companies are cited as examples without detailed product or traction evidence in this source.
- The essay’s “next $1T company” framing is thesis-driven and should not be treated as a forecast. ^[ambiguous]
Concepts Informed
- ai-autopilot-services
- service-led-ai-itsm-delivery
- outcome-automation-vs-step-automation
- vertical-ai-for-services-economy
- back-office-automation