# TEL501: Services: The New Software (Sequoia Capital)
https://www.telosready.com/skills/TEL501?v=1
Summary of Sequoia Capital's March 2026 article arguing that the next trillion-dollar companies will sell AI-delivered outcomes (autopilots) rather than AI tools (copilots), with a vertical-by-vertical opportunity map.

## Instructions
# Services: The New Software (Sequoia Capital)

**Source:** [Services: The New Software](https://sequoiacap.com/article/services-the-new-software/) — Julien Bek, Sequoia Capital, March 2026

---

## Core Thesis

The next $1T company will be a software company masquerading as a services firm. Companies that sell AI tools are in a race against the model. Companies that sell the *work* benefit from every model improvement — it makes their service faster, cheaper, and harder to displace.

> "A company might spend $10K a year for QuickBooks and $120K on an accountant to close the books. The next legendary company will just close the books."

---

## Intelligence vs Judgement

- **Intelligence work** — rule-based tasks: translating specs, testing, coding, form-filling. AI has crossed the threshold to handle this autonomously.
- **Judgement work** — experience and taste: deciding what to build, when to ship, whether to take on debt. This still requires humans — for now.

The higher the intelligence ratio in a field, the sooner AI autopilots will win. Software engineering got there first; every profession will follow.

---

## Copilots vs Autopilots

| | Copilot | Autopilot |
|---|---|---|
| **What's sold** | The tool | The work |
| **Customer** | The professional | The company buying the outcome |
| **Budget captured** | Tool budget | Work/labour budget |
| **Example** | Harvey (sells to law firms) | Crosby (sells to the company needing an NDA) |

For every $1 spent on software, $6 are spent on services. Autopilots capture the work budget from day one.

---

## The Autopilot Playbook: Outsourcing as the Wedge

Start where outsourcing already exists. Outsourced work signals three things:
1. The company accepts this work can be done externally
2. A budget line already exists that can be substituted cleanly
3. The buyer is already purchasing an outcome — not headcount

**Replacing an outsourcing contract is a vendor swap. Replacing headcount is a reorg.**

The playbook:
1. Start with the outsourced, intelligence-heavy task
2. Nail distribution
3. Expand toward insourced, judgement-heavy work as AI compounds

---

## Opportunity Map by Vertical

| Vertical | Outsourced TAM (US) | Intelligence ratio | Notable companies |
|---|---|---|---|
| Recruitment & staffing | $200B+ | High (top of funnel) | Juicebox, Mercor, Jack & Jill |
| Management consulting | $300–400B | Lower (mostly judgement) | TBD |
| Supply chain & procurement | $200B+ | High (long-tail suppliers) | Magentic, AskLio, Tacto |
| IT managed services | $100B+ | High (patching, monitoring, provisioning) | Edra, Serval |
| Insurance brokerage | $140–200B | Very high (shopping carriers, filling forms) | WithCoverage, Harper |
| Accounting & audit | $50–80B | High (structural shortage accelerating adoption) | Rillet, Basis |
| Healthcare revenue cycle | $50–80B | Very high (medical coding is rules, not judgement) | Anterior |
| Claims adjusting | $50–80B | High (policy language + actuarial tables) | Pace, Strala |
| Tax advisory | $30–35B | 80–90% intelligence | TaxGPT, Skalar, Ravical |
| Legal (transactional) | $20–25B | High (contract drafting, filings) | Harvey, Crosby, Lawhive |

---

## The Convergence

Today's judgement becomes tomorrow's intelligence. As autopilots accumulate proprietary data on what good judgement looks like in their domain, the frontier shifts. Copilots and autopilots will eventually converge — but starting position matters. It determines where autopilots can win customers now and begin compounding the data advantage that will let them handle judgement too.

---

## The Innovator's Dilemma for Copilots

In 2025, the fastest-growing AI companies were copilots. In 2026, many will attempt the transition to autopilot. The challenge: selling the work means cutting existing customers out of doing it. That tension is the opening for pure-play autopilots built from scratch.

# Services: The New Software (Sequoia Capital)

**Source:** [Services: The New Software](https://sequoiacap.com/article/services-the-new-software/) — Julien Bek, Sequoia Capital, March 2026

---

## Core Thesis

The next $1T company will be a software company masquerading as a services firm. Companies that sell AI tools are in a race against the model. Companies that sell the *work* benefit from every model improvement — it makes their service faster, cheaper, and harder to displace.

> "A company might spend $10K a year for QuickBooks and $120K on an accountant to close the books. The next legendary company will just close the books."

---

## Intelligence vs Judgement

- **Intelligence work** — rule-based tasks: translating specs, testing, coding, form-filling. AI has crossed the threshold to handle this autonomously.
- **Judgement work** — experience and taste: deciding what to build, when to ship, whether to take on debt. This still requires humans — for now.

The higher the intelligence ratio in a field, the sooner AI autopilots will win. Software engineering got there first; every profession will follow.

---

## Copilots vs Autopilots

| | Copilot | Autopilot |
|---|---|---|
| **What's sold** | The tool | The work |
| **Customer** | The professional | The company buying the outcome |
| **Budget captured** | Tool budget | Work/labour budget |
| **Example** | Harvey (sells to law firms) | Crosby (sells to the company needing an NDA) |

For every $1 spent on software, $6 are spent on services. Autopilots capture the work budget from day one.

---

## The Autopilot Playbook: Outsourcing as the Wedge

Start where outsourcing already exists. Outsourced work signals three things:
1. The company accepts this work can be done externally
2. A budget line already exists that can be substituted cleanly
3. The buyer is already purchasing an outcome — not headcount

**Replacing an outsourcing contract is a vendor swap. Replacing headcount is a reorg.**

The playbook:
1. Start with the outsourced, intelligence-heavy task
2. Nail distribution
3. Expand toward insourced, judgement-heavy work as AI compounds

---

## Opportunity Map by Vertical

| Vertical | Outsourced TAM (US) | Intelligence ratio | Notable companies |
|---|---|---|---|
| Recruitment & staffing | $200B+ | High (top of funnel) | Juicebox, Mercor, Jack & Jill |
| Management consulting | $300–400B | Lower (mostly judgement) | TBD |
| Supply chain & procurement | $200B+ | High (long-tail suppliers) | Magentic, AskLio, Tacto |
| IT managed services | $100B+ | High (patching, monitoring, provisioning) | Edra, Serval |
| Insurance brokerage | $140–200B | Very high (shopping carriers, filling forms) | WithCoverage, Harper |
| Accounting & audit | $50–80B | High (structural shortage accelerating adoption) | Rillet, Basis |
| Healthcare revenue cycle | $50–80B | Very high (medical coding is rules, not judgement) | Anterior |
| Claims adjusting | $50–80B | High (policy language + actuarial tables) | Pace, Strala |
| Tax advisory | $30–35B | 80–90% intelligence | TaxGPT, Skalar, Ravical |
| Legal (transactional) | $20–25B | High (contract drafting, filings) | Harvey, Crosby, Lawhive |

---

## The Convergence

Today's judgement becomes tomorrow's intelligence. As autopilots accumulate proprietary data on what good judgement looks like in their domain, the frontier shifts. Copilots and autopilots will eventually converge — but starting position matters. It determines where autopilots can win customers now and begin compounding the data advantage that will let them handle judgement too.

---

## The Innovator's Dilemma for Copilots

In 2025, the fastest-growing AI companies were copilots. In 2026, many will attempt the transition to autopilot. The challenge: selling the work means cutting existing customers out of doing it. That tension is the opening for pure-play autopilots built from scratch.