In the rapidly evolving landscape of artificial intelligence, many founders of AI-driven companies share a common concern: the fear that their innovative tools might soon be relegated to mere features in the next big model release. In a thought-provoking essay circulating within venture capital circles, Sequoia partner Julien Bek articulates this anxiety while also presenting a pathway forward.
Signage with logo at headquarters of venture capital investment firm Sequoia Capital, on Sand Hill Road in the Silicon Valley town of Menlo Park, California, August 25, 2016. (Photo via Smith Collection/Gado/Getty Images).
Getty Images
Currently, enterprises allocate about six dollars for services for every dollar spent on software. This creates a vast market for firms that focus on delivering AI-driven services rather than just AI-enhanced tools. Bek argues that the next groundbreaking company will simplify and streamline complex outcomes, similar to closing financial accounts, underscoring the immense opportunity in automating tasks traditionally handled by human labor.
Copilots Versus Autopilots
Bek categorizes AI solutions into two main types: copilots and autopilots. A copilot provides tools tailored for professionals, such as Harvey for law firms and Rogo for investment banks. Conversely, autopilots directly deliver the final product to clients without intermediary professionals. For instance, Crosby drafts NDAs directly for companies that require them, bypassing external legal counsel, while WithCoverage sells insurance solutions straight to CFOs.
Professionals generally purchase software within a specific tools budget, but autopilots target a much broader work budget from the outset.
As AI technology matures, the distinction between copilots and autopilots becomes increasingly relevant. Once, copilots were necessary as AI was still developing the capabilities to perform structured tasks effectively. However, according to Bek, we’ve now reached a point where AI can independently initiate tasks that previously relied on human intervention. Software engineering now constitutes over half of AI tool usage, illustrating the sector’s dominant reliance on intelligence.
Where VCs Are Placing Bets
In 2025, global funding for AI surged to an impressive $203 billion, marking a 75% increase from the previous year, as reported by Crunchbase. While foundation model companies secured a sizeable portion of this funding, there has been a noticeable acceleration in investments directed at niche vertical AI firms in early 2026.
One standout example is Harvey, a legal AI company that raised an impressive $200 million in March 2026, achieving an $11 billion valuation co-led by Sequoia and GIC. The firm’s annual recurring revenue skyrocketed from $100 million in August 2025 to over $190 million by January, with more than 100,000 attorneys leveraging the platform across 1,300 organizations. Initially positioned as a copilot, Harvey’s trajectory now points toward complete workflow automation, embodying the autopilot concept.
Sequoia’s analysis sheds light on various sectors attracting investor interest, primarily focusing on insurance brokerage, with a total addressable market ranging from $140 to $200 billion. The fragmented nature of the distribution landscape and the high standardization of commercial lines create a ripe environment for disruption. Additionally, healthcare revenue cycle management, valued between $50 and $80 billion in the U.S. and involving highly standardized medical coding processes, presents an almost pure intelligence domain ripe for AI integration. With a significant staffing shortage in accounting and a looming retirement crisis among CPAs, AI adoption here is gaining momentum as well.
The Outsourcing Wedge
Bek advocates that the best starting point for developing autopilots lies in tasks that have already been outsourced. This outsourcing trend indicates three key factors: companies recognize that the work can be efficiently completed by external providers, a specific budget allows for straightforward replacement, and clients are already focusing on outcomes rather than just manpower. As Bek emphasizes, “Replacing an outsourcing contract with an AI-native services provider is a vendor swap; replacing headcount is a reorganization.”
Crosby effectively utilized the drafting of NDAs as a wedge into a well-defined, increasingly intelligence-driven task, which is routinely outsourced. This approach offers clear scopes, immediate ROI, and seamless transitions. Bek foresees that autopilots will likely begin by addressing outsourced responsibilities before further venturing into more complex, insourced work as their AI capabilities evolve.
The transition from relying on human judgment to automated intelligence represents a longer arc in this thesis. As AI systems gather proprietary insights into effective professional judgment within specific domains, tasks once requiring human expertise will increasingly become automatable. The evolution of copilots and autopilots is expected to converge over time.
The Innovator’s Dilemma for Incumbents
Bek articulates the significant challenge that copilot companies face when attempting to pivot to autopilots: selling outcomes directly often means sidelining existing customers who typically handle these tasks. For instance, a law firm using Harvey’s tools has lawyers involved in reviewing and billing for the work. An autopilot, however, bypasses the firm entirely, creating opportunities for startups that do not have incumbent relationships to safeguard.
The overall economic landscape further emphasizes the urgency. Gartner projects global IT spending to reach $6.15 trillion in 2026, reflecting a 10.8% year-over-year growth. Software spending is also expected to rise by 14.7%, reaching $1.43 trillion. However, Sequoia’s framework highlights an even larger opportunity: global IT services spending stands at around $1.73 trillion and is on the rise. If Bek’s predictions hold true, autopilots will aim to capture a share of this expansive services market rather than limiting themselves to software.
For investors, this insight facilitates a prioritization strategy. Targeting service verticals that are heavily driven by intelligence, primarily outsourced, and lacking sufficient human talent becomes essential. Key areas include legal transactional work, insurance brokerage, healthcare billing, tax compliance, and managed IT services. Management consulting, a sector valued between $300 and $400 billion, might pose challenges due to its heavy reliance on judgment, though Bek notes that segments involving intelligence-driven components like benchmarking and data gathering could be addressable.
Map of market opportunities, source: https://sequoiacap.com/article/services-the-new-software/
Sequoia Capital
What It Means for Founders
In closing, Bek issues a direct challenge to founders, sharing his email and X handle. His message signals a need to rethink what a scalable AI business looks like in 2026. While revenue from a $10,000 software seat is typically limited to the tool budget, a $120,000 accounting engagement’s revenue is determined solely by how many engagements an autopilot can efficiently handle. For the first time, scalability, rather than price sensitivity, emerges as the defining constraint.
However, the framework does pose risks concerning concentration. Autopilots entering mature outsourcing processes will have to contend with established service providers, existing outsourcing firms, and professional service networks that have built strong customer relationships and compliance infrastructures. Successfully capturing these vendor switches means not only matching service quality but also aligning with existing contractual and regulatory frameworks, which can be significantly challenging.
In conclusion, the insights from Sequoia underline where the leverage lies: for every dollar spent on software, six are allocated to services, a segment eager for innovation and improvement.
