When “Services Are the New Software” Stops Being a Metaphor
For the past three years, venture capital has debated whether AI would augment professional services or replace them. In 2026, that debate has been settled by data: a class of startups is generating real revenue by delivering service outcomes — completed insurance claims, reconciled accounts, processed medical codes — without the human labor that previously made those outcomes expensive.
Sequoia Capital’s early-stage investor Julien Bek published “Services: The New Software,” laying out the thesis explicitly: the next trillion-dollar company will be a software company masquerading as a services firm. The key distinction Bek draws is between copilots (tools that assist humans) and autopilots (AI that delivers outcomes). For every software dollar spent globally, six go to services — meaning the total addressable market for autopilots is roughly six times the SaaS market.
Emergence Capital has gone further, publishing an AI-Native Services Playbook and a market map covering 77 companies across 8 sectors, identifying a $3 trillion addressable market across outsourced professional work. The Emergence thesis frames “AI-Native Services” (AINS) as businesses that “leverage AI to deliver services faster, better, and cheaper than incumbents” — and achieves 5-10x improvements in speed or throughput while operating at 50+ percent gross margins, making them structurally more investable than traditional services firms that generate 15-20 percent margins.
The Business Process Outsourcing market exceeded $300 billion in 2024 and was projected to surpass $525 billion by 2030 before the AI-native disruption thesis arrived. That growth projection is now under serious revision.
The Sectors Under Pressure — and the Startups Moving In
The disruption is not uniform. Emergence Capital’s analysis identifies the most vulnerable verticals as those with high-volume, repeatable workflows — the kind where a trained professional follows established rules most of the time and exercises genuine judgment rarely. These are precisely the workflows that AI agents can reliably automate today.
Insurance brokerage and claims is the leading battleground. The insurance services market spans $140–200 billion in brokerage and $50–80 billion in claims adjusting. Sequoia has backed Pace, which raised a $10 million Series A to replace manual insurance BPO workflows with autonomous AI agents targeting the $70 billion insurance BPO market. Emergence Capital led the Series A of Harper, bringing its total funding to $47 million for an AI-native commercial insurance brokerage. WithCoverage targets the same category from the consumer side.
Healthcare revenue cycle management (RCM) is the second major battleground, with a $50–80 billion market in medical coding, claims processing, and prior authorization. The workflow is essentially rule-based translation — converting medical procedures into billing codes under regulatory frameworks that change regularly — making it a near-perfect target for AI that can ingest regulatory updates and apply them consistently at scale. Anterior and Prosper AI are the companies gaining traction here.
Accounting and audit is a $50–80 billion market facing a structural CPA shortage that has been building for a decade. Rillet is positioning as an AI-native accounting firm that can deliver month-end closes and financial statement preparation faster and at lower cost than Big 4 junior teams. Emergence notes that the Big 4 firms alone generate over $200 billion in annual revenue, much of it from increasingly commoditized work that is “characterized by high-volume, repeatable workflows.”
Legal transactional work ($20–25 billion), recruitment and staffing ($200+ billion), and management consulting ($300–400 billion) complete the map — though the last category is weighted toward judgment-heavy work that AI disrupts more slowly.
Advertisement
What This Means for Founders and Investors Watching This Space
The AI-native services wave creates a specific set of opportunities and risks that differ from traditional SaaS investing. Understanding the structural dynamics separates companies positioned to capture genuine value from those replicating the pattern without the underlying defensibility.
1. The Outsourcing Wedge Is the Right Entry Strategy — Not the Disintermediation Play
Sequoia’s analysis makes a critical distinction that many founders miss: replacing an outsourcing contract is a vendor swap (low friction), while replacing internal headcount is an organizational restructuring (high friction). The fastest path to revenue is targeting existing outsourced workflows first — where the decision-maker is a procurement officer comparing vendors, not an HR director managing a workforce reduction. Companies that pitch themselves as “we replace your BPO contract” land customers faster than companies that pitch “we replace your team.” Build for the outsourcing wedge first.
2. Gross Margin Architecture Determines Investability
Traditional services firms operate at 15-20 percent gross margins because human labor is their primary cost. AI-native services firms can operate at 50+ percent gross margins because their primary cost is compute — and compute costs are declining. This margin difference is what makes the category venture-backable. Founders in this space should model their gross margins explicitly from day one and design their service delivery architecture to minimize the human-in-the-loop interventions that erode margins. Every workflow step that requires human review is a gross margin leak.
3. Regulatory Moats Are the New Competitive Advantage
The most defensible AINS companies are building in regulated verticals where AI must demonstrate compliance with specific rules, certifications, and audit trails. Insurance AI must satisfy state-level regulatory requirements. Healthcare coding AI must comply with CMS and ICD coding standards. This regulatory complexity — which deters casual entrants — is precisely what creates sustainable competitive advantage for first movers who invest in compliance infrastructure early. Founders should treat regulatory approval not as a cost but as a moat-building investment.
4. The Incumbent Response Timeline Gives Startups a 2-3 Year Window
Traditional BPO firms — Accenture, Cognizant, Infosys, Teleperformance — are beginning to deploy AI in their workflows. But their business model relies on billable human hours, and transitioning to outcome-based pricing requires restructuring compensation, delivery models, and client contracts simultaneously. Emergence Capital notes that “when the world’s largest consultancy is more worried about a blog post than its own innovation pipeline, incumbent competitive moats have weakened.” The typical incumbent response time in professional services is 3-5 years from thesis recognition to scaled deployment — giving well-capitalized AI-native startups a meaningful head start.
The Regulatory Question: What Happens When AI Makes Mistakes at Scale
The AI-native services thesis has one structural vulnerability that its proponents acknowledge: when an AI agent makes an error in a traditional service context, the liability is human and therefore insurable and remediable. When an AI agent makes an error at scale — miscoding 10,000 medical claims, mispricing 5,000 insurance renewals — the liability exposure is potentially catastrophic.
The companies that will survive this question are those that build explicit human-in-the-loop review at the highest-stakes decision points while automating the routine majority. Harper, the Emergence-backed insurance brokerage, uses licensed human brokers to review edge cases while AI handles standard commercial lines — a hybrid model that maintains regulatory compliance while delivering cost advantages over pure-human brokerage.
This is not a failure of the thesis. It is the maturation of it. The first generation of AI-native services companies will define the liability frameworks, the audit trail standards, and the human-AI workflow designs that make the second generation structurally safer. The founders building in 2026 are writing the playbook that their successors will inherit.
The $3 trillion opportunity is real. The competitive window is open. The question is which founders build service models disciplined enough to survive the inevitable error events that will define how regulators and enterprise buyers understand this new category.
Frequently Asked Questions
Q: What distinguishes an AI-native services company from a traditional SaaS company with AI features?
A traditional SaaS company sells a tool — customers pay for access to software and use it to do their work. An AI-native services company sells an outcome — customers pay for completed tasks (processed claims, reconciled accounts, coded procedures) and the AI does the work. The pricing model shifts from seat-based or usage-based licensing to outcome-based contracts, and the company takes on delivery responsibility rather than merely providing a tool.
Q: Which sectors are safest for AI-native services investment?
According to Emergence Capital’s framework, the safest sectors are those with high-volume, repeatable, rules-based workflows in regulated industries — insurance, healthcare revenue cycle, accounting/audit, and legal transactional work. Management consulting and creative professional services remain more resistant because they involve judgment, relationship, and creativity components that AI cannot reliably replicate at required quality thresholds.
Q: How does the AI-native services model protect against model commoditization?
If the underlying AI model (GPT-5, Claude, Gemini) becomes a commodity, companies that have built proprietary workflow automation, regulatory compliance layers, data pipelines, and client integration infrastructure retain their competitive advantage. The model is an input, not the product. Companies that treat their regulatory knowledge, audit trail systems, and client-specific workflow automations as the core IP are more defensible than companies that compete primarily on model quality.
Sources & Further Reading
🔗 Related Intelligence
OpenAI’s Deployment Company: $10B to Embed AI Engineers Inside Your Business
The AI-Native Startup Stack: Infrastructure Choices That Define the Next Generation of
Anthropic’s $30B Run Rate: How Enterprise AI Demand Tripled Revenue in Four Months
