Key takeaways
- Enterprise software revenue scaled with headcount for thirty years. AI native companies now need fewer people to produce similar outcomes — that assumption is breaking.
- Executives think they're buying AI software. In reality, they're beginning to buy digital labor — and the target market shifts from software budgets to labor budgets, which are dramatically larger.
- AI quietly changes the leverage of individual roles — a PM coordinating two engineering pods instead of one — without changing the org chart.
- Giving every department its own AI makes each function faster individually, but the company doesn't accelerate unless the workflow itself is redesigned around AI, not bolted onto the old one.
- The strategic shift: three decades of selling software licenses gives way to a decade of selling completed work. Not software. Not seats. Not users. Outcomes.
For nearly thirty years, enterprise software followed a remarkably consistent formula. A company hired more people. More people needed software. Software vendors sold more seats. Revenue scaled with organizational growth. The larger the customer became, the larger the software contract became.
Entire SaaS businesses — and their valuations — were built on this assumption. That assumption is beginning to break. Not because companies suddenly need less software, but because AI native companies increasingly need fewer people to produce similar outcomes. That distinction matters. It changes everything.
01The new unit of competition isn't software
Over the last twelve months, we've worked with organizations ranging from venture-backed startups to Fortune 500 companies redesigning how products are built, launched, sold, and supported. One pattern keeps appearing.
Executives think they're buying AI software. In reality, they're beginning to buy labor. Not human labor, but digital labor. The difference sounds semantic and it isn't.
Historically companies bought CRM software, help desk software, and project management software. Today they're buying AI SDRs, AI customer support agents, and autonomous delivery management workflows.
The product category hasn't changed. The economic model has.
02Investors have already changed their bet
The most interesting signal isn't coming from enterprise buyers. It's coming from the investors funding tomorrow's competitors. Across recent cohorts from YC and the application-layer companies highlighted by a16z, a pattern is emerging. Many of the fastest-growing startups aren't trying to build better versions of existing SaaS. They're trying to eliminate the labor those SaaS products supported.
Look across the categories. Instead of CRM software, investors fund AI sales employees. Instead of IT ticketing software, they fund AI IT departments. Instead of legal workflow software, they fund AI legal operators. Instead of finance software, they fund AI finance operations.
Notice the pattern. The target market isn't software budgets anymore. It's labor budgets. Labor budgets are dramatically larger.
03The economics are already showing up
This isn't theoretical. We've watched it happen inside client organizations. One of the world's largest consulting firms recently deployed enterprise-approved AI tooling across product and engineering organizations. The technology worked.
Product managers suddenly generated Jira stories, analyzed research, drafted product specifications, created mockups, summarized meetings, and prepared sprint plans in a fraction of the previous effort. One experienced product manager who traditionally managed a single engineering pod (1 PM for every 5 engineers) was now effectively coordinating two or more pods.
Not because the engineers doubled in productivity. Because AI removed much of the coordination work that historically limited a product manager's capacity. The organization didn't hire better PMs. It quietly changed the economics of product management. The official organizational chart barely changed. The leverage did.
04AI is quietly changing professional economics
We have seen similar patterns across entirely different industries.
A healthcare company preparing for a Series A financing needed to rebuild its financial model after a major GTM pivot. Historically, that project would have required roughly three months of strategy and fractional CFO work. Using AI-assisted financial analysis and modeling, the work was completed in about two weeks.
The financial model improved. The operating model did not. The company still thought in terms of buying expert hours. It had actually purchased an outcome.
05AI doesn't just replace work. It rewrites management
The more interesting changes happen one layer higher. Inside another client organization, AI adoption became widespread. Marketing had AI. Sales had AI. Product had AI. Engineering had AI.
Each function became individually faster. Yet the company itself barely accelerated. Why? Because every department optimized its own work. Nobody redesigned the system connecting them. Product managers generated better specifications. Engineering generated better code. Marketing produced more campaigns. Executives received more dashboards.
Meetings remained the same. Approvals remained the same. Hand-offs remained the same. Management remained the bottleneck.
The organization had purchased AI tools. It hadn't redesigned work.
06The companies pulling away think differently
The organizations making the largest gains don't start with tools. They start with workflows.
Recently we worked with a Fortune 500 company testing an AI-orchestrated product launch system. Instead of asking "Where should we use AI?" they asked "What would the product launch process look like if AI existed from day one?"
The answer looked radically different. Human experts focused on judgment. AI agents handled research, competitive analysis, pricing scenarios, consumer segmentation, campaign generation, operational planning, and coordination.
The innovation leader compared this AI-orchestrated launch against the company's traditional launch methodology. The comparison surprised people. Not because AI generated better PowerPoint slides. Because AI continuously incorporates real customer signals — including social media and direct-to-consumer data — rather than relying primarily on historical research and periodic consumer studies.
The management conversation shifted. The question stopped being "Can AI write?" It became "Why are humans still coordinating this workflow manually?"
07AI as labor is already reshaping go-to-market
Instead of expanding SDR headcount and agency relationships, leadership began redesigning its go-to-market system around coordinated AI agents. Those agents continuously analyzed CRM activity, pilot feedback, customer conversations, sales cycles, and historical conversion data.
Rather than replacing salespeople, they continuously reprioritized opportunities, refined messaging, surfaced objections earlier, and recommended next actions. Sales became progressively less dependent on individual heroics. It became an improving system.
That's the difference between buying AI software and deploying AI labor.
08The companies investors are funding start here
This explains why many AI-native startups look strange through traditional SaaS lenses. The goal isn't to build software people use. The goal is to build organizations that require dramatically fewer people.
Andreessen Horowitz's recent analysis of AI startup spending found a growing class of companies explicitly positioning themselves as end-to-end AI employees rather than productivity tools. Customer service, legal operations, IT support, sales development, and software engineering increasingly appear as autonomous services rather than software categories.
Software augments labor. AI labor increasingly competes with labor.
09CEOs must ask a different question
Most executive conversations still begin here: "How should we use AI?" That was the right question eighteen months ago. Today it is incomplete.
The more important question is: which functions and tasks should continue to depend on human coordination, and which should become AI-orchestrated systems?
That is no longer a technology decision. It is an operating model decision. It determines organizational structure. Management layers. Revenue per employee. Margins. Speed. Competitive positioning. Ultimately, valuation.
10The strategic shift
The software industry spent three decades selling licenses. The next decade may belong to companies selling completed work.
Not software. Not seats. Not users.
Outcomes.
And the companies that recognize that shift early won't simply deploy more AI. They will redesign how their businesses create value.