Software is no longer just about organizing information; its new target is the $13 trillion U.S. labor market itself. Venture capitalist Alex Rampell argues that while traditional software digitized the world’s “filing cabinets”—turning records for travel, sales, and accounting into databases—it still required humans to operate the systems. The coming wave of AI fundamentally changes this dynamic by performing the actual work, moving from a tool for labor to a replacement for it. This shift isn’t merely about efficiency; it represents a complete rethinking of what software companies sell, transitioning from licensed seats to pricing based on outcomes delivered.

This evolution turns the entire software business model on its head. Using the example of a company like Zendesk, Rampell illustrates a stark crossroads: if AI agents can answer all customer support queries, a per-seat pricing model collapses to zero. However, if the same company charges for the outcome—handling all customer support for a fixed fee—it can capture a much larger share of the client’s budget previously spent on human labor. This outcome-based approach applies across industries, from legal software drafting contracts to accounting software making collection calls, transforming software from a system of record into an active, performing agent.

The implications extend far beyond cost-cutting. AI unlocks opportunities in markets previously too small or complex for traditional software, such as providing compliance-as-a-service or serving a small optometry office that spends almost nothing on software but $45,000 on a receptionist. Crucially, AI excels in areas humans find difficult: handling intermittent demand spikes, performing demoralizing jobs like collections, ensuring regulatory compliance, and operating across dozens of languages on demand. This allows entirely new businesses to emerge—ideas like an “Airbnb for bicycles” that were once economically unviable due to high human operational costs can now be built on the backbone of cheap, scalable AI labor.

Surprising Insights

• AI expands the total addressable market for software. Instead of just selling into the existing software budget, AI companies can pitch themselves against the vastly larger line item for human labor, even in industries that traditionally spend very little on software.
• Demoralizing and intermittent jobs are prime targets. AI isn’t just for cost reduction; it’s uniquely suited for tasks with high turnover (like collections), fluctuating demand (like holiday retail support), or strict regulatory requirements where human error or emotion is a liability.
• Non-AI businesses become viable. Previously impractical business models, like peer-to-peer rentals for niche items, can now work because AI dramatically lowers customer acquisition and operational costs (like sales and support), making the unit economics feasible.
• The fastest-growing human job in the U.S. is a compliance officer, a role that exemplifies a large labor spend in an area without dominant software—presenting a massive opportunity for AI to do the job end-to-end.

Practical Takeaways

• Re-evaluate software pricing. For builders, consider outcome or transaction-based models instead of per-user seats. For buyers, negotiate contracts where you pay for results (e.g., problems solved, sales closed) rather than software access.
• Analyze job listings for automation opportunities. Look at open roles in any industry—like the optometry receptionist—and identify discrete tasks (calling patients, arguing with insurance) that could be bundled into an AI service sold directly to that business.
• Prioritize AI solutions for “moments of pain.” Focus implementation on areas with intermittent demand spikes, high turnover, multilingual needs, or strict compliance requirements, where AI offers capabilities beyond mere cost savings.
• Use AI as a wedge to enter underserved markets. Target industries with high labor costs but low existing software spend by offering an AI agent that performs a specific job, then expand into providing a broader software platform.