Conference Presentation, Keynote
Software is Eating Labor
- The global SaaS market is valued at approximately $300 billion annually, with a worldwide market cap of $2.2 trillion, whereas the U.S. labor market alone represents $13 trillion.
- Software is transitioning from a model of digitizing "filing cabinets" (creating digital records) to one that executes labor tasks end-to-end (acting as an autonomous workforce).
- The speaker posits a new economic equation: capital invests in GPU infrastructure and engineering talent to produce software that performs the function of human labor, effectively replacing wage costs with software fees.
- Historical precedents for automation (e.g., the loom, steamships, the printing press, and assembly lines) increased efficiency but still required humans to operate the machinery; current AI software performs the entire workflow autonomously.
- The speaker identifies a "Tall-Grande-Venti" pricing crisis for SaaS companies that charge per seat, noting that if AI makes agents 9,000 times more productive, revenue models based on human headcount could collapse to zero.
- Zendesk Case Study:
- Currently generates $1.4 million in software revenue annually against $75 million in human labor costs for a 1,000-seat call center.
- Faces a binary outcome: revenue drops to $0 if customers require zero seats, or potentially rises to $5 million+ if the company shifts to outcome-based pricing (charging for the value of resolved support rather than seats).
- The company is currently piloting outcome-based pricing models in New Zealand to test this transition.
- Nursing Market Example: Registered nurses in the U.S. earn $650 billion annually, a figure exceeding the entire global software market, indicating a massive addressable market for software to automate nursing-adjacent tasks.
- Target Industries for Automation:
- Travel: Software can now book, rebook, and manage travel itineraries directly without human agents.
- Sales: Potential shift from charging per seat (Salesforce) to selling completed outcomes, such as generating a specific number of qualified leads or conducting customer renewal calls.
- Manufacturing: ERPs can autonomously audit supply chains, calculate tariff exposures, and contact suppliers regarding shipping delays.
- Legal: Software can draft contracts and bill for the completion of legal work rather than maintaining a system of record.
- Healthcare: AI can conduct post-op follow-up calls, assess pain levels, and direct patients to emergency care based on symptoms.
- HR/Payroll: Systems like Workday can autonomously verify employment history, explain benefits, and manage enrollment.
- New Business Models: Startups are now scanning job boards (e.g., Craigslist) for unfilled positions (e.g., $45,000/year receptionist roles) and applying as AI agents to perform specific non-physical tasks for $20,000/year.
- Specific Portfolio Examples:
- Happy Robot: An AI agent successfully negotiated freight trucking rates between 6 AM and 2 PM, closing a deal at $735 against a target of $775.
- Salient: An AI collections agent handles debt recovery, speaking multiple languages (including Tagalog, Vietnamese, and Mandarin) and managing accounts 51+ days past due.
- Drivers of AI Labor Adoption:
- Intermittent Demand: AI solves staffing volatility issues (e.g., Black Friday retail or weather-related airline disruptions) where hiring/training humans is inefficient.
- Demoralizing Tasks: AI absorbs high-stress, abusive interactions (e.g., collections) without suffering burnout or morale issues.
- Regulatory Compliance: AI ensures strict adherence to regulations (e.g., UDAP laws) by removing human error and emotional variance in high-risk communications.
- Multilingual Scalability: AI provides instant access to languages (e.g., Farsi, Mongolian, Serbian) that are cost-prohibitive to source via human hiring pools.
- Market Expansion: The speaker notes the "compliance officer" as the second fastest-growing job in America; AI allows software companies to enter this market by offering end-to-end compliance solutions rather than just tools.
- Viability of New Ideas: AI reduces Customer Acquisition Cost (CAC) and Cost of Goods Sold (COGS) to the point where previously unviable businesses (e.g., an "Airbnb for bicycles") could theoretically become profitable.
- Forward-Looking Statement: The speaker concludes that the venture capital mission is to identify companies that will make the current software market appear small by fully capitalizing on the labor market's $13 trillion U.S. and larger global value.