Why $300B of SaaS Revenue Is Trapped & How the $19.9T AI Agent Economy Unlocks It

Manny identifies a critical gap in the market: existing monetization systems (seat-based or metering) don't work for AI agents that perform complete workflows and deliver outcomes rather than individual tasks. This mismatch is trapping $300B in SaaS revenue and forcing agent businesses to operate at 30-40% gross margins without proper revenue capture mechanisms.

• •Traditional seat-based and metering models are built for 10-year-old paradigms and don't fit agentic systems
• •AI agents operate on 30-40% gross margins (sometimes lower), requiring better revenue capture to maintain viable businesses
• •Paid is building the 'engine of growth' for AI agent companies to manage costs, monetization, and business operations
• •The problem emerged from Manny's direct experience scaling AI agents at Outreach with no market solution available

Manny chose London as Paid's headquarters, comparing it to Seattle 10 years ago. With 12M people versus SF's 1M, easier immigration laws, and four top computer science schools nearby, London offers superior talent density and access. The decision reflects a contrarian bet on undervalued talent markets.

• •London has 12M people vs Seattle's 1.5M and SF's 1M, dramatically increasing talent pool size
• •Easier immigration laws allow hiring from Germany, Spain, and across Europe without visa friction
• •Four top-tier computer science schools in the London area provide direct talent pipeline
• •Positioning in an 'undervalued' market like Seattle was 10 years ago creates competitive hiring advantage

After raising nearly $400M for Outreach, Manny's fundraising process was dramatically faster. EQT gave a term sheet after one conversation, Sequoia's Pat Grady committed on a single call. The key was validating market timing with GTM Fund's Max before approaching investors, using his network to confirm the idea matched current zeitgeist.

• •EQT closed with term sheet after single conversation; Sequoia committed same-day via Pat Grady call
• •First call was to GTM Fund's Max to validate 'hot vs not hot' and confirm market timing before fundraising
• •Raising $10-11M felt 'adorable' after negotiating $100M rounds, but matched actual capital needs for early validation
• •Leveraged existing Sequoia relationship from Outreach growth round and C round discussions

The market is in early 'green shoots' phase with agents evolving from deterministic workflows to high-autonomy agentic architectures. Customer support alone has five strong competitors all growing simultaneously due to bottomless demand. Agents are succeeding in semi-repetitive work requiring imagination: complex customer service, mortgage processing, car dealerships, HR compensation systems.

• •Evolution from deterministic agents (step-by-step) to agentic architectures with autonomous decision-making and tool-calling
• •LLM advancement is outpacing agent architecture development, creating sustained improvement runway
• •Five competitors in customer support space all growing fast simultaneously - sign of massive market expansion
• •Successful use cases: complex customer service (warranty decisions, upselling), mortgages, car dealerships, HR/compensation
• •Like early Stripe, many of Paid's future customers haven't been born yet - building for tomorrow's market

Paid serves two distinct segments: new agent startups with free/cheap self-service implementation, and legacy SaaS companies ($300B TAM) needing to pivot from seat-based to agent-based models. The mission is 'rescuing SaaS from the cliff' by helping them transition before growth stalls. Currently 40 customers, 80% startups by count but majority revenue from SaaS transitions.

• •Free/cheap self-service for weekend projects and new agent startups (like Global's early implementation)
• •Consulting-heavy approach for legacy SaaS companies transitioning from seat-based to agent monetization
• •Current customer base: 40 total, 80% startups by count, but majority of revenue from SaaS companies
• •Positioning as infrastructure to prevent SaaS companies from hitting growth cliff as seat model dies

Manny's core lesson from Outreach: you cannot grow out of inefficiencies - they're set in DNA from day one. Only he and his co-founder are selling to perfect the playbook before scaling to 1, 2, 20, or 100 sellers. Self-serve must be ready before scaling to become a growth engine rather than requiring manual hunting for every deal.

• •Fundamental principle: inefficiencies are 'set in stone in your DNA from the beginning' - cannot be fixed by scaling
• •Only founders selling currently to create repeatable, tight playbook before hiring any sales team
• •Self-serve platform must be fully operational before scaling to avoid dependency on manual deal hunting
• •Efficiency focus driven by experience, not capital constraints - saw firsthand that inefficient companies don't scale successfully

Paid immediately disqualifies prospects where agents aren't priority 1 or 2. For qualified leads, they engage both CTO (building agents) and CRO (monetizing them) simultaneously. The key education: agents tap into headcount budget pools, not IT budgets, enabling $500K contracts instead of $50K seat-based deals by capturing BPO contract savings and open headcount value.

• •Hard qualification filter: agents must be priority 1 or 2 on roadmap, otherwise immediate disqualification
• •Dual-stakeholder engagement required: CTO understands technology, CRO must understand new monetization model
• •Value pool shift: agents access headcount budgets and BPO contract savings, not just IT software budgets
• •Contract size transformation: $50K seat-based deals become $500K agent deals by capturing labor replacement value
• •Heavy professional services/consulting in early implementation to ensure customer wins and proper instrumentation

Manny predicts less backfilling of traditional roles, with every job redefined to operate with agents. Entire job categories will disappear while new ones emerge. His vision: a world where capitalism creates many millionaires and hundred-thousandaires rather than few billionaires, with entrepreneurs building and selling agents without traditional career paths or college grinding.

• •Traditional open roles will see dramatically less backfilling as agent capabilities expand
• •Every existing role will be redefined around agent collaboration rather than eliminated entirely
• •New category of entrepreneurs: individuals building and selling agents without traditional career paths
• •Vision of distributed wealth: 'capitalism wins' with many millionaires/hundred-thousandaires vs few billionaires
• •Democratization of business ownership through low-barrier agent creation and monetization

Paid operates with strong Palantir DNA and high-agency culture. New hires receive a problem space, not specific tasks, and must forward-deploy with customers to discover what to build. The philosophy: building for one customer who needs it guarantees finding more, making you 'less wrong' than not talking to customers. Team is AI-native, immediately using Cursor and Claude without training.

• •Strong Palantir cultural DNA with high-agency expectations for all team members
• •New hires given problem spaces, not tasks - must forward-deploy to customers and discover solutions
• •Contrarian advice: building for one customer's need is better than avoiding single-customer builds
• •Entire team is AI-native, immediately adopting Cursor and Claude Code without training required
• •Iterative AI-assisted development accelerates build speed across engineering organization

Manny is most excited about the self-service launch currently in soft beta. A PE firm CEO shared that agents deployed across portfolio companies are failing with poor returns, wrong numbers, and complaints. The next 12-24 months will transition agents from experimental exceptions to normalized deployments, with focus shifting from 'does it work?' to 'what's your agent's performance?'

• •Self-service platform in soft launch with limited pipeline, significant work remaining to iron out user experience
• •Real market feedback: large PE firm deploying agents across holdings seeing failures, poor ROI, and operational complaints
• •Current phase: working through 'hard miles' of hallucination issues, bad tests, and deployment problems
• •12-24 month vision: agents shift from experimental exception to normalized standard practice
• •Future evaluation criteria will be agent performance metrics, not whether agents work at all