<p>Is the traditional SaaS model officially dead ? On this episode of BILLIONS, I’m sitting down with Matthew Fitzpatrick, the man Fortune 500 CEOs called when they didn’t know what to do with AI.</p><p>Matthew walked away from one of the most prestigious roles in tech, leading 1 000 engineers at McKinsey’s QuantumBlack Labs to lead <a href="https://invisibletech.ai/" target="_blank" rel="ugc noopener noreferrer">Invisible Technologies</a>.</p><p>Invisible is the "invisible" engine behind the AI revolution. </p><p>They don't just build software; they provide the RLHF (Reinforcement Learning from Human Feedback) and the data that trains the models the entire world is building on. </p><p>With $100M raised at a $2B+ valuation, Matthew is proving that the future isn't in selling tools, but in selling outcomes.</p><p>In this masterclass, we break down:</p><ul><li>The McKinsey Exit: Why a top AI leader "jumped ship" for a $2B startup.</li><li>The Death of SaaS: Why "Outcome-based pricing" is replacing the subscription model.</li><li>The Enterprise Gap: Why 90% of companies are failing to get AI into production.</li><li>The Scaling Laws: The truth about data bottlenecks and the future of AI training.</li><li>Process as Code: How Invisible integrates human intelligence with AI to solve "impossible" problems.</li></ul><p><br /></p><p><strong>TIMELINE </strong>: </p><p>00:00 The data bottleneck: Why Enterprise AI is currently "stuck"</p><p>01:01 Why McKinsey’s AI chief left to lead a $2B unicorn</p><p>02:33 The "Four Platforms": How Invisible actually works</p><p>05:58 SaaS vs. Outcomes: The pricing model of the future</p><p>09:19 Why the "AI Bubble" reality check is coming</p><p>15:12 The "Capability Gap" holding back the Fortune 500</p><p>22:15 RLHF &amp; Data: Building the workforce behind the major models</p><p>31:42 "Process is Code": The new architecture for billion-dollar companies</p><p>41:10 Matthew’s advice for founders: Don't just build a "wrapper"</p><p>48:20 The future of the "Invisible" empire</p><p><br /></p><p><strong>REFERENCES</strong><strong> :</strong></p><ul><li><p><a href="https://www.linkedin.com/in/mary-meeker-5823ba48/" rel="ugc noopener noreferrer" target="_blank"><strong>Mary Meeker</strong></a></p></li><li><p><a href="https://x.com/elonmusk" rel="ugc noopener noreferrer" target="_blank"><strong>Elon Musk</strong></a> <a href="https://sloanreview.mit.edu/article/beyond-the-hype-the-real-state-of-ai/" rel="ugc noopener noreferrer" target="_blank"><strong>Étude MIT Sloan</strong></a></p></li></ul><ul><li><p><a href="https://www.nber.org/papers/w31161" rel="ugc noopener noreferrer" target="_blank"><strong>Étude NBER (National Bureau of Economic Research)</strong></a></p></li><li><p><a href="https://www.bloomberg.com/news/articles/2019-03-06/mary-meeker-s-1999-internet-predictions-how-did-they-turn-out" rel="ugc noopener noreferrer" target="_blank"><strong>Article Bloomberg</strong></a></p></li><li><p><a href="https://www.mckinsey.com/" rel="ugc noopener noreferrer" target="_blank"><strong>McKinsey &amp; Company</strong></a> </p></li><li><p><a href="https://www.mckinsey.com/capabilities/quantumblack/how-we-help-clients" rel="ugc noopener noreferrer" target="_blank"><strong>Quantum Black</strong></a> </p></li></ul><ul><li><p><a href="https://www.invisible.co/" rel="ugc noopener noreferrer" target="_blank"><strong>Invisible Technologies</strong></a> </p></li><li><p><a href="https://www.swissgear.com/" rel="ugc noopener noreferrer" target="_blank"><strong>SwissGear</strong></a> </p></li><li><p><a href="https://www.ycombinator.com/" rel="ugc noopener noreferrer" target="_blank"><strong>Y Combinator</strong></a></p></li><li><p><a href="https://www.wecreateproblems.com/" rel="ugc noopener noreferrer" target="_blank"><strong>WeCP (We Create Problems)</strong></a></p></li></ul><ul><li><p><a href="https://www.databricks.com/" rel="ugc noopener noreferrer" target="_blank"><strong>Databricks</strong></a> </p></li><li><p><a href="https://www.snowflake.com/" rel="ugc noopener noreferrer" target="_blank"><strong>Snowflake</strong></a></p></li></ul><ul><li><p><a href="https://en.wikipedia.org/wiki/Jevons_paradox" rel="ugc noopener noreferrer" target="_blank"><strong>Jevons paradox</strong></a> </p></li><li><p><a href="https://en.wikipedia.org/wiki/Reinforcement_learning_from_human_feedback" rel="ugc noopener noreferrer" target="_blank"><strong>Reinforcement learning from human feedback (RLHF)</strong></a></p></li><li><p><a href="https://arxiv.org/abs/2201.11903" rel="ugc noopener noreferrer" target="_blank"><strong>Chain-of-thought reasoning</strong></a> </p></li></ul><ul><li><p><a href="https://www.revolut.com/" rel="ugc noopener noreferrer" target="_blank"><strong>Revolut</strong></a></p></li></ul>