In this episode, I break down how AI is actually going to hit the income statement for a handful of companies investors keep circling—lending (Better.com, Upstart), enterprise workflows (Box), customer support (Intercom), and voice AI (SoundHound). I focus on what matters in the next 12 months: AI showing up in products, cost bases, and competitive moats. Think 24/7 borrower experiences with voice, chat, and even avatars; support teams moving toward “answer it once” knowledge that gets reused forever; and voice as the lowest-friction interface for real-world transactions. Then I zoom out to the five-year arc, because that’s where Future Tech Investing gets interesting. In customer support, the battleground shifts from reactive resolution to proactive support—detect the problem, then show (or do) the solution. In lending, the claim is bigger: AI moving into home, auto, and revolving credit, with a world where consumer loan decisions asymptotically approach 100% AI-driven. And across the stack, I’m watching value migrate from chips and hyperscalers toward the application layer as enterprises absorb the behavior change of working with “AI labor.” Finally, I cut through AI-washing. My takeaway: do the homework yourself—test the product, push it hard, and look for real usage, real business value, deep differentiation, and a credible path to positive margins (token economics matter).