Emergence Capital

Enterprise software is entering a new era: from static tools to dynamic learning systems. Introducing the Emergence Rate: our framework for evaluating AI software based on its ability to evolve with users over time. Learn more from @gordonritter : emcap.com/thoughts/the-e…

Before AI agents and CLM became commonplace in legal tech, @jakesaper was already placing bold bets on companies of the future. As General Partner at @emergencecap, he backed game-changers like Ironclad and SimpleLegal — long before the hype even existed. In the latest episode, @maryshenocarro1 sits down with Jake to unpack: 🤖 What does the future of the billable hour hold? 🔁 CLM isn't dead - just evolving 🚀 How to vet startups in the AI gold rush 🎲 Where the next billion-dollar SaaS bets are being placed Tune in to join the conversation shaping the future of legal tech. spoti.fi/4mvzo8F apple.co/4kgk1zm #LegalTech #Startups #AI #SaaS

"Finally a startup podcast where you actually learn something that matters!" That's what I set out to do when I started having discussions with founders and operators around potentially doing a podcast. Why? Because most startup podcasts are basically a highlight reel that barely scratch the surface across the 20+ topics they cover. They're entertaining - but you don't actually learn anything useful. Thus, meet the focal podcast where I go deep with some of today's best founders / operators on ONE crucial lesson they learned the hard way early on. No fluff, just the real, actionable insights you'd get if these founders were mentoring you 1-on-1. Our debut episode - launched today - tackles a critical mistake most founders make early on: “Most startups don’t focus on growth nearly early enough.” This comes from @austinh___ who knows startup growth better than almost anyone else. As a former early growth lead at @tryramp and now the co-founder and CEO of @unifygtm (>$30M raised from @ThriveCapital , @emergencecap, @OpenAI), Austin has both built a growth organization from scratch as well as helped scale one faster than almost anyone else. A few weeks ago, I had a fascinating conversation with him that changed how I think about growth at startups, including: 1. Experimentation speed is everything 2. Why you have to start with outbound from day 1 3. To win, move from cold to warm outbound 4. Why to build your growth team as early as possible 5. The ultimate startup moat: Building in public Starting tomorrow, I'll also share insights from these conversations in writing. For more, check out our conversation on the newly launched focal podcast and subscribe to my Substack for the write ups (LINK IN COMMENTS).

This essay from @emergencecap hits harder now than when it was first published. For years, GenAI models have been trained on public internet data—unstructured, noisy, and not always high-quality. It got us surprisingly far, but we’re quickly approaching the ceiling of what that data can offer. What’s next is already unfolding: proprietary data. ✨ McKinsey just launched a generative AI platform built on 150+ years of internal research, cases, and insight. ✨OpenAI has signed content licensing deals with The Wall Street Journal, The Atlantic, Vox Media, and more. These partnerships are quietly opening the second tier—where models get access to structured, filtered, high-value content. And as more of this comes online, the quality of AI output won’t just improve—it’ll jump. We’re stepping into a new innovation window. The data that used to live behind the firewall is becoming the differentiator.

A full circle moment at @emergencecap.

9/ AI success = user behavior × proprietary data × continuous learning. Don’t settle for “demo AI.” Build the perpetual algorithm.

8/ If you’re building in AI: Track your Emergence Rate. If you’re buying AI software: Ask how quickly it evolves. And if you’re investing: Look for products where customers are the secret proprietary training data.

1/ Most AI software isn’t evolving. It’s just performing. That’s why we built the Emergence Rate — a framework to measure whether AI products are actually learning and improving in the wild. 🧵 Here’s why it matters (and how to use it):

We have @alexrlevin (CEO of Regal) on the Infinite Curiosity Pod today. They've raised funding from amazing investors such as @emergencecap, @fcollective, and @homebrew. Here are 7 key moments from the episode: “Now people are going, wait a second, at 10 cents a minute, I want voice everywhere.” — On how the cost breakthrough flipped the economics of voice from expensive to desirable. “Why are we convincing everyone to try to get humans to do the right thing? Let’s just walk in and say the agent will do it.” — On how AI agents can replace human contact center actions more effectively. “Every time you build a new AI agent, you can’t just test it once. It’s not deterministic, it’s stochastic.” — On the challenge of reliably evaluating AI phone agents in real-world scenarios. “The best companies in our market will be 90% AI agents in the next 12 months.” — A prediction on the near future of AI adoption in customer service. “If somebody’s charging you a penny a minute, they’re not on the new models.” — On the quality divide between true LLM-based agents and legacy systems. “Most people are still thinking about IVAs. There’s a new technology—LLMs have changed the game.” — On widespread misconceptions about what AI phone agents actually are today. “You're just going to talk, and it's going to immediately engage and respond and know everything about you.” — On the natural language interface future enabled by AI voice agents.

Excited for the AI x Marketing Summit! Grateful to speakers like @KathleenEstreich,@meisenberg @tejasmanohar , Meaghan Moore, Jen Rapp, Kaushik Patel, Chris O'Neill, @ToobaDurraze, @jakesaper & Arjun Saksena. 📅 Apr 23–24 📍San Francisco 🔗shorturl.at/E4FMp

For many AI users, the “just use GPT-4o” mindset feels safe. Until you realize how much you’re spending on responses 𝘵𝘩𝘢𝘵 𝘥𝘪𝘥𝘯’𝘵 𝘯𝘦𝘦𝘥 𝘵𝘩𝘦 𝘱𝘰𝘸𝘦𝘳 𝘰𝘧 𝘢 𝘧𝘳𝘰𝘯𝘵𝘪𝘦𝘳 𝘮𝘰𝘥𝘦𝘭. That’s where ⚡Arcee Conductor⚡ comes in. It’s a smart model router that selects the right model for each task, automatically. Instead of sending everything to one model, Conductor evaluates the prompt and routes it to the optimal model: ✔️ Smaller, faster models for everyday queries ✔️ Larger models for tasks that require deeper reasoning. The result? ✅ Faster responses ✅ Quality where it counts. ✅ Dramatically reduced prices (we're talking 50-200x savings per prompt). In a world with more AI options than ever, Arcee Conductor helps you get the most out of each prompt, with zero guesswork. Sign up for Arcee Conductor here bit.ly/4kQ80l1 and get a free credit of ~400 million tokens (a $200 value)!

🚀 PearX 😍 DevAI When @susiewee & Edwin Zhang joined PearX, they had no product—just rare expertise & a bold vision to reinvent dev experience. Fast build. Real user feedback. Clear story. A short few days after Demo Day: $6M raised from @emergencecap , @Base10Partners Benchstrength This is what PearX is for. 💥 Apply 👇

Wonderful to support Chris Shutts and Logik.ai team and partner with @emergencecap to build 3x #CPQ winner leading to ServiceNow acquistion emcap.com/article/congra…

To Chris, Godard, and everyone at Logik.ai: Thank you for letting us be part of the ride. You’ve built something extraordinary, and we can’t wait to see what’s next.

This partnership with @servicenow means even greater resources behind Logik’s vision — and a path to scale the platform across the enterprise. The mission continues.

Huge congratulations to the entire Logik.ai team on their agreement to be acquired by @servicenow! 👏 This is a well-earned milestone — and a powerful moment for the future of enterprise CPQ. 🧵