Ryan R.

Evals aren't a technical problem - they're a product management opportunity. Hot take from Brooke Hopkins and Hamel Husain's podcast on evals that immediately resonated with my experience managing Voice AI initiatives. My top 8 takeaways: 1. **Lead with outcomes.** If your team can't articulate what "excellent" looks like for your agent, get the team aligned on what your agent does (and doesn't!) do. 2. **Evals are product management in disguise.** They're cross-functional and touch every team that drives inputs and downstream workflows (eng, marketing, sales, CS, operations, etc). 3. **Most teams underinvest.** 60-80% of your time should go toward defining and testing for success. Evals are the bottleneck for most teams. 4. **Fix upstream first.** Focus on the earliest error in a trace, especially in cascading architectures like voice stacks. 5. **n-1 testing is naive.** Reproducing errors in fewer turns reduces complexity and gives you a cleaner signal. 6. **Sims with states get expensive fast.** Managing evals at scale means balancing speed, latency, signal, and cost. 7. **Evals are a spectrum.** You want "hill climbing" problems just above and below current capabilities, plus happy path and adversarial cases for both existing wins and aspirational paths. 8. **Criteria drift is real.** Users adapt expectations based on capabilities they experience. Approached right, this is a feature, not a flaw! High yield conversation for just 30 minutes, and spot on from what I'm seeing at the frontlines of building and deploying enterprise Voice AI. I'm looking forward to reading Hamel's blog and scope his evals course. 🔗 Link to podcast in comments

12

1 Comment

Very much agree with this strategy. You can't fix "Everything"at once, you have to work through the most important devices and applications in order. One Battle After Another, if you like. Is there an Oscar yet for cryptographic discovery? @BlueMeshSolutions has our own process named {CODE}. We follow the same philosophy. Thanks for your information Brian C.

6

1 Comment

Yesterday Clay hit $100M ARR and I shared a post on the GTM bets that got us here. Today, I'm sitting down with our Head of Finance to dive deep on the business - metrics, our view on the AI market, and the path to $1B. Karan often gets this question from candidates: "Look, I have to dedicate my livelihood to Clay. All the upside of my equity is baked into one asset. How should I frame the real risk/reward?" It's a great question! So I wanted to use this video to go beyond the tactics and share how we think about the business holistically. The fundamentals, market dynamics and where we're really going. Some of what we cover: - The AI market and why GTM is actually the least crowded - Cohort data that shocked even us (enterprises never churn!) - How we've built an entire economy around Clay - The three things that create our customer flywheel - Why we're not burning cash while growing at an insane rate - The vectors to $1B: new geos, personas, use cases, channels, and verticals At the end, we get candid about the real risks and why World of Warcraft helps explain them. Watch here 👇

460

44 Comments

Your Team's Attention Is the Actual Bottleneck, Not Your Data Every private lending org runs into the same ceiling eventually: reps can only make so many calls in a day, marketing has a fixed budget, and leadership only has so much bandwidth to review the pipeline. Given that ceiling, "how do we find everyone who could possibly borrow" is the wrong question to be optimizing for. You already know approximately who could borrow. That list has existed for years. The question that actually determines revenue is: where does the next unit of attention go? That reframe changes what "good data" even means. A system that hands your team 5,000 more qualified names hasn't solved anything if your reps can only work 200 of them this week. What actually helps is a system that tells you which 200 to work first, and, just as importantly, which names to leave alone until the timing changes. Attention is the scarce resource in this business. Everything else is just infrastructure around it. Full breakdown – link below 👇

2

1 Comment

Brett Queener (Partner at Bonfire Ventures) joins GTMnow to share what three decades across Siebel, early Salesforce, co-founding, and seed-stage investing has taught him about what actually wins now that software is cheaper and faster than he ever imagined. Brett was one of the earliest GTM hires at Salesforce when it had seven employees. He helped build the go-to-market playbook that defined an entire generation of SaaS. Now, he's investing at pre-seed and seed out of Bonfire, watching that same playbook get dismantled in real time, and placing bets on the founders willing to throw it out. In this episode, we cover: • Why "the king of explaining" is no longer a competitive advantage • How AI collapses the traditional SaaS sales motion from SDR to CSM • Why your right to win becomes table stakes every 30 days • What Brett looks for in founders and why he calls the best ones "psychos" • Why vertical software is more defensible than ever in the AI era • Why face-to-face still matters when buyers are betting their careers • How the Salesforce and HubSpot ecosystems could become their own boat anchors • Why events create 75%+ of pipeline for his early-stage portfolio companies • What "deploy before you close" means in practice • Why the functional GTM leader of today has to be a builder not just a smooth talker • What changes in VC when software is no longer scarce to build • Why seven out of ten bets should die and what that means for seed-stage investing Thank you to our VC series partner AngelList, who have been instrumental in helping GTMfund scale and evolve since day one. The episode is available through the links in the comments, or wherever you get your podcasts 🎧

9

1 Comment