AI Builders

Episode

• Why EverWorker targets "boring billion-dollar companies" | Anton Antich 07.04.2026 21mnt

  Most AI companies in 2023 raced to own a vertical. EverWorker made the opposite bet — build a horizontal platform that lets anyone create agents for any purpose, no code required. In this episode of BUILDERS, ⁠Anton Antich⁠, CPO and Co-Founder of ⁠EverWorker⁠, gets into what it actually takes to sell AI inside enterprises that are stuck between the hype and the reality, why he's making the case against SaaS entirely, and how an early PLG motion gave way to deep consultative selling once they realized the market wasn't where Silicon Valley thought it was. Anton helped scale a company from $0 to $1B ARR — and he's direct that most of what he learned there no longer applies.Topics Discussed:Why the $0–$1B scaling playbook is obsolete in the AI eraEverWorker's pivot from PLG to enterprise consultative sellingTargeting "boring billion-dollar companies" as a deliberate ICPWhy most AI pilots never reach production — and the services motion that fixes itThe org-chart model for AI agent teams and the "Chief of Staff" productThe case for replacing SaaS entirely with agents, databases, and markdown filesGoing down-market in 2026 and why community is the lead growth channelWhy instant product access has replaced "contact us for a demo" as the conversion standardGTM Lessons For B2B Founders:Audit your team's DNA before choosing your GTM motion. EverWorker launched PLG, then quickly realized their entire founding team came from enterprise — Microsoft, VMware, Veeam. The pivot wasn't a failure; it was an honest read of where their unfair advantages actually lived. Before committing to a motion, map your team's network, sales instincts, and domain depth. Those signals will outperform market trend-chasing every time.Build a services layer or watch your pilots die. The gap between AI pilot and production is where most deals go to die — Anton cites the widely-reported stat that the vast majority never make it through. EverWorker's solution was to build a services organization that identifies two or three mundane, high-friction processes — Anton's example is data entry, work humans find demeaning and AI handles well — automates them fast, and uses that visible win to build organizational trust. The services layer isn't a concession. For complex AI sales right now, it's the mechanism that actually converts pilots into production.Your ICP should be defined by who won't default to "we'll build it ourselves." EverWorker learned this the hard way in enterprise. Walk into a Fortune 500 or a tech-forward company and IT shows up in the room and kills the conversation. Anton's team shifted toward what he calls "boring billion-dollar companies" — industries doing real, essential work that don't get the spotlight and can't afford to staff AI expertise internally. These buyers need the outcome, not the platform, and they don't have an internal team to rationalize building around. That dynamic is a structural GTM advantage.// Sponsors: Front Lines — We help B2B tech companies launch, manage, and grow podcasts that drive demand, awareness, and thought leadership.⁠ www.FrontLines.io⁠The Global Talent Co. — We help tech startups find, vet, hire, pay, and retain amazing marketing talent that costs 50-70% less than the US & Europe.⁠ www.GlobalTalent.co⁠//Don't Miss: New Podcast Series — How I Hire Senior GTM leaders share the tactical hiring frameworks they use to build winning revenue teams. Hosted by Andy Mowat, who scaled 4 unicorns from $10M to $100M+ ARR and launched Whispered to help executives find their next role. Subscribe here:⁠ https://open.spotify.com/show/53yCHlPfLSMFimtv0riPyM⁠
• How Yutori landed enterprise contracts without a sales team by letting prosumer word of mouth do the work | Abhishek Das, Co-CEO at Yutori 02.04.2026 18mnt

  Yutori⁠ is building web agents — AI that can monitor, navigate, and eventually act on the web on your behalf. Their first product, Scouts, launched in beta in June 2024 with one deliberate constraint: read-only web monitoring. No booking, no form-filling, no write actions. Just signal extraction from the open web. That narrow framing, paired with a $25K launch video that went viral on Twitter, drove 20–30K waitlist signups in a single week. M1 retention held above 80%. Enterprise contracts followed — entirely bottom-up, entirely unsolicited. In this episode of Unicorn Builders, Co-CEO Abhishek Das breaks down the thinking behind all of it.Topics Discussed:Why scoping Scouts to read-only monitoring at launch was a GTM decision, not just a product oneThe $25K launch video that went viral — what was in it and why it workedHow unsolicited enterprise contracts emerged from a prosumer productRunning two parallel GTM motions simultaneously with no dedicated marketing teamHow hackathons became a developer acquisition channelThe browser automation API: a separate product with a separate motion, and why the two audiences cross-pollinateWhat's next: authenticated browsing and write-action agents currently in alphaGTM Lessons For B2B Founders:Constrain your launch scope to match what you can actually deliver. The AI agent space is full of products that promise to do everything and fail at anything. Yutori's answer was the inverse: launch Scouts as read-only monitoring only — no purchasing, no reservations, no form submissions. Abhishek was explicit that this was intentional: lower stakes for errors, a cleaner value prop, and a more honest promise to early users. The constraint wasn't a limitation — it was the pitch. If you're launching in a crowded category where trust is already eroded, scoping tightly is a competitive move.Let retention data — not your roadmap — trigger monetization. Scouts launched free with no fixed plan to charge. When M1 retention held above 80%, the team pulled their monetization timeline forward and shipped a flat monthly subscription. No elaborate pricing research, no staged rollout. The data gave them the signal. For founders debating when to introduce pricing: retention is the clearest leading indicator that your product has earned the right to charge. Set a retention threshold before you launch, and let it make the call for you.A $25K launch video beat the market — because the message did the work. The video was Abhishek on camera, directly explaining what Scouts can and cannot do. No cinematic production. It went viral because prominent builders — Guillermo Rauch from Vercel, Scott Belsky — reshared it organically. Abhishek is candid that going viral involves luck and that Twitter feels significantly more saturated today than it did at launch. The takeaway isn't "spend $25K on a video." It's that precise articulation travels further than high production value, and distribution through trusted voices matters more than raw reach.// Sponsors: Front Lines — We help B2B tech companies launch, manage, and grow podcasts that drive demand, awareness, and thought leadership.⁠ www.FrontLines.io⁠The Global Talent Co. — We help tech startups find, vet, hire, pay, and retain amazing marketing talent that costs 50-70% less than the US & Europe.⁠ www.GlobalTalent.co⁠//Don't Miss: New Podcast Series — How I Hire Senior GTM leaders share the tactical hiring frameworks they use to build winning revenue teams. Hosted by Andy Mowat, who scaled 4 unicorns from $10M to $100M+ ARR and launched Whispered to help executives find their next role. Subscribe here:⁠ https://open.spotify.com/show/53yCHlPfLSMFimtv0riPyM⁠
• How Cassidy achieved 90% content performance consistency across TikTok and Instagram | Justin Fineberg 04.03.2026 15mnt

  Justin Fineberg⁠ built a 500,000+ follower audience on TikTok and Instagram before launching ⁠Cassidy⁠, an AI automation platform for non-technical users. By consistently creating content about AI and technology, he turned inbound interest into his initial customer base and market validation. In this episode of BUILDERS, Justin breaks down how he leveraged short-form video to identify product opportunities, the mechanics of maintaining authentic audience relationships while monetizing, and how to transition from social-led distribution to scalable B2B SaaS go-to-market.Topics Discussed:Leveraging ChatGPT's launch as an inflection point to ride mainstream AI interestConverting consultant requests into product insights and early customer signalsThe platform mechanics of TikTok vs Instagram for B2B contentTransitioning from 100% social-sourced revenue to multi-channel B2B salesBuilding repeatable content systems that survive founder time constraintsTesting product messaging and features through content before formal launchGTM Lessons For B2B Founders:Timing content focus with market inflection points compounds growthInbound consulting requests are product requirement documents in disguiseContent systems must be friction-free or they'll die under operational loadGood content transcends platform-specific algorithm hackingSocial distribution creates unfair launch advantages, not permanent moats//Sponsors:Front Lines — We help B2B tech companies launch, manage, and grow podcasts that drive demand, awareness, and thought leadership.⁠ www.FrontLines.io⁠The Global Talent Co. — We help tech startups find, vet, hire, pay, and retain amazing marketing talent that costs 50-70% less than the US & Europe.⁠ www.GlobalTalent.co⁠//Don't Miss: New Podcast Series — How I HireSenior GTM leaders share the tactical hiring frameworks they use to build winning revenue teams. Hosted by Andy Mowat, who scaled 4 unicorns from $10M to $100M+ ARR and launched Whispered to help executives find their next role. Subscribe here:⁠ https://open.spotify.com/show/53yCHlPfLSMFimtv0riPyM
• How Positron AI is driving sales ahead of product | Mitesh Agrawal 20.02.2026 26mnt

  Positron AI is a 2+ year old silicon company targeting decode-heavy AI inference workloads where memory bandwidth, not compute, is the bottleneck. Launching end of 2025/early 2026, their architecture delivers 2TB of on-chip memory capacity versus Nvidia Rubin's 0.4TB—enabling 3-5x better performance per dollar and per watt for reasoning models, code generation, and video generation. In this episode, ⁠Mitesh Agrawal⁠ shares how ⁠Positron⁠ identified the memory bandwidth gap in a market where Nvidia controls 90%+ share, why they're prioritizing anchor customer commitments over product completion, and the hard lessons from Lambda Labs about rapid iteration and customer-driven optionality.Topics Discussed:Positron's technical approach: focusing on memory bandwidth and capacity over compute for inference workloadsWhy decode-heavy applications (reasoning models, video generation, code generation) are becoming memory-boundThe challenge of selling silicon to hyperscalers when Nvidia controls 90%+ of the marketBuilding optionality into product strategy: air cooling vs. liquid cooling as unexpected GTM advantageLearning to sell hardware before the product ships and why anchor customers matterLambda Labs experience: lessons on rapid iteration and thoughtful hiring during hypergrowthMaintaining engineering-centricity: 47 of 50 employees focused on product developmentGTM Lessons For B2B Founders:Find technical bottlenecks in high-growth markets: Positron identified that memory bandwidth wasn't scaling as fast as compute, creating a bottleneck for inference workloads. While Nvidia dominates with 90%+ market share, they optimize for training revenue. B2B founders should analyze where dominant players are constrained by their own economics or existing roadmaps, then build specifically for those underserved segments.Markets default to oligopoly, not monopoly: Mitesh observed that customers actively seek alternatives even when one vendor is superior. "Markets want oligopoly structure to exist," he explained. B2B founders shouldn't be discouraged by dominant incumbents—customers want optionality for leverage, supply chain resilience, and risk management. Position yourself as the credible alternative in specific use cases.Discover optionality through customer conversations: Positron initially pitched performance per watt without realizing air cooling capability was a major advantage. Only after selling their first product did they learn customers valued deploying in existing data centers without infrastructure overhauls. B2B founders should systematically debrief early customers to uncover which features solve problems you didn't anticipate.Sell before shipping in hardware: The biggest priority between now and product launch is securing anchor customers willing to commit purchase orders. "If you have someone to build for, the fillip it gives the engineering team, the confidence it gives operations and supply chain vendors—we underwrite that," Mitesh emphasized. Pre-sales derisk production, prove demand, and create momentum. Build storytelling into technical sales: Convincing customers to buy unshipped hardware requires months of narrative work. "It becomes like, if I sell it to you, why will it be useful to you? Is it going to save cost? Attract new customers? Drive growth?" Success means co-creating the internal business case your champion will present. Maintain rapid iteration cadence: Nvidia ships every 12-15 months versus the industry standard of 3-4 years. "If you tell me that in 10 years you've launched 10-12 products in silicon, I will give much more probability we will be successful," Mitesh stated. Delay non-engineering hires until product proves itself: With 47 of 50 people in engineering, Positron has consciously prioritized product over go-to-market. "It was a very conscious decision," Mitesh emphasized. For deep-tech companies, this focus ensures you can actually deliver before scaling sales.
• Why aiOla targets CFOs — not IT buyers | Amir Haramaty, Co-Founder at aiOla 28.01.2026 28mnt

  aiOla is pioneering speech-to-data technology that transforms unstructured speech into actionable data for enterprise operations. As a serial entrepreneur on his sixth startup, Co-Founder ⁠Amir Haramaty⁠ built ⁠aiOla⁠ after witnessing firsthand how traditional AI implementations fail to deliver ROI in enterprise settings. The company has developed proprietary technology that achieves near-100% accuracy in challenging environments with heavy jargon, multiple languages, and difficult acoustics. With strategic investors including a major airline and partnerships with Nvidia, Accenture, and USG, aiOla is addressing the fundamental challenge that 95% of enterprise AI pilots fail to show value by focusing on immediate, measurable ROI through speech-based data capture.Topics Discussed:The genesis of aiOla from consulting work revealing AI's implementation gaps in traditional enterprisesSolving the triple challenge of speech recognition: accuracy in jargon-heavy environments, separating signal from noise, and converting speech to structured workflow dataaiOla's "jargonic" approach: creating hyper-personalized language models for specific processes without retrainingEarly customer acquisition through serendipitous encounters and demonstrating immediate ROIVertical expansion strategy from food manufacturing to aviation, travel, hospitality, and retailChannel partnership strategy refined from previous startups to achieve scaleThe shift from convincing customers about speech technology to being pulled into diverse use casesBuilding the aiOla Intelligate orchestration layer to dynamically select optimal speech recognition modelsGTM Lessons For B2B Founders:Make CFOs your best friend, not IT departments: Amir explicitly targets CFOs rather than IT as primary buyers because "it doesn't matter how small or big you are, you still have to do more with less." While IT serves as facilitators, CFOs control budgets focused on operational efficiency and ROI. B2B founders should identify which executive truly owns the pain point and budget authority, even if IT will implement the solution.Deploy capital strategically to remove obstacles before they emerge: aiOla convinced their airline investor to provide working capital specifically to fund POCs for prospects without existing budgets. This eliminated the "we don't have pilot budget" objection before it arose. B2B founders should proactively identify and neutralize common barriers in their sales process, whether through creative deal structures, proof-of-concept funding, or implementation support.Prioritize instant ROI over long-term transformation promises: Amir explicitly avoids "digital transformation" conversations, instead selecting use cases delivering "biggest impact within shortest period of time with minimum obstacle possible." The airline baggage tracking example saved 110,000 hours immediately, creating momentum for expansion. B2B founders should resist selling comprehensive transformation and instead identify narrow use cases with quantifiable, rapid returns that create internal champions.Replicate proven use cases across customers rather than customizing: Once aiOla achieved success with specific applications like CRM data entry or pre-op inspections, they "stop, print, replicate" rather than reinventing for each customer. This approach reduced a two-hour inspection process to 34 minutes in food manufacturing, then replicated across industries. B2B founders should document successful implementations as repeatable playbooks and resist the urge to over-customize for each prospect.Channel success requires speaking the partner's economic language: When working with telcos, Amir demonstrated that his solution increased ARPU by 34% and reduced churn by 17%—the only two metrics telcos prioritize. He built predictable models showing exactly how many units each channel rep would sell by geography.
• How Parable achieved a 100% POC win rate in enterprise AI sales | Adam Schwartz 16.01.2026 24mnt

  Parable⁠ is building an end-to-end intelligence platform that quantifies how organizations spend their collective time—the foundation for measuring real AI impact. With a thousand data connectors ingesting activity and log data across the enterprise software stack, Parable constructs proprietary knowledge graphs that size opportunities and measure outcomes in hard dollars, not adoption metrics. In this episode of BUILDERS, I sat down with ⁠Adam Schwartz⁠, Co-Founder & CEO of Parable, to explore why 95% of CFOs see no AI ROI, how his decade running profitable businesses under resource constraints shaped his focus on inputs over outcomes, and why 2026 requires moving AI from CapEx experimentation to measured OpEx.Topics Discussed:Why the 95% CFO stat on AI ROI matters as an arbiter of truth, despite backlashBuilding knowledge graphs from activity data to quantify collective time allocation across hundreds of peopleThe fundamental problem: enterprises lack quantitative frameworks for operational efficiency pre-AIRunning parallel ICP experiments to achieve sales-market fit before product-market fitWhy Parable has never lost a POC once leaders see quantitative baselinesMarket dynamics creating false signals—unprecedented curiosity without buying intentThe demarcation between companies treating AI as product work versus those waiting for vendor solutionsWhy AI transformation demands century-old management structures to be questionedGTM Lessons For B2B Founders:Engineer disqualification in momentum markets: Market-wide AI enthusiasm creates pipeline illusion. Prospects will engage indefinitely for education without purchase intent. Adam's framework: "How do we get people to say no to us and not drag us along... They want to keep talking because they want to learn and they want to know what's going on and they are genuinely interested." Use go-to-market as ICP discovery mechanism: Adam intentionally pursued multiple customer segments simultaneously—different company sizes and AI maturity stages—to let data reveal fit rather than rely on hypothesis. His memo to the team: "We're going to go after these three, you know, many different sizes of companies in order for us to decide like, who we like best." Qualify on organizational structure, not verbal commitment: Every enterprise claims AI is strategic. Adam's hard filter: "Who in the organization is responsible for AI transformation? And if you don't have a one person answer to that question, you're not serious." Serious buyers have a named owner reporting to C-suite with dedicated budget and team. Buying Gemini, Glean, or other point solutions isn't a seriousness KPI—it's often passive consumption of AI as a byproduct of existing software relationships. Target post-experimentation, pre-scale buyers: Adam discovered the sweet spot isn't companies beginning their AI journey—it's those who've deployed initial programs and now need to prove value. "The market of people that have started to build AI into their operating model or into their strategy in like a coherent way, there's a team, there's an owner, there's budget... those are the people that we really want to be talking to." These buyers understand the problem viscerally because they're living it. They do product work daily—talking to stakeholders, generating use cases, building briefs, triaging roadmaps. Build measurement into your category narrative: The AI tooling market has over-indexed on soft efficiency claims that won't survive renewal cycles. Adam's warning: "There is too much hand waving around soft efficiency gains... you're going to have to renew and you need NRR and I don't think it's going to be that usage of the tool internally by employees and adoption is going to be enough." The last decade over-rotated to "everything drives revenue" due to VC pressure. This decade requires precision: does your product save time, reduce headcount needs, or accelerate revenue?
• How Datawizz discovered the chasm between AI-mature companies and everyone else shaped their ICP | Iddo Gino 18.12.2025 29mnt

  Datawizz⁠ is pioneering continuous reinforcement learning infrastructure for AI systems that need to evolve in production, not ossify after deployment. After building and exiting RapidAPI—which served 10 million developers and had at least one team at 75% of Fortune 500 companies using and paying for the platform—Founder and CEO ⁠Iddo Gino⁠ returned to building when he noticed a pattern: nearly every AI agent pitch he reviewed as an angel investor assumed models would simultaneously get orders of magnitude better and cheaper. In a recent episode of BUILDERS, we sat down with Iddo to explore why that dual assumption breaks most AI economics, how traditional ML training approaches fail in the LLM era, and why specialized models will capture 50-60% of AI inference by 2030.Topics Discussed:Why running two distinct businesses under one roof—RapidAPI's developer marketplace and enterprise API hub—ultimately capped scale despite compelling synergy narrativesThe "Big Short moment" reviewing AI pitches: every business model assumed simultaneous 1-2 order of magnitude improvements in accuracy and costWhy companies spending 2-3 months on fine-tuning repeatedly saw frontier models (GPT-4, Claude 3) obsolete their custom workThe continuous learning flywheel: online evaluation → suspect inference queuing → human validation → daily/weekly RL batches → deploymentHow human evaluation companies like Scale AI shift from offline batch labeling to real-time inference correction queuesEarly GTM through LinkedIn DMs to founders running serious agent production volume, working backward through less mature adoptersICP discovery: qualifying on whether 20% accuracy gains or 10x cost reductions would be transformational versus incrementalThe integration layer approach: orchestrating the continuous learning loop across observability, evaluation, training, and inference toolsWhy the first $10M is about selling to believers in continuous learning, not evangelizing the categoryGTM Lessons For B2B Founders:Recognize when distribution narratives mask structural incompatibility: RapidAPI had 10 million developers and teams at 75% of Fortune 500 paying for the platform—massive distribution that theoretically fed enterprise sales. The problem: Iddo could always find anecdotes where POC teams had used RapidAPI, creating a compelling story about grassroots adoption. Qualify on whether improvements cross phase-transition thresholds: Datawizz disqualifies prospects who acknowledge value but lack acute pain. The diagnostic questions: "If we improved model accuracy by 20%, how impactful is that?" and "If we cut your costs 10x, what does that mean?" Companies already automating human labor often respond that inference costs are rounding errors compared to savings. Use discovery to map market structure, not just validate hypotheses: Iddo validated that the most mature companies run specialized, fine-tuned models in production. The surprise: "The chasm between them and everybody else was a lot wider than I thought." . Target spend thresholds that indicate real commitment: Datawizz focuses on companies spending "at a minimum five to six figures a month on AI and specifically on LLM inference, using the APIs directly"—meaning they're building on top of OpenAI/Anthropic/etc., not just using ChatGPT. Structure discovery to extract insight, not close deals: Iddo's framework: "If I could run [a call where] 29 of 30 minutes could be us just asking questions and learning, that would be the perfect call in my mind." He compared it to "the dentist with the probe trying to touch everything and see where it hurts." Avoid the false-positive trap in well-funded categories: Iddo identified a specific risk in AI: "You can very easily run these calls, you think you're doing discovery, really you're doing sales, you end up getting a bunch of POCs and maybe some paying customers. So you get really good initial signs but you've never done any actual discovery.
• How Wisdom AI reduces enterprise trial time-to-value from weeks to minutes | Soham Mazumdar 14.11.2025 18mnt

  Wisdom AI⁠ sells to enterprise data teams, empowering them to deploy AI data analysts that automate analytics functions traditionally handled by human analysts. As a former Rubrik co-founder and Google search ranking engineer, Soham identified the analytics problem firsthand while scaling Rubrik from intuition-driven to data-driven operations. In this episode of Category Visionaries, ⁠Soham⁠ shares how four Rubrik alumni are building a category-defining solution in the data analytics space, the tactical insights from targeting mid-market accounts to optimize deal velocity and onboarding experience, and how AI buying committees shifted from experimental budgets in 2024 to gatekeepers requiring departmental champions in 2025.Topics Discussed:Leveraging mid-market focus to compress sales cycles while refining onboarding as core product differentiationThe transition from gut-based decisions to data-driven operations and why analytics remains unsolvedTaming LLMs for precision and explainability requirements in enterprise analytics contextsStrategic navigation of the data ecosystem following the FiveTran-DBT merger and positioning against Snowflake, Databricks, and cloud providersOverlaying product-led trial motions on enterprise sales to maintain momentum during extended procurement cyclesAI committee evolution from 2024's experimental phase to 2025's security-focused consolidation mandatePursuing 10x productivity gains versus incremental improvement in established analytics marketsGTM Lessons For B2B Founders:Use mid-market to build onboarding velocity as moat: Rubrik deliberately targeted mid-market accounts despite being an enterprise product that closed eight-figure deals. This served two strategic purposes: compressed sales cycles enabled faster learning loops, and the necessity of quick onboarding forced the team to build exceptional admin experiences that became their primary differentiation. Find problems through operational scar tissue, not market research: Wisdom AI originated when Soham tried moonlighting as engineering's data analyst during Rubrik's scaling phase and discovered he couldn't do it effectively. This wasn't a customer interview insight—it was firsthand recognition that even sophisticated technical leaders with dedicated focus couldn't wrangle data for operational decisions. The problem proved ubiquitous across every business leader optimizing top line, bottom line, and operations. Engineer time-to-value in minutes for PLG overlay on enterprise sales: Wisdom AI's experiential quality—users get excited when they try it, not when they see slides—creates PLG opportunity despite enterprise positioning. The critical difference: sales-led motions tolerate weeks to first value and build confidence through process, but self-serve requires hook-to-value in minutes with zero support. Soham's insight is using PLG not for credit card swipes but to maintain champion enthusiasm during lengthy procurement processes. Treat ecosystem navigation as first-class GTM workstream: Wisdom AI's success depends on partnership execution with Snowflake, Databricks, and cloud providers—all potential competitors with their own AI initiatives. The FiveTran-DBT merger created immediate dynamic shifts requiring repositioning. Rather than viewing partnerships as business development, Soham frames ecosystem navigation as core GTM infrastructure requiring dedicated strategy and repeatable playbooks. Architect for AI committee gatekeepers with departmental executive sponsorship: The market fundamentally shifted from mid-2024's "experimental AI budgets, try everything" to 2025's centralized AI committees focused on security, tool consolidation, and preventing organizational wild west scenarios. Soham's tactical response: secure champions owning specific important departments who can navigate approval hierarchies while trial experiences maintain grassroots excitement.
• How TwelveLabs sells AI to federal agencies: Mission alignment over process optimization | Jae Lee 15.10.2025 21mnt

  TwelveLabs is building purpose-built foundation models for video understanding, enabling enterprises to index, search, and analyze petabytes of video content at scale. Founded by three technical co-founders who met in South Korea's Cyber Command doing multimodal video understanding research, the company recognized early that video requires fundamentally different infrastructure than text or image AI. Now achieving 10x revenue growth and serving customers across media, entertainment, sports, advertising, and federal agencies, TwelveLabs is proving that category creation through extreme focus beats trend chasing. In this episode, Jae Lee shares how the company navigated early product decisions, built specialized GTM motions for established industries, and maintained technical conviction during years of building in relative obscurity.Topics Discussed:How military research in multimodal video understanding led to founding TwelveLabs in 2020 The technical thesis: why video deserves purpose-built foundation models and inference infrastructure Targeting video-centric industries where ROI justifies early-stage pricing: media, entertainment, sports, advertising, and defense Partnership-driven distribution strategy and AWS Bedrock integration results Specialized sales approach: generalist leaders, vertical-specific AEs and solutions architects Maintaining extreme focus and avoiding hype cycles during the first three years of building Federal GTM lessons: why In-Q-Tel partnership and authentic mission alignment matter more than process optimization The discipline of saying no to large opportunities that don't fit ICP Keeping hiring bars high when the entire team is underwaterGTM Lessons For B2B Founders:Hire vertical specialists on the front lines, not just at the top: TwelveLabs structures its GTM team with generalist leaders (head of GTM and VP of Revenue) who can sell any technology, but vertical-specialized AEs, solutions architects, and deployment engineers. These front-line team members come directly from the four target industries and understand customer workflows, buying patterns, and integration points without ramp time. Infrastructure plays require integration partnerships, not displacement: In established industries with layered technology stacks, positioning as foundational infrastructure demands partnership-first distribution. Jae explained their approach: integration with media-specific GSIs, media asset management platforms, and cloud providers ensures TwelveLabs fits into existing workflows rather than forcing wholesale replacement. Extreme focus on first-principles product development beats fast-follower tactics: While competitors built quick demos by wrapping existing models, TwelveLabs spent three years building proprietary video foundation models and indexing infrastructure from scratch. Jae was explicit about the cost: "It was painful journey in the first like two and a half, three years because folks are flying by." The payoff came from solving actual customer problems—indexing 2 million hours of content in two days, enabling semantic search at scale, building agent workflows for specific use cases. Federal requires cultural alignment before GTM optimization: TwelveLabs' federal success stems from authentic mission alignment, not just process execution. With In-Q-Tel as an investor providing interface to agencies and founders with military backgrounds, the company established credibility through shared values rather than sales tactics. ICP discipline protects product focus and team morale: Saying no to large early opportunities that don't fit ICP is operationally painful but strategically essential. Jae acknowledged the difficulty: "Early on saying no to customers is hard... as a founder you want to grow your business and you know that's going to be good for the morale. But that's only true when the customers are actually their ideal customers."
• How Freeplay built thought leadership by triangulating insights across hundreds of AI implementations | Ian Cairns 15.10.2025 28mnt

  Freeplay AI emerged from a precise timing insight: former Twitter API platform veterans Ian Cairns and Eric Schade recognized that generative AI created the same platform opportunity they'd previously captured with half a million monthly active developers. Their company now provides the observability, evaluation, and experimentation infrastructure that lets cross-functional teams—including non-technical domain experts—collaborate on AI systems that need to perform consistently in production.Topics Discussed:Systematic customer discovery: 75 interviews in 90 days using jobs-to-be-done methodology to surface latent AI development pain pointsCross-functional AI development: How domain experts (lawyers, veterinarians, doctors) became essential collaborators when "English became the hottest programming language"Production AI reliability challenges: Moving beyond 60% prototype success rates to consistent production performanceEnterprise selling to technical buyers: Why ABM and content worked where ads and outbound failed for VPs of engineeringCategory creation without precedent: Building thought leadership through triangulated insights across hundreds of implementationsOffline community building: Growing 3,000-person Colorado AI meetup with authentic "give first" approachGTM Lessons For B2B Founders:Structure customer discovery with jobs-to-be-done rigor: Ian executed a systematic 75-interview program in 90 days, moving beyond surface-level feature requests to understand fundamental motivations. Using Clay Christensen's framework, they discovered engineers weren't just frustrated with 60% AI prototype reliability—they were under career pressure to deliver AI wins while lacking tools to bridge the gap to production consistency. This deeper insight shaped Freeplay's positioning around professional success metrics rather than just technical capabilities.Exploit diaspora networks from platform companies: Twitter's developer ecosystem became Ian's customer research goldmine. Platform company alumni have uniquely valuable networks because they previously interfaced with hundreds of technical teams. Rather than cold outreach, Ian leveraged existing relationships and warm introductions to reach heads of engineering who were actively experimenting with AI. This approach yielded higher-quality conversations and faster pattern recognition across use cases.Target sophistication gaps in technical buying committees: Traditional SaaS tactics failed because Freeplay's buyers—VPs of engineering at companies building production AI—weren't responsive to ads or generic outbound. Instead, Ian invested in deep technical content (1500-2000 word blog posts), speaking engagements, and their "Deployed" podcast featuring practitioners from Google Labs and Box. This approach built credibility with sophisticated technical audiences who needed education about emerging best practices, not product demos.Build authority through cross-portfolio insights: Rather than positioning as AI experts, Ian built trust by triangulating learnings across "hundreds of different companies" and sharing pattern recognition. Their messaging became "don't just take Freeplay's word for it—here's what we've seen work across environments." This approach resonated because no single company had enough AI production experience to claim definitive expertise. Aggregated insights became more valuable than individual case studies.Time market entry for the infrastructure adoption curve: Ian deliberately positioned Freeplay for companies "3, 6, 12 months after being in production" rather than competing for initial AI experiments. They recognized organizations don't invest in formal evaluation infrastructure until they've proven AI matters to their business. This patient approach let them capture demand at the moment companies realized they needed serious operational discipline around AI systems.
• How Cerebrium generated millions in ARR through partnerships without a sales team | Michael Louis 29.09.2025 24mnt

  Cerebrium is a serverless AI infrastructure platform orchestrating CPU and GPU compute for companies building voice agents, healthcare AI systems, manufacturing defect detection, and LLM hosting. The company operates across global markets handling data residency constraints from GDPR to Saudi Arabia's data sovereignty requirements. In a recent episode of Category Visionaries, I sat down with Michael Louis, Co-Founder & CEO of Cerebrium, to explore how they built a high-performance infrastructure business serving enterprise customers with high five-figure to six-figure ACVs while maintaining 99.9%+ SLA requirements.Topics Discussed:Building AI infrastructure before the GPT moment and strategic patience during the hype cycleScaling a distributed engineering team between Cape Town and NYC with 95% South African talentPartnership-driven revenue generation producing millions in ARR without traditional sales teamsAI-powered market engineering achieving 35% LinkedIn reply rates through competitor analysisTechnical differentiation through cold start optimization and network latency improvementsRevenue expansion through global deployment and regulatory compliance automationGTM Lessons For B2B Founders:Treat go-to-market as a systems engineering problem: Michael reframed traditional sales challenges through an engineering lens, focusing on constraints, scalability, and data-driven optimization. "I try to reframe my go to market problem as an engineering one and try to pick up, okay, like what are my constraints? Like how can I do this, how can it scale?" This systematic approach led to testing 8-10 different strategies, measuring conversion rates, and building automated pipelines rather than relying on manual processes that don't scale.Structure partnerships for partner success before revenue sharing: Cerebrium generates millions in ARR through partners whose sales teams actively upsell their product. Their approach eliminates typical partnership friction: "We typically approach our partners saying like, look, you keep the money you make, we'll keep the money we make. If it goes well, we can talk about like rev share or some other agreement down the line." This removes commission complexity that kills B2B partnerships and allows partners to focus on customer value rather than internal revenue allocation conflicts.Build AI-powered competitive intelligence for outbound at scale: Cerebrium's 35% LinkedIn reply rate comes from scraping competitor followers and LinkedIn engagement, running prospects through qualification agents that check funding status, ICP fit, and technical roles, then generating personalized outreach referencing specific interactions. "We saw you commented on Michael's post about latency in voice. Like, we think that's interesting. Like, here's a case study we did in the voice space." Position infrastructure as revenue expansion, not cost optimization: While dev tools typically focus on developer productivity gains, Cerebrium frames their value proposition around market expansion and revenue growth. "We allow you to deploy your application in many different markets globally... go to market leaders love us and sales leaders because again we open up more markets for them and more revenue without getting their tech team involved." Weaponize regulatory complexity as competitive differentiation: Cerebrium abstracts data sovereignty requirements across multiple jurisdictions - GDPR in Europe, data residency in Saudi Arabia, and other regional compliance frameworks. "As a company to build the infrastructure to have data sovereignty in all these companies and markets, it's a nightmare." By handling this complexity, they create significant switching costs and enable customers to expand internationally without engineering roadmap dependencies, making them essential to sales teams pursuing global accounts.
• How OpenInfer discovered unexpected government traction by focusing on data ownership pain points | Behnam Bastani 16.09.2025 21mnt

  OpenInfer addresses the enterprise infrastructure gap that causes 70% of edge AI deployments to fail. Founded by system architects who previously built high-throughput runtime systems at Meta (enabling VR applications on Qualcomm chips via Oculus Link) and Roblox (scaling real-time operations across millions of gaming devices), OpenInfer applies proven architectural patterns to enterprise edge AI deployment. The company targets three specific customer pain points: cost reduction for AI-always-on applications, data sovereignty requirements in regulated environments, and reliability for systems that must function regardless of connectivity. In this episode of Category Visionaries, CEO and Founder Behnam Bastani reveals how external market catalysts like DeepSeek's efficiency breakthrough transformed investor perception and validated their compute optimization thesis.Topics Discussed:System architecture pattern replication from Meta's Oculus Link to Roblox to OpenInferThe compute efficiency gap: why "throwing hardware" at AI problems creates market inefficienciesHow DeepSeek's January 2025 breakthrough shifted investor sentiment from skepticism to oversubscriptionCustomer targeting methodology: focusing on business unit leaders facing career consequencesGovernment market discovery: air-gapped environments and data sovereignty requirementsTechnical demonstration strategies for overcoming the 70% edge deployment failure ratePrivacy-first AI positioning unlocking previously inaccessible use casesGTM Lessons For B2B Founders:Target decision-makers with career-level consequences: Rather than pursuing prospects who might "take a risk," Behnam focuses on "those that lose their jobs if they're not solving the problem" - specifically business unit leaders whose profit margins or sales metrics directly impact their career trajectory. This creates urgency that comfortable cloud users lack and accelerates deal cycles by aligning solution adoption with personal survival incentives.Leverage external market catalysts for thesis validation: OpenInfer initially faced investor pushback ("Nvidia's got everything working well. Why you think you can do anything better?") until DeepSeek's efficiency breakthrough provided third-party validation. "January hits and then there's DeepSeek... People called us, hey, you're DeepSeek on edge." Founders should identify potential external events that could validate their contrarian thesis and be prepared to capitalize when these catalysts occur.Lead with technical proof points over explanations: In markets with high failure rates, demonstrations eliminate skepticism faster than education. "We definitely have metrics, demos, and we go with those. We demonstrate what's possible... we remove this skepticalism in terms of ease of deployments, power of edge in one shot." This approach recognizes that technical buyers need confidence before curiosity.Pursue unexpected traction sources aggressively: Despite targeting enterprise ISVs, government demand emerged due to air-gapped environment requirements. "Government is actually becoming huge traction primarily because data ownership was a major topic to them." Rather than forcing initial market hypotheses, founders should redirect resources toward segments showing organic product-market fit signals, even when they require different sales processes.Build credibility through architectural pattern repetition: Investors backed OpenInfer because "we are the people that have built this twice, scaled it to millions." Repeating proven technical patterns across different contexts creates sustainable competitive advantages that new entrants cannot replicate without similar experience depth.
• How Hamming AI accidentally created a new category by focusing on customer problems instead of category creation | Sumanyu Sharma ($3.8M Raised) 12.09.2025 20mnt

  Hamming AI has emerged as a pioneer in voice agent quality assurance, creating what founder Sumanyu Sharma calls a "new category" of QA for conversational voice agents. After spending a decade building data products at scale at companies like Tesla and Citizen, Sharma recognized an acute pain point as voice agents began proliferating: enterprises desperately needed confidence that their voice agents would work reliably before launching to production. In this episode of Category Visionaries, Sharma shares how his team accidentally created a new category by following their instincts and leveraging a decade of expertise in reliability testing, audio processing, and machine learning.Topics Discussed:The evolution from Tesla's data science team to founding a voice agent QA companyHow "wandering the desert" for months led to finding the perfect problem-solution fitBuilding a completely inbound-driven go-to-market strategy in an emerging categoryThe decision to launch before feeling ready and building alongside customersWhy the voice agent market skeptics were wrong about market sizeCreating enterprise trust through reliability testing at scaleGTM Lessons For B2B Founders:Follow your instincts when you have deep domain expertise: Sharma spent months "wandering the desert" looking for the right problem until voice agent QA clicked. He emphasizes that when you have a decade of relevant expertise, you can recognize the perfect problem when it appears. As he put it, "when you see it, you kind of know... I am perfectly equipped to solve this specific problem. I'm built for this." Founders should trust their instincts when they have genuine domain expertise rather than overthinking market validation.Build something people want before focusing on category creation: Unlike many founders who start with category creation in mind, Hamming AI "accidentally" created their category by obsessively solving customer problems. Sharma notes, "We weren't looking to create a category. We were just looking to solve a problem that we feel passionate about, that we are already experts at." This customer-first approach led to organic category emergence and sustainable demand.Launch before you feel ready and build with customers: Sharma's biggest learning was launching with a "half-baked" product rather than perfecting it in isolation. "We didn't have a product that we thought was incredible. We just thought, hey, it kind of works, but let's actually build the product together with customers." This approach accelerated learning cycles and created stronger product-market fit than months of internal development would have achieved.Leverage contrarian insights from deep market proximity: While others dismissed voice agent QA as "too small," Sharma's data science background and proximity to builders gave him conviction. He analyzed the fundamentals: "Voice is a universal API for people. Voice agents are just becoming possible. They will be unreliable. Therefore, testing is very important. That's the math." Founders should develop conviction through first-principles thinking rather than consensus market opinions.Focus obsessively on customer success over marketing in emerging categories: Hamming AI remains completely inbound-driven, focusing entirely on making existing customers successful rather than traditional marketing. Sharma explains, "The voice space is so small where if you are doing a good job and if you build a product that people love, they will tell their friends about it." In nascent categories, product excellence and word-of-mouth can be more effective than broad marketing campaigns.
• How Nevermined coined "AI commerce" in 2023 to create category language before market adoption | Don Gossen 11.09.2025 17mnt

  Nevermined is pioneering the infrastructure for AI commerce, building payment rails specifically designed for agent-to-agent transactions. With a vision of trillions of AI agents functioning as both merchants and consumers, Don Gossen brings 20 years of AI experience to solving what he believes will be the foundational payment challenge of the next era of computing. In this episode of Category Visionaries, Don shares insights on creating an entirely new category—AI commerce—and the unique go-to-market challenges of building for a future that's rapidly becoming reality.Topics Discussed:The emergence of two distinct agent modalities: agent as proxy and agent as independent economic actorWhy existing payment infrastructure cannot handle the scale and velocity of AI agent transactionsNevermined's commission-based business model focused on agent-to-agent paymentsThe fundamental cost model differences between SaaS and AI agentsCreating the "AI commerce" category and the strategic importance of early categorizationGo-to-market strategy targeting verticalized AI agent builders with Series A+ fundingThe infrastructure investment phase versus deployment challenges in AI adoptionGTM Lessons For B2B Founders:Target customers who have proven business models, not just potential: Don's go-to-market strategy specifically targets AI agent companies that have raised Series A or later rounds. His reasoning: "Hopefully the VCs that are backing them have done some due diligence. And the money they're earning is actually real." Rather than chasing every potential customer, focus on those who have already validated their revenue model and can immediately benefit from your solution.Understand the fundamental cost structure of your customer's business model: Don identified that AI agents have an inverted cost model compared to traditional SaaS—most costs are operational (OpEx) rather than capital (CapEx). He explains: "The cost model is basically flipped. Most of your cost is actually on the opex... Your operating costs fluctuate based on the request." This insight shaped Nevermined's entire value proposition around cost monitoring and settlement rather than just payment processing.Create category language early, even before market adoption: Don coined "AI commerce" in 2023 when "people were like, what the hell's an AI agent?" His approach: "It always helps to categorize and provide language that's going to allow people to understand what it is that you're talking about... It's the memeification of the category." Don't wait for your market to mature—create the vocabulary that will define it.Focus on the operational reality, not the theoretical use case: While competitors focus on connecting bank accounts to AI agents for consumer purchases, Don focuses on the underlying workflow costs: "How much does the workflow cost to actually render that outcome?" Understanding the true operational mechanics of your customers' business—not just their surface-level needs—can create significant competitive differentiation.Leverage deep domain expertise to identify non-obvious problems: Don's 20 years in AI revealed that variable AI agent responses create variable operational costs—a problem most founders wouldn't recognize. He notes: "Until recently most people didn't realize that is a major issue in operating these solutions." Deep industry experience can help you spot problems that newer entrants miss entirely.
• Why Typedef starts go-to-market activities during the design partner phase instead of after | Kostas Pardalis ($5.5M Raised) 19.08.2025 27mnt

  Typedef is building an inference-first data engine designed for the new era of AI agents and machine-to-machine interactions. With $5.5 million in funding, the company is reimagining data infrastructure for a world where both humans and AI systems need seamless access to data processing capabilities. In this episode of Category Visionaries, I sat down with Kostas Pardalis, Co-Founder & CEO of Typedef, to explore how the company is addressing the fundamental shift from traditional business intelligence platforms to AI-native data infrastructure that treats inference as a first-class citizen alongside traditional compute resources.Topics Discussed:Typedef's vision for inference-first data infrastructure in the AI eraThe transition from human-only to machine-to-machine data interactionsWhy infrastructure companies take longer to reach revenue but build deeper moatsThe evolution from pre-AI data platforms to AI-native solutionsDesign partner strategies for infrastructure companiesGo-to-market approaches that combine bottom-up (engineers) and top-down (decision makers) strategiesCategory creation challenges in rapidly evolving AI marketsThe importance of open source and education in developer-focused go-to-marketGTM Lessons For B2B Founders:Start go-to-market activities during the design partner phase: Kostas emphasized that go-to-market isn't something you switch on after product development. "It's okay to go out there and talk about something that it's not very well defined or it might change, but actually it doesn't matter... go to market like just like everything else, it's an interactive process." B2B founders should begin building awareness, creating content, and engaging with potential customers even while their product is still evolving.Design partners must have real pain, not just time: The biggest insight about design partnerships is treating them like real customer relationships. "A design partner is still someone who has a problem that needs to be solved... no one is just donating their time out there... There still has to be value there." Don't approach design partnerships as charity work - ensure there's genuine mutual value exchange where your solution addresses real business pain.Product-market fit requires both product AND market innovation: Kostas challenged the common engineering mindset about product-market fit: "Many times, especially engineers, think that when we say product, market fit is that we have market, which is a static thing and we just need to iterate over the product until we find the right thing that matches exactly the market. No, that's not right." B2B founders must innovate on both the product and go-to-market sides simultaneously, including defining their target vertical and building appropriate sales motions.Infrastructure sales require dual-persona strategies: When selling to developers and technical infrastructure, you need both bottom-up and top-down approaches. "Even if you go to the manager and they love what you are saying, you still have to convince the engineers to use this thing... And they have a lot of leverage and vice versa." The bottom-up motion involves open source adoption and education, while the top-down involves traditional outbound sales to decision makers.Category creation doesn't guarantee category dominance: Having witnessed category creation firsthand, Kostas shared that defining a category doesn't ensure winning it. "It doesn't necessarily mean that because you define the categories that you are going to win at the end... Vercel was not actually the company that invented the category there." Focus on solving real problems and building sustainable competitive advantages rather than just being first to market with category messaging.
• How Personal AI scales enterprise contracts by selling to COOs and business users first | Suman Kanuganti ($16M Raised) 19.08.2025 25mnt

  Personal AI is pioneering the next generation of artificial intelligence with their memory-first platform that creates personalized AI models for individuals and organizations. Having raised over $16 million, the company has evolved from targeting consumers to focusing on enterprise customers who need highly private, precise, and personalized AI solutions. In this episode of Category Visionaries, we sat down with Suman Kanuganti, CEO and Co-Founder of Personal AI, to explore the company's journey from early AI experimentation in 2015 to building what he envisions as the future AI workforce for enterprise organizations.Topics Discussed:Personal AI's evolution from consumer-focused to enterprise B2B platformThe technical architecture behind personal language models vs. large language modelsPrivacy-first approach and competitive advantages in regulated industriesGo-to-market pivot and scaling from small law firms to enterprise contractsUnit economics advantages and 10x cost reduction compared to traditional LLMsVision for AI workforce integration in public companies within 3-5 yearsGTM Lessons For B2B Founders:Recognize when market timing doesn't align with your vision: Suman's team was building AI solutions as early as 2015, nearly a decade before the ChatGPT moment. When ChatGPT launched in November 2022, Personal AI faced confusion from investors and customers about their differentiation. Rather than forcing their sophisticated personal AI models on consumers who wanted simpler solutions, they recognized the market mismatch and pivoted. B2B founders should be prepared to adjust their go-to-market approach when market readiness doesn't match their technical capabilities, even if their technology is superior.Find your wedge in enterprise through specific pain points: Personal AI discovered their enterprise entry point by targeting "highly sensitive use cases that LLMs are not good for" where companies would be "shit scared to put any data in the LLM." They focused on precision and privacy pain points that large language models couldn't address. B2B founders should identify specific enterprise pain points where their solution provides clear advantages over existing alternatives, rather than trying to be everything to everyone.Let customer expansion drive revenue growth: Personal AI's enterprise strategy evolved organically as existing contracts "started growing like wildfire as more people had a creative mindset to solve the problem with the platform." They discovered that their Persona concept allowed enterprises to consolidate multiple AI use cases into one platform. B2B founders should design their platforms to naturally expand within organizations and reduce vendor fragmentation, creating stickiness and increasing average contract values.Leverage architectural advantages for unit economics: By positioning their personal language models between customer use cases and large language models, Personal AI achieved "10x lower cost" per token. This architectural decision created both privacy benefits and economic advantages. B2B founders should consider how their technical architecture can create sustainable competitive advantages in both functionality and economics, not just features.Geography matters more than you think for fundraising: Suman identified his biggest fundraising mistake as not moving to San Francisco earlier, stating "back in 2022 or 2023 is when I should have moved to San Francisco, period." He learned that being part of the Silicon Valley ecosystem and conversation is critical for fundraising success. B2B founders should consider the strategic importance of physical presence in key markets, especially when raising capital, and not underestimate the value of in-person relationship building.
• How Wispr Flow manufactured viral moments by personally onboarding 500 users on Google Meet | Tanay Kothari ($56M Raised) 15.08.2025 27mnt

  Wispr Flow has transformed voice dictation from a frustrating novelty into a seamless productivity tool that users trust implicitly. With a recent $30 million Series A led by Menlo Ventures, the company has achieved remarkable product-market fit through 90% word-of-mouth growth and users who share the product organically without prompting. In this episode, I sat down with Tanay Kothari, CEO and Co-Founder of Wispr Flow, to learn about the company's pivot from hardware to software, their approach to manufacturing viral moments, and their strategy for competing against tech giants with distribution advantages.Topics Discussed:Wispr Flow's pivot from building voice assistant hardware to focusing on voice-to-text softwareThe company's unique approach to achieving sub-half-second latency and exceptional accuracyBuilding viral growth through manufactured "aha moments" and exceptional user onboardingCompeting against OpenAI and Apple through speed of execution and user experience focusThe challenge of building for mainstream users beyond Silicon Valley's tech-savvy populationStrategic decisions around cutting non-essential growth channels to maintain focusGTM Lessons For B2B Founders:Manufacture viral moments through obsessive user research: Tanay personally onboarded the first 500 users via Google Meet, watching their facial expressions, mouse movements, and emotional reactions in real-time. This intensive observation allowed him to identify and systematically reproduce moments of user delight. He explained, "Find the things that repeatedly create delight, make sure that never dies, and then find the other places where there's confusion and kind of take them out." B2B founders should invest heavily in understanding the micro-moments of user experience, as these compound into organic growth at scale.Leverage authentic product usage by your target buyers during fundraising: When Wispr Flow raised their Series A, every VC in Silicon Valley was already using the product daily. Tanay noted, "I didn't need to convince them about why the product was good. All I had to tell them about if you believe why Whisper is good today, here is where we can take the company." This eliminated the typical product demonstration phase and shifted conversations to vision and execution capability. B2B founders should prioritize getting their product into the hands of potential investors as users before ever pitching them as investors.Build anti-fragile technology that improves as the industry evolves: Rather than competing directly with AI model capabilities, Wispr Flow built infrastructure that gets better as underlying AI models improve. Tanay instructs his team: "If at some point that you feel afraid of a new model launching, you're doing something wrong." This philosophy led them to focus on latency, user experience, and integration rather than competing on raw AI performance. B2B founders in AI-adjacent spaces should identify where they can create value that compounds with industry improvements rather than being displaced by them.Cut aggressively to maintain focus during rapid growth: Despite conventional wisdom, Wispr Flow eliminated SEO efforts entirely because "no one is searching for voice dictation" and most people don't know the technology has reached usability thresholds. Tanay applies an extreme 80/20 rule: "You can cut the 80% of the things that are not giving you the results... You find a new 20% that's going to give you 80% more results and you can just keep doing that again and again." Design for mainstream adoption beyond early adopters: While most AI tools target Silicon Valley technologists, Tanay identified that 95% of the population represents the real market opportunity. He noted these users "end up being your most loyal users" because they have less churn and higher lifetime value than tech-savvy early adopters.
• Sauraj Gambhir, Co-Founder of Prior Labs: $9 Million Raised to Build Foundation Models for Structured Data 31.07.2025 16mnt

  Prior Labs is pioneering foundation models for structured data, bringing transformer technology from the generative AI world to tabular data that sits in databases and spreadsheets across every business. With $9 million in funding and over 1.5 million downloads of their open-source model, Prior Labs is revolutionizing how data scientists work with structured data by creating universal models that can handle multiple use cases instead of requiring custom models for each specific application. In this episode, I sat down with Sauraj Gambhir, Co-Founder of Prior Labs, to explore how they're transforming machine learning workflows from taking days to seconds and building a global community around their breakthrough technology.Topics Discussed:Prior Labs' mission to bring transformer technology to structured data and tabular datasetsThe transition from traditional 20-year-old machine learning methods to universal foundation modelsBuilding a horizontal product that serves data scientists across finance, healthcare, and scientific researchThe company's open-source strategy with 1.5 million downloads and community-driven developmentSocial media and community-building tactics that drove adoption across LinkedIn, Twitter, and DiscordScaling from 3 to 16 team members in seven months while maintaining technical focusFundraising strategy for AI companies and the balance between raising enough capital without over-inflatingPlans for geographic expansion from Berlin to the US marketGTM Lessons For B2B Founders:Lead with open source for technical audiences: Prior Labs built their entire go-to-market strategy around an open-source model that anyone can download and use for free. Sauraj explained, "We've got like over one and a half million downloads and it is open source. You just need to attribute us that you're using our model." This approach allowed them to achieve massive adoption while building credibility with their technical audience. B2B founders targeting developers or technical users should consider how open source can accelerate adoption and community building before monetization.Build community ownership, not just engagement: Sauraj approaches community building like team building, saying, "If you think about your team as a founder, like when you build a team, you want them to feel like it's their company... I'm trying to take that same philosophy towards community building." He creates biweekly Discord updates where half the content showcases community contributions, leading members to actively submit their use cases and request features. Leverage co-founder networks strategically for different audiences: Prior Labs uses each co-founder's unique network to reach different segments. Sauraj noted, "One of my co founders has been a professor of machine learning for the last 12 years. So he already had a pretty good following of let's say the data science community... when we need to generate inbound, I'm the one pushing when we need to like generate more technical applications for people to apply for jobs with us. We're going through my co founders networks." Focus adoption over monetization in emerging categories: Despite having paying customers, Prior Labs keeps their API free and focuses entirely on adoption metrics. Sauraj explained, "Right now we are offering it for free because we just want like adoption is really the biggest use case at the moment... when we have like the next versions of the models, that's really when we're going to be able to flip the switch." Use technical documentation as brand building: Instead of focusing on traditional marketing materials, Prior Labs invested heavily in developer-focused assets. Sauraj said, "We were really focused on getting really good docs in place, really good, like GitHub read me in place. And the brand was really kind of like building this community and being like open and honest with the community."
• Tony Zhang, Founder & CEO of Tera AI: $8M Raised to Build the Future of Robotics Operating Systems 31.07.2025 23mnt

  Tera AI is pioneering a software-centric approach to robotics, moving away from traditional hardware-dominated solutions toward a unified operating system for robotic platforms. After raising $8 million and transitioning from insurance applications to robotics, the company is building what founder Tony Zhang envisions as "a general purpose operating system for robot platforms" powered by spatial foundation models. In this episode of Category Visionaries, Tony shares his journey from Google X to founding Tera AI, including hard-won lessons about market validation, customer discovery, and the critical importance of understanding buyer priorities.Topics Discussed:Tera AI's evolution from geospatial foundation models in insurance to robotics applicationsThe challenges of customer discovery in regulated industries like insuranceTony's experience at Google X and the ChatGPT moment that sparked entrepreneurial actionFirst Round's Product Market Fit program and structured customer discovery methodologyThe transition from hardware-centric to software-centric robotics architectureFundraising strategies and developing instincts for investor feedbackBuilding a team of top-tier AI researchers in a competitive talent marketGTM Lessons For B2B Founders:Lead with priority validation, not pain discovery: Tony learned the hard way that not every pain point can be solved on a VC timeline. His breakthrough insight was asking upfront: "Tell me if this is one of your top three priorities. If not, tell me what are those three priorities." He discovered that many insurance prospects liked their solution but had more pressing infrastructure problems unrelated to AI. B2B founders should qualify buyer priorities before presenting solutions to avoid getting trapped in lengthy sales cycles for non-critical problems.Understand regulatory constraints early in enterprise markets: Tera AI spent nearly a year in insurance before realizing that regulatory barriers made technology adoption extremely difficult, regardless of product-market fit. Tony explains: "Because of the regulations in America, it is incredibly difficult for an insurer or carrier to adopt new technology, especially technology that was as new as the stuff that we were building." Founders entering regulated industries should map compliance requirements and adoption timelines before committing significant resources.Structure customer discovery to eliminate waste: Through First Round's PMF program, Tony discovered they were doing discovery calls inefficiently, often requiring multiple meetings with the same prospects. The key insight was asking the right qualifying questions upfront rather than leading with solutions. This approach eliminated unnecessary follow-up meetings and accelerated their discovery process by 5x. Founders should develop structured discovery frameworks with clear qualifying criteria before scaling outreach efforts.Market timing requires both technology readiness and buyer urgency: Tony's "ChatGPT moment" wasn't just about technological possibility—it was about recognizing the convergence of technical capability and market readiness. He emphasizes: "It wasn't too early, it wasn't too late." The key was understanding that spatial AI could finally deliver value that buyers were ready to adopt. Founders should evaluate both technical feasibility and market timing when deciding on startup opportunities.Attract talent with novel technical challenges, not just compensation: Despite intense competition for AI talent in Silicon Valley, Tera AI successfully recruits top researchers by offering genuinely innovative work. Tony explains: "We genuinely try to innovate across the entire stack. We build our own models, we build our own datasets, we can write papers on the things we're doing." They target researchers who are "bored to death by the LLM world" and want to work on groundbreaking spatial AI problems.
• David Reger, CEO of NEURA Robotics: €185M Raised to Power the Future of Cognitive Robotics 16.07.2025 34mnt

  NEURA Robotics is transforming the robotics industry by building cognitive robots powered by physical AI. With €120 million raised and 5,000-10,000 robots already deployed, the company has set an ambitious goal of deploying 5 million robots by 2030. In this episode, I sat down with David Reger, CEO and Founder of NEURA Robotics, to explore how his company is solving the reliability and adoption challenges that have kept robotics a niche market, and his vision for making robots as ubiquitous as smartphones. Topics Discussed:NEURA's partnership-driven go-to-market strategy using horizontal and vertical partnersThe company's unique physical AI model built specifically for embodied intelligenceCurrent deployment of household robots starting with elderly care applicationsThe challenge of raising hardware funding in Europe versus Japan and ChinaBuilding cognitive robots that can operate with limited compute and bandwidthCreating a platform ecosystem where partners can download skills and applicationsThe regulatory and cultural barriers to robot adoption in different marketsNEURA's recent partnership with SAP and strategy to become Europe's next €100 billion companyGTM Lessons For B2B Founders:Leverage established channels for reliability-critical products: David built NEURA's entire go-to-market strategy around partnering with established robot companies rather than direct sales. He recognized that for reliability-critical hardware like robots, startups face an inherent trust deficit. "If you're talking about robots, there's all about reliability, it's all about trust because it has to run 24/7... And if you're looking into strength of a startup, that's exactly the point. Like this is something you don't have." B2B founders in hardware or mission-critical software should consider white-label partnerships with established players who already have the service infrastructure and customer trust.Build horizontal and vertical partnership ecosystems simultaneously: NEURA created a dual partnership model - horizontal partners (robot manufacturers) for broad distribution and vertical partners (domain specialists like welding or household task companies) for specialized applications. This creates a platform effect where "our partners don't have to have the knowledge, but they can simply download, let's say an app or a skill and they can use the robot like in all kinds of different domains." B2B founders should consider how to enable both broad distribution and deep specialization through complementary partnership types.Target markets where regulatory shifts create urgency: David identified that China's 2030 goal of transforming 5% of working labor to robotics (40 million robots) would force global competition. "The whole world has to, let's say, also wake up in the same time... because if we don't want to end up, let's say as a museum, we have to also contribute." B2B founders should identify geopolitical or regulatory shifts that create market urgency and position their solutions as necessary responses to competitive pressure.Raise capital in markets that understand your technology: When European and US investors were skeptical of hardware, David found receptive investors in Japan who "believe in robots" and understood the market potential. He eventually had to pivot to China for speed, then later successfully raised €120 million in Europe when the market shifted. B2B founders should be willing to pursue capital in non-obvious geographies where their technology vision is better understood, even if it requires navigating different business cultures.Focus on physical AI differentiation for embodied products: David emphasized that NEURA's competitive advantage lies in their physical AI model: "I do believe that like our AI model is one of the, let's say it's the best in the world in that space, because simply it's much more efficient and actually built for being physical, while the most other models are not."