Building an Autonomous Enterprise for Real-World Services with Netic Founder Melisa Tokmak

No Priors (hosts Sarah Guo and Elad Gil, guest Melisa Tokmak) · podcast · 2026-08-01 Reader 🖥️

practical · mixed — A VC-podcast founder interview: argumentative (defends a specific strategic choice — product platform vs. AI roll-up — and dismisses two competitive threats) interleaved with expository description of the business and narrative anecdotes (the -20°F outage, her origin story, a new hire's interview answer) that carry the argument implicitly.

Structure

Key terms

• Netic-first / N1
  A customer whose very first interaction with the essential-service business is with a Netic AI agent rather than a human — her metric for AI-native adoption depth.
• Autonomous enterprise
  An essential-service company where every function except the actual hands-on service labor itself — intake, diagnosis, routing, revenue generation, customer engagement — runs autonomously on AI.
• Essential services
  A deliberately broad category: billion-dollar home-services companies (HVAC, plumbing, electric), consumer wellness/membership businesses, hospitality, automotive, and pet services — large, high-volume, real-world businesses distinguished from tech-serving-tech startups.
• AI roll-up
  Elad's term, used as the explicit foil throughout: a firm (e.g. Long Lake) buying operating businesses in a vertical and optimizing them with AI, as opposed to Netic's single scalable product sold across many independently owned operators.
• The three layers
  Her framework for where competitive defensibility lives in an AI-for-real-world-services product: the underlying model, the harness/orchestration built on top of it, and the product/software layer — all three, not any one alone, and specifically not the model.
• Permanent underclass mentality / 'AGI pill' view
  A critical label for a mindset (which she attributes especially to Gen Z candidates) that treats the next 18–36 months as existentially decisive — believing that if you don't make money or learn everything now, AI will permanently erase your economic value.

Propositions

• explicitNetic operates as an AI intermediary between essential-service enterprises and their end customers — understanding customer need, matching it against the business's operational rules, and deploying the service or labor — across the full interaction from first contact to dispatch.conf: high
• explicitOver 70% of Netic's customers are now 'AI-first' (Netic-first / N1) — their very first interaction with the business is with a Netic AI agent rather than a human.conf: high
  extendsHow Gemini is speaking the language of Southeast Asia · p6 — Same self-reported, methodology-free adoption-metric pattern, extended to a new genre: a VC-podcast founder interview (with a disclosed investor as host) rather than a corporate blog post.
• explicitEssential-service businesses historically could not grow without continually adding headcount, which caps growth for the many that are private-equity-owned and margin-focused (EBITDA businesses).conf: high
• explicitBuilding one scalable, compounding product platform usable by any operator in the vertical creates more durable value than an 'AI roll-up' strategy of buying and optimizing individual operating businesses, because a roll-up's product doesn't generalize beyond the companies it already owns.conf: high
• explicitRobotics is far from being able to replace human labor in these industries in the current 'chapter,' due to dexterity limits, the huge physical variation across buildings and homes, and the emotional stakes present when a customer is having 'the worst day of their lives.'conf: high
• explicitFrontier AI labs (OpenAI, Anthropic, Google, Meta) are not a competitive threat to Netic, because enterprises in these industries want durability rather than the labs' fast-build/fast-kill product cadence, because the enterprise vertical is fragmented (~20 competing point solutions) rather than lab-dominated, and because labs are incentivized toward the most generalizable solution ('wait for AGI') rather than the industry-specific 'last mile' work.conf: high
  extendsThe Founder's Playbook: Building An AI-Native Startup · p7 — Both locate AI-native defensibility outside the model itself (accumulated depth / harness+orchestration+product); Melisa extends the claim specifically to argue frontier labs themselves aren't a threat, not just other startups.
• explicitDefensibility for an AI-for-real-world-services product requires being good at all three layers together — the model, the harness/orchestration on top of it, and the product/software layer — not the model alone.conf: high
  supportsBuilding AI Agents For The Enterprise · p10 — Near-identical framing: defensibility lives in everything around the model (data/trust/domain expertise there; harness/orchestration/product here), with the model as enabler rather than moat.
• explicitThe private-equity playbook has shifted from 'find an undervalued gem, flip it for a multiple' toward creating genuine, tangible ROI in the businesses they hold — yet PE conversations still 'always' start focused on cost-cutting, and she has to actively redirect them toward net-new revenue.conf: high
  supportsThe Tidy House · p13 — Her own admission that PE conversations 'always' start focused on cost-cutting corroborates the more skeptical claim that PE captures AI efficiencies as profit — in tension with her own more optimistic framing that the playbook has already shifted.
• explicitDemonstrated, sustained agency and follow-through across a person's whole life — not credentials, a single project, or a rehearsed answer — is the primary signal Netic screens for in hiring, probed through open-ended questions about the hardest thing a candidate has ever done.conf: high
  supportsWhy cultivating agency matters more than cultivating skills in the AI era | Max Schoening (Head of Product, Notion) · p1 — Netic's hiring practice operationalizes exactly this claim, screening for demonstrated agency and follow-through over credentials or skills.
• explicitA mindset that treats the next 18 months as existentially decisive for one's economic value (the 'permanent underclass' / 'AGI pill' view) is dangerous, because building genuinely good things and durable skill requires years of committed follow-through, not urgency-driven shortcuts.conf: high
  supportsLeadership Summit 2026: Inspiring Grit and Growth Amid Unprecedented Change · p1 — Her critique of short-termist '18-month' thinking, and her own multi-year commitment narrative, match the grit definition's emphasis on sustained consistency over intensity or urgency.
• explicitFounders today are unusually cautious about entering verticals they perceive as being on a frontier lab's roadmap, more so than in prior tech cycles, partly because many are now building companies to exit quickly rather than commit to them for decades.conf: medium
• explicitNetic-enabled interactions have generated over $600 million for its customers over the company's roughly two-year existence.conf: high
• explicitEvery engineer at Netic is required to visit customer sites in person before building product back in San Francisco, treating direct field exposure as necessary to design the right product for these industries.conf: high

Arguments

Narratives

• A customer's furnace breaks in -20°F weather; they need to find a trustworthy HVAC provider fast, and Netic agents handle the intake, diagnosis, and urgency/lifetime-value-aware routing.
  → Makes the operational complexity behind propositions 1 and 7 concrete — shows why 'is it that simple' understates what the agent actually has to reason about.
• Melisa's own origin story: growing up in a small town in Turkey without a computer, arriving on a full scholarship to Stanford, and getting informal mentoring via Facebook Messenger from earlier alumni of her town who had already emigrated to study in the US.
  → Grounds her closing claim that AI/education removes 'resourcing as a question' as a barrier to opportunity, and supports her stated motivation to build for 'real world' industries rather than serve other tech companies.
• A new hire's answer to Melisa's standard interview question ('the hardest thing you've ever done') — maintaining a strict personal health regimen every day for over 15 years without losing focus.
  → Operationalizes and supports proposition 9 (sustained agency/follow-through, not credentials or a single project, as the hiring signal).
• A large roofing company that previously sent door-knockers neighborhood-to-neighborhood now has Netic connect satellite hurricane/roof-material data into its agents, so the same platform both handles inbound/outbound conversations and proactively identifies which neighborhoods to target.
  → Supports proposition 7 (the 'three layers' / 'last mile' argument) with a concrete instance of domain-specific context a general-purpose model alone would not supply.

Questions addressed

• ✅ How should an AI company serve fragmented, operationally complex real-world service industries — as a product platform, or by buying and optimizing existing operators (a roll-up)?
  Build a single compounding, multi-tenant product platform: it fits her skill set, generalizes across customers, and isn't capped by deal flow the way a roll-up is.
• ✅ Will physical robotics displace this human-plus-AI labor model?
  Not in the current 'chapter' — dexterity limits, building/home variation, and the emotional stakes of the interactions push robotics into a future, separate chapter.
• ✅ Are frontier AI labs a competitive threat to vertical AI-for-real-world-services companies?
  No, in her view — labs are focused on generalizable model/product progress and fast iteration, not the durable, operationally embedded 'last mile' work these industries need.
• ⭕ Is private equity's posture toward AI still purely cost-cutting, or has it genuinely shifted toward value creation?
  She asserts the industry playbook and rhetoric have shifted toward wanting real ROI and net-new revenue, but also states PE conversations 'always' start focused on cost-cutting and that she has to actively redirect them — leaving the actual extent of the shift unresolved.
• ✅ What screens reliably for high-performing hires at an AI-native company?
  Demonstrated, sustained agency/follow-through across a candidate's whole life, probed through open-ended 'hardest thing you've done' interviewing rather than a fixed rubric or credential check.

Gaps

• incompleteThe '70%+ of customers are AI-first (N1)' figure and the '$600 million generated for customers' figure are both self-reported by the founder in a podcast whose host explicitly discloses being an investor in her company; neither figure comes with a disclosed methodology, denominator, time window, or independent verification.
  Read next: Compare against the corpus's recurring vendor-self-reported-metrics pattern (e.g. extract-gemini-southeast-asia, extract-google-gemini-coffee-shop-case-study) for what independent corroboration of adoption/impact claims typically looks like versus what's missing here.
• incompleteThe claim that frontier labs aren't competitive threats is asserted from the speaker's own incentive position (a founder pitching to a host who is already an investor in her company) and isn't tested against the concrete counter-scenario of a lab, or a well-capitalized generalist agent platform, choosing to build vertical-specific orchestration for essential services — exactly the 'last mile' work she claims only Netic is positioned to do.
  Read next: extract-founders-playbook-building-ai-native-startup's proposition that 'well-funded incumbents' (not other AI-native startups) are the real competitive threat for AI-native companies is a more skeptical framing to weigh against her claim.
• incompleteNo discussion of what happens to the humans who previously handled the '70%' of first-contact interactions AI now handles — whether headcount shrank, roles were redefined, or hiring simply slowed. The episode frames AI-first adoption entirely from the operator's efficiency/growth side, never from the displaced- or redefined-labor side.
  Read next: Read against the corpus's ai-labor-disruption material for how other sources handle (or fail to handle) the labor-impact side of 'AI-first' transitions in physical/blue-collar-adjacent service work.
• illogicalShe claims the PE playbook has 'already changed' from flip-for-multiple to genuine value creation, but in the same answer says PE conversations are 'always' first about cost-cutting and that she has to actively 'expand their horizon' toward revenue — these two claims sit in tension and she doesn't reconcile them.
  Read next: extract-tim-oreilly-tidy-house's proposition that private equity tends to capture AI-generated efficiencies as profit rather than reinvest in service quality is a more skeptical read of the same dynamic, worth weighing against her more optimistic framing of an already-underway shift.

Related documents

• The Tidy House
• Leadership Summit 2026: Inspiring Grit and Growth Amid Unprecedented Change
• Why cultivating agency matters more than cultivating skills in the AI era | Max Schoening (Head of Product, Notion)
• How Gemini is speaking the language of Southeast Asia
• Andreessen Horowitz's Post
  AI-Native Company Moats & Defensibility
• Agents Over Bubbles
  Agency Moats & Defensibility
• AI Is Changing the Rules of Entrepreneurship
  AI-Native Company Moats & Defensibility
• Designing the AI-native engineering organization
  Enterprise AI Adoption & Readiness Agency
• Travel Through the Lens of AI with Booking.com CEO Glenn Fogel
  Agency Moats & Defensibility
• Brain Food: Answers Without Effort
  Grit (Passion & Perseverance) Agency