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

2026-07-31 - 35 min - source - Read full transcript

Sarah Guo (host)Elad Gil (host)Melisa Tokmak

Key insights

Netic positions itself as the AI interaction and orchestration layer between essential-service businesses and their end customers, not just a chatbot vendor.

For a company like an HVAC provider, Netic's agents handle the call or text, gather context (home type, service history, urgency), match it to the right technician and timing, and optimize for both customer delight and revenue - work Tokmak says is operationally complex because service triage, lifetime value, and technician specialization all factor into the decision.

vertical-ai-agents

Over 70% of Netic customers' end-user interactions are now AI-first, a state Netic calls N1.

Tokmak says most customers started by using Netic agents only to handle call overflow, but adoption has grown to the point where AI is the primary first touchpoint for the large majority of interactions across Netic's enterprise customer base.

vertical-ai-agents

Tokmak deliberately built a software platform rather than an AI roll-up, because her skill set is product and engineering, not M&A, and because roll-up products don't generalize.

She contrasts her approach with firms like Long Lake that buy and operate service businesses directly with AI. Her three stated reasons for not doing a roll-up: it isn't the personal mission she wants, M&A isn't her core skill, and roll-up-built products only ever serve the specific companies acquired rather than compounding across an entire industry.

build-vs-roll-up

Full robotic automation of home-service trades is far off because of environmental variation and required dexterity, so the near-term AI opportunity is the software/orchestration layer, not physical labor.

Tokmak argues that for robots to replace technicians, buildings would essentially need to be standardized or 3D-printed, and current robotics can't reliably handle the fine-grained dexterity of variable tasks (different screws, wall types, cramped access). She also notes a human element - customers are often having one of the worst days of their lives - that current robotics isn't equipped to navigate.

vertical-ai-agents

Tokmak doesn't see foundation model labs as a competitive threat to Netic because winning in these industries requires a full stack - models, orchestration/harness, and deep vertical product work - that generalist labs aren't focused on building.

She argues labs optimize for the most generalizable version of a problem ("we'll ask AGI how to solve it") rather than the industry-specific last-mile work of handling accents, local context, and repeat-customer relationships. She also notes enterprises in her space don't want a vendor with OpenAI's pattern of shipping fast and killing products fast, or the multi-product sprawl she associates with some enterprise AI offerings.

ai-lab-competition

Tokmak screens for agency by asking about the hardest thing a candidate has ever done and probing whether they sustained commitment through difficulty over time, not just cited a single anecdote.

She looks for a track record of continuous agency - projects candidates stuck with when it got hard - rather than an isolated example. She cites a recent hire whose answer was maintaining a strict personal health regimen and work discipline for over 15 years, which she found more compelling than a conventional "built a startup" story.

founder-and-hiring-philosophy

Tokmak criticizes a growing founder and hiring-pool mindset she frames as an "AGI pill" belief that one must extract value or learn everything within 18 months before AI makes people obsolete.

She says this belief, common in some younger candidates she interviews, discourages the patient, decades-long commitment she thinks is required to build something real, and contrasts it with her view that founders today are too often optimizing for a fast exit rather than dedicating themselves to a mission.

founder-and-hiring-philosophy

Private equity's AI playbook has shifted from finding undervalued "gems" to arbitrage, toward generating tangible new revenue in existing portfolio companies, though conversations still often open with cost-cutting.

Tokmak says cheap, undiscovered acquisition targets no longer exist, so PE firms increasingly hire AI-focused operating partners and look for platforms that create measurable new revenue rather than pure margin extraction. She contrasts this with treating an AI product like typical deterministic software that should show ROI within a week, arguing AI adoption is the start of a relationship whose value should compound over the year rather than a one-time trial.

private-equity-ai-adoption

Essential-service industries (HVAC, roofing, pet care) are commonly mischaracterized as slow, old-school adopters, when in practice many owners are highly tech-forward, value-driven buyers.

Tokmak cites large enterprise customers that ran thorough evaluations (one half-million-dollar contract took 14 days end-to-end) and describes roofing companies that already combine door-to-door sales ('door knockers') with satellite data on storm damage and material needs, feeding that context into Netic's agents to prioritize outreach.

vertical-ai-agents

Netic reports generating over $600 million in value for customers from AI-handled interactions, which Tokmak treats as concrete proof of ROI versus AI "vaporware" demos.

She contrasts this with sales conversations built on demos alone, saying Netic instead shows live deployments and real dollar outcomes so customers - and the private equity firms guiding them - can evaluate tangible impact rather than a one-time pilot.

private-equity-ai-adoption

Companies

Netic - Tokmak's company: an AI layer that sits between large essential-service businesses (HVAC, plumbing, roofing, pet care, wellness, hospitality, automotive) and their customers, handling inbound/outbound interaction and dispatch end to end.
Scale AI - Tokmak spent about four years there before founding Netic, building the government business unit and large-enterprise verticals (logistics, manufacturing, financial services, healthcare).
Meta - Paid roughly $28-30 billion for a licensing/talent arrangement with Scale AI, per Tokmak; also her employer prior to Scale AI.
Long Lake - Cited as an example of an AI roll-up firm that buys home-service businesses and optimizes them with AI, contrasted with Netic's platform-for-everyone approach.
Harvey - Cited by Elad Gil as one of the vertical AI applications that emerged roughly four years ago.
Perplexity - Cited alongside Harvey as an early wave of AI-native vertical/consumer applications.
Decagon - Cited as part of a more recent wave of vertical AI applications, alongside Netic.
Abridge - Cited (name garbled in the ASR transcript as "a bridge") alongside Decagon as part of the recent wave of vertical AI applications.
OpenAI - Discussed as a leading lab that ships fast but also discontinues products fast, which Tokmak says doesn't fit what enterprises in her industries want from a vendor.
Anthropic - Discussed as a leading lab known for focus, credited with pulling ahead in coding agents via Claude.
Google - Referenced as the earlier version of today's "can the labs do this" competitive question - a decade ago the question was "can Google do this?"
SpaceX - Cited as a company Tokmak respects for long-term craftsmanship and durability of vision.
Notion - Cited alongside SpaceX as a company known for craftsmanship and serving its audience over many years.

Techniques and frameworks

N1 (Netic-first) metric - Netic's internal measure of the share of a customer's end-user interactions where a Netic AI agent is the first point of contact; over 70% of Netic's customers are now N1.
Agency interview question - Tokmak's hiring technique: ask candidates about the hardest thing they've ever done in life, then dig into whether they sustained effort through difficulty, rather than judging a single anecdote.
Christian shoemaker analogy - A quote Tokmak attributes to Martin Luther - a shoemaker honors God not by putting crosses on shoes but by building the best shoe - used as her framing for craftsmanship-first building.

Summary

Melisa Tokmak, founder and CEO of Netic, describes her company as the AI layer that sits between large essential-service businesses - HVAC, plumbing, roofing, pet care, wellness, hospitality, automotive - and their end customers. Netic's agents handle inbound calls, texts, and online scheduling, gather context about the customer and the job, and route work to the right technician at the right time, all while optimizing for customer satisfaction and revenue. Adoption has grown from handling call overflow to becoming the primary point of contact: over 70% of Netic's enterprise customers' end-user interactions are now AI-first, a state the company calls "N1."

Tokmak explains why she built a horizontal platform rather than pursuing an AI roll-up strategy - buying and directly operating service businesses, as firms like Long Lake do. She cites three reasons: mission (she wants to build a product, not run M&A), skill set (she is an engineer and product builder, not a dealmaker), and scalability (roll-up products only ever serve the specific companies acquired, while a platform can compound across an entire industry). Her path to this conviction ran through four years at Scale AI, where she built the government and large-enterprise business units, and a personal history growing up in a small town in Turkey and arriving at Stanford on a full scholarship without ever having owned a computer - an experience she says shaped her interest in AI that creates tangible impact outside the tech industry's usual customer base.

On the question of whether full automation - robotics, self-driving - eventually converges with what Netic does, Tokmak is skeptical on any near-term timeline. She argues that the physical variation in buildings and the dexterity required for trade work (different screws, wall types, cramped spaces) puts robotic replacement of technicians far in the future, and that the emotional stakes of a customer's "worst day" call for human labor that AI orchestrates rather than replaces. On competition from foundation model labs, she isn't worried: winning in these verticals requires a full stack of model, orchestration/harness, and deep product work that she believes labs aren't focused on building, and she notes that enterprises in her space specifically don't want a vendor with the shipping-and-sunsetting cadence she associates with a company like OpenAI.

A recurring theme is Tokmak's hiring and founder philosophy. She screens for "agency" by asking candidates about the hardest thing they've ever done and digging into whether they sustained effort through difficulty over time, rather than accepting a single anecdote. She's critical of what she calls an "AGI pill" mindset she sees in some younger candidates - the belief that they must extract all possible value or learn everything within 18 months before AI renders them obsolete - arguing this short-termism undermines the patient, decades-long commitment required to build something real. She frames her own philosophy through a quote she attributes to Martin Luther about a shoemaker honoring God through craftsmanship rather than decoration, and names Notion and SpaceX as companies she respects for sustaining that kind of long-term craft.

On the buyer side, Tokmak pushes back on the idea that essential-service industries are slow, old-school technology adopters; she describes highly tech-forward, value-driven enterprise buyers and a roofing company that already combines door-to-door sales with satellite storm-damage data. She also describes private equity's AI playbook shifting from cheap-arbitrage acquisitions toward generating tangible new revenue in portfolio companies, though initial conversations still tend to start with cost-cutting. Netic's proof point, she says, is roughly $600 million generated for customers from AI-handled interactions - real deployments and dollar outcomes shown directly, rather than demos. She closes by naming what excites her most beyond Netic: AI's potential to expand access to education and health information for people who, like her own younger self, don't have access to expensive resources or expert help.

Notable Quotes

"The Christian shoemaker doesn't honor God by putting little crosses on the shoes. He does so by building the best shoe... because God cares about craftsmanship." - Melisa Tokmak

"I grew up with nothing and really came here only for college. When I got a full scholarship to Stanford, I didn't even own a computer before." - Melisa Tokmak

"I think 10 years ago, that same exact question was, can Google do this? And then now it became can labs do this?" - Melisa Tokmak

"We have made so far, I think, over $600 million for our customers that have been really generated from AI-handled interactions." - Melisa Tokmak