AI in Real-World Services: The Hard Part Is Operations, Not Conversation
The service problem is connecting judgment to dispatch
Netic founder Melisa Tokmak does not describe the product as simply an AI that answers calls. She says the system sits between a business and its customer, understands a need, applies operational rules, and then determines who should deliver service, when, and how. In HVAC, plumbing, and similar businesses, urgency, equipment, available technicians, and customer value can change a dispatch decision. Melisa Tokmak 00:07 “every single thing to understand the customer need or want” Direct Audio Anchor Listen from 00:07
Tokmak says more than 70 percent of Netic customers are AI-first and that the customer's first interaction is handled by a Netic agent. That is the guest's company metric, not independently verified in the episode. It nevertheless defines the product direction: not only handling overflow calls, but operating at the first customer entry point. Melisa Tokmak 06:26 “over 70% actually all of their customers first interaction with the company is with NetIC agents.” Direct Audio Anchor Listen from 06:26
A platform boundary is not a model boundary
Tokmak did not choose an acquisition-and-optimization roll-up. She says products in that model usually serve the acquired company, while she wants one product layer that many real-world businesses can use, leaving each business to focus on labor and service quality. Melisa Tokmak 09:28 “how every real world business can run on netic” Direct Audio Anchor Listen from 09:28
She likewise does not see robotics or frontier labs as direct near-term replacements. Variation in buildings, equipment, and human situations makes physical service difficult to standardize, while models still need harnesses, orchestration, software, and product work to complete the last mile. This is her judgment about the industry's timeline and competitive boundary, not a verified forecast. Melisa Tokmak 12:56 “There's quite a bit of lost mile that you really have to do” Direct Audio Anchor Listen from 12:56
Long-lived products need long-lived commitment
Tokmak uses a quote attributed to Martin Luther about a shoemaker to discuss craft: the point is not the symbol, but making the best shoe. She applies that principle to product work, which she describes as solving a problem for people one genuinely cares about. Melisa Tokmak 18:27 “It really is about building the best product that are solving the problems of the people you care about.” Direct Audio Anchor Listen from 18:27
Her hiring standard extends the same idea. She looks not only for one conspicuous act of initiative, but for a person who has started, sustained, and finished something over time. Her five-year vision is for Netic to manage the operational layer of service businesses while people continue to deliver the physical work; the company therefore also needs people willing to stay with hard work. Melisa Tokmak 22:25 “We are building an autonomous enterprise” Direct Audio Anchor Listen from 22:25
Enterprise AI is ultimately bought on observable value
Tokmak rejects the premise that essential-service industries are slow technology buyers. She says large enterprises are focused on value and gives an example of a roughly $500,000 contract closing in 14 days. This is a guest-supplied case, not independent verification of the transaction. Melisa Tokmak 23:53 “it can be like a half a million contract” Direct Audio Anchor Listen from 23:53
On private equity, she argues that the discussion should move beyond demos and cost cutting to whether results persist throughout the year and produce net-new revenue. She says Netic has generated more than $600 million for customers, also a company statement. She closes by naming education as an important positive AI application: technology can widen access, though action remains a human choice. Melisa Tokmak 27:27 “I think over $600 million for our customers” Direct Audio Anchor Listen from 27:27