After a decade launching products at Google and YouTube, Yann Goarin went looking for the markets Silicon Valley ignores. He thinks the next defensible AI companies will be built in the trillion-dollar industries nobody is looking at, and he is now building the playbook to go after them.
For a decade at Google and YouTube, Yann Goarin worked at the center of the digital economy, running go-to-market across more than 20 product launches and shipping software to millions of users. It was the arena every ambitious product person wanted to be in, and it taught him something that only became useful once he left: the most critical part of building a successful product is rarely the engineering. It is knowing what to build, and for whom.
He left in 2023 to start Zag Labs, an advisory firm that helps startups find product-market fit, and over the next two years worked with more than a hundred early-stage founders across a range of industries. Out of that came a repeatable system for reaching fit and the pattern recognition to walk into a new vertical and quickly find where the pull already exists. It also surfaced an interesting insight: the startups chasing the largest, most obvious markets with horizontal software were the ones grinding hardest for traction, while giant problems remained unsolved in industries very few in Silicon Valley were looking at.
“Everyone is chasing horizontal plays because the addressable market looks bigger on paper,” Goarin said. “That’s a misread. Horizontal AI platforms will not solve legacy industry problems, the domain specificity and tribal knowledge required are too high. What works is operators who understand these businesses and workflows deeply who can productise vertical intelligence. There are dozens of billion-dollar companies waiting to get built. You just need to know where to look.”
Energy, materials, industrial distribution, and much of the physical-world economy are trillion-dollar verticals that have carried the same operational inefficiencies for decades: sourcing, forecasting, inventory, pricing, sales, customer service. The problems were always there, and until large language models and cheap data moved the line they were too complicated or too costly to solve, so most of them went untouched.
There are reasons these sectors have been left alone, and Goarin is candid that none of them are secret. They are relationship-based and closed to outsiders. Sales cycles run long. Buyers are often skeptical about technology and need education, and many have been burned by bloated ERP and CRM rollouts that stretched for years and under-delivered. They are also unglamorous. Silicon Valley’s technical talent has long preferred building for itself and for already-digital industries like technology, finance, media, and travel, and the bulk of venture funding has gone to horizontal software for these mature markets.
To Goarin, that neglect is the opportunity. These are enormous industries with pervasive, unsolved problems, and the technology to solve them finally exists. Capturing them requires an edge most AI builders lack: real knowledge of the space, and a way into rooms that stay closed to outsiders. Whoever holds that starts ahead.
A year ago, he found a place to test his conviction at scale. AAXIS, a 26-year enterprise technology consultancy that has built digital commerce systems for Fortune 500 industrial clients including Ferguson Enterprises and QXO, was steering its core business toward custom vertical AI and standing up a venture arm to incubate AI companies of its own. It named Goarin its first Founder in Residence, leading the development of its first venture. The appointment pairs the firm’s assets, a long industrial client base and an in-house AI engineering practice, with an operator who knows how to find demand and generate early traction. “Yann brings a rare combination of product development and go-to-market expertise that rounds out our technical capabilities and deep industry relationships,” Asad Durrani, AAXIS’s chairman and chief executive, said when the appointment was announced.
Building materials distribution is the first move, with HVAC as the beachhead vertical: a mature, fragmented industry that’s consolidating fast as private equity rollups multiply. These distributors run on software from a handful of old vendors that haven’t innovated in 20 years, leaving operations riddled with inefficiency from top to bottom. Yet it is a difficult space to enter for a young startup. “The only way in is to earn it,” Goarin said. “Build relationships with the buyers and users, spend time on-site, map the workflows, quantify what it is actually costing them, and show them something that works in their environment.”
That inverts how most startups operate. The common path is to build something, then go looking for a market that pulls it out of your hands. Goarin’s approach is the opposite: find where the pull already exists, then reverse-engineer what needs to be built to satisfy it. Customers will only commit time, access, and sometimes even payment as design partners if what gets built is shaped around a need that has been identified, quantified, and demonstrated. “Demand pull beats product push in any market, but the gap is even wider in these legacy industries that are skeptical of technology,” Goarin said.
AAXIS’ first venture, Century Intelligence, was incorporated in April 2026 and just shipped its first product, BranchKinect, a sales automation platform for HVAC distributors. By Goarin’s count, the company signed three design partnerships and is in active discussions with more than a dozen warm prospects. Comparable AI-native competitors in adjacent categories took four to five years to reach roughly twenty customers, a mark Goarin aims to hit in eighteen months.
None of it is specific to building materials. “We can replicate the same playbook in other verticals or industries,” Goarin said. He is already eyeing plumbing distribution as an adjacent vertical, and pricing optimisation within building materials as a second, separate venture. Each starts the same way: a niche where the pain is acute, the pull is real, and few have gotten there yet.
The pattern is legible enough for the founders and investors chasing the same opening: a narrow niche studied for pull, industry insiders who can open doors, design partners who co-build the product, and early deployments that become the case studies that sell the next one. It looks less scalable than a horizontal play at first. But once the product is embedded deeply in an enterprise’s workflows and systems, scale opens in two directions at once: expanding to other problem areas in the same account, and outward into adjacent verticals facing the same problem.
It is a wager against Silicon Valley’s instincts, and on Goarin’s own arc, from the Big Tech to an industry that still runs on 1990s systems and handshakes. “In the age of AI, industry knowledge is more valuable than technical skill for building defensible software companies,” he said. “Execution is commoditised. Knowing what to build and for whom is the scarce resource.”
