Your Technology Is Not A Business
What Value-First Founders Do Differently
In 2016, one of the godfathers of artificial intelligence co-founded a company in Montréal designed to turn world-class AI research into commercial products. Element AI eventually attracted $340 million in investment, recruited hundreds of researchers and engineers, opened offices across three continents, and operated within one of the world’s deepest AI talent pools. By late 2020, the company was sold at a discount to ServiceNow, a California-based cloud services firm. Most employees received termination notices. A signed government funding agreement was cancelled. The technology was extraordinary. However, the business never achieved its potential.
Element AI did not fail because its people lacked talent or ambition. It failed because it started with the technology and worked outward, searching for problems its models could solve rather than embedding itself in the value chains where those problems lived. The founders built a research engine and assumed commercial traction would follow. It did not. And the pattern that produced that outcome is repeating itself right now, at scale, across a new generation of AI ventures armed with even more powerful tools.
I made a version of this argument in 2020, presenting to colleagues on the opportunities and challenges facing AI innovators. One slide in that presentation carried a line I still use with founders today: technology innovators must consider how to be the linchpin in their end users’ value creation chain. Six years later, the slide feels less like advice and more like a warning. The tools have improved enormously. The underlying mistake has not changed.
Value first, model second
The mistake is a specific one, and it is worth naming precisely. AI founders, guided by an engineering mindset, overwhelmingly begin with technological innovation. They build a model, refine its capabilities, demonstrate what it can do, and then go looking for customers whose problems might fit what the model already does. This sequence feels logical — but is also backwards. The founders who build lasting businesses start from the other direction. They begin with value innovation: understanding what the end user needs, mapping the workflow the AI will enter, and identifying where in that value chain the technology can deliver an outsized return. The model comes second. The value proposition comes first.
This distinction is not academic. It shows up in how founders spend their first twelve months. A technology-first founder spends that year building a proof of concept, improving model performance, running benchmarks, and pitching demos. A value-first founder spends that year inside a customer’s operation, learning the bottlenecks, understanding the costs, and mapping the points where a well-placed AI capability could change the economics of the customer’s business. The first founder has a better model at the end of the year. The second founder has a business.
The data confirms the pattern at a systemic level. A January 2026 report from Bain & Company and Montréal’s Mila – Québec Artificial Intelligence Institute found that Canada holds roughly 10 percent of the world’s top AI researchers but captures less than 2 percent of global AI venture capital investment. Two-thirds of high-potential Canadian-led startups that raised more than a million dollars were headquartered outside the country. The ecosystem’s response was to launch a new venture capital fund. More capital for more ventures.
The instinct is understandable. It is also incomplete. The gap is not primarily a funding gap. It is a value creation gap. Pouring more capital into ventures that start with the technology and work outward will produce more failed Element AIs, not fewer.
Consider the human dimension of that gap. Nearly 95 percent of AI researchers surveyed expressed interest in entrepreneurship, but the infrastructure to support them remains oriented toward attracting venture capital rather than toward value creation. Brilliant people want to build things that people will use; however, they are funnelled into a system that optimizes for raising money rather than earning revenue. This is the gap.
The ecosystem’s response does not just fail to close this gap; it also fails to address it. It reproduces it, training each new cohort of founders to start with the technology and chase capital before they have earned revenue. The accelerators reward technical sophistication. The pitch competitions reward fundraising milestones. The demo days celebrate what the model can do, not what the customer will pay for. Founders do not independently arrive at the wrong sequence. They absorb it from the system that is supposed to help them.
What value-first founders do differently
The founders I work with who get this right share a set of habits that look unremarkable from the outside but represent a fundamentally different orientation.
They get close to the use case before they build anything. Not close in the sense of reading industry reports or attending conferences. Close in the sense of spending time in the end user’s operation, watching how work actually gets done, understanding where the friction is and what it costs. An AI founder who wants to serve manufacturing does not start by training a predictive maintenance model. That founder starts by standing on a factory floor, talking to the maintenance crew, and learning that the real problem is not predicting failure but coordinating the response when failure is predicted. The model is a component. The value is in the workflow it enables.
They look across the entire value-creation ecosystem, not just the task their model performs. Most AI solutions address a single step in a longer chain: data collection, model inference, or output delivery. Founders who build viable businesses see the entire chain and position themselves at the point where their intervention creates the greatest leverage. That point is rarely where the most impressive technology sits. It is where the biggest cost, delay, or risk sits. Those are different questions, and they lead to different products.
Value-first founders design for the economics of maintenance, not just the economics of deployment. One of the challenges I flagged in 2020 was that AI solutions have a short half-life and need constant maintenance. This remains true. Models drift. Data changes. Regulatory environments shift. A founder who prices only for initial deployment is building a business that becomes unprofitable the moment the model needs retraining. The viable AI business accounts for ongoing costs from the start, building maintenance into the value proposition rather than treating it as an afterthought.
They focus on scalable business models, not scalable technology. A model that can process a million transactions per minute is impressive. A business model that captures value from each of those transactions is what pays the bills.
Too many AI founders confuse technical scalability with commercial scalability. They assume that because the model can scale, the business will scale with it. But scaling a business requires repeatable sales, predictable revenue, defensible positioning, and a customer base that renews. These are business problems, not engineering problems, and they require business thinking from the earliest stages.
None of this means the technology does not matter. It matters enormously. But it matters in the way that an engine matters to a car: it is necessary but not sufficient. A founder who builds the best engine in the world and installs it in a vehicle that no one wants to drive has not built a transportation business. That founder has built an engine. The AI ecosystem is full of extraordinary engines sitting in vehicles going nowhere.
Better tools, same mistake
The agentic AI wave has made this problem more urgent, not less. When the barrier to building a capable model was high, founders who lacked technical depth were filtered out early. Now that powerful foundation models are accessible through APIs, the barrier has dropped. More founders than ever can build something that works. But fewer founders than ever are asking whether what works is something someone will pay for, repeatedly, at a price that sustains the business. The democratization of AI technology has not democratized the thinking required to turn that technology into a viable venture. If anything, it has widened the value innovation gap.
The AI founders who will still be standing five years from now are not the ones with the most sophisticated models. They are the ones who earned the right to scale by first building credibility and capability inside a specific value chain, who understood their end user’s problem before they wrote a line of code, and who built a business that accounts for the real costs of delivering AI over time.
Your technology is not a business. Your business is the value your technology creates for someone willing to pay for it. The founders who learn that early will build something that lasts.
Davender’s passion is to guide innovative entrepreneurs in developing the clarity, commitment, confidence and courage to enter, engage and lead their markets in a world that refuses to hold still, by thinking strategically and acting tactically.
Find out more at https://coachdavender.substack.com/about and https://linkedin.com/in/coachdavender