SwitchPoint CEO Ray Guzman on venture building, the messy truth of innovation, and the timeless principles behind building what matters.
What if the most interesting AI companies are not the ones moving fastest, but the ones thinking longest?
It’s a thought that surfaced during recent discussions I’d been having with the SwitchPoint Ventures team about the company’s evolution. When I sat down with CEO Ray Guzman, we assessed how the team’s approach to building with AI had evolved, and why the usual startup categories no longer seemed to fit.
Outgrowing the old label
First, there was this: SwitchPoint had outgrown the language it was using to describe itself. What should they even be calling themselves nowadays?
Up to that point, SwitchPoint had referred to itself as an AI venture studio, which worked just fine for awhile. The team had launched multiple companies around AI and data science, enjoyed four successful exits, and was increasingly sought after for innovation partnerships by major healthcare systems and enterprise operators.
But the broader the company’s work became, the less the old label seemed to fit. “AI venture studio” started to feel too narrow and, frankly, a tad indistinguishable from the AI blare now coming from every direction.
The reality was that the company had matured significantly since it began as an early AI pioneer, over a dozen years ago, long before AI or Nashville became fashionable. Some of the work now sitting in SwitchPoint’s pipeline looked less like traditional venture studio creation and more like the slow, complicated work of building enduring solutions to entrenched problems, as well as the heavy lift involved with helping large organizations take the risks of innovation.
The old venture “studio” category no longer described the full shape of the business. The phrase that kept coming back was venture builder.
The tools evolve
“Venture builder” captured a nuance that was missing. SwitchPoint wasn’t simply incubating ideas, creating marketable prototypes, and moving on to the next concept. The team was digging into the long-term work of building only what mattered. That meant pressure-testing not just whether something could launch successfully, but whether entirely new approaches to longstanding industry problems could actually change outcomes in the real world. The emphasis was never on generating startup volume, but on building solutions and companies that held up over time.
Ray summed up the thinking behind this outlook. One of the problems with today’s buzzy tech culture is that people tend to assume that AI changes the fundamental nature of building itself. As the SwitchPoint team sees it, this is wishful thinking and a clear mistake. The tools change and the speed accelerates, but building anything meaningful still requires the same inconvenient steps for success: patience, iteration, operational discipline, flexibility, and the ability to wrestle with difficult problems long after the novelty has worn off.
The philosophy also explains why the new SwitchPoint site (under construction) might look a bit unexpected.
During the rebrand process, we found ourselves pulling references from a different visual world than what you see everywhere with AI. Sleek dashboards, glowing webs of dots and lines, futuristic AI stock imagery…that wasn’t SwitchPoint.
Instead, the mood board began evoking an era the team felt a kinship with: mid-century industrial America, a golden age of optimism and faith in building great things. Our conversations took us back to NASA mission control, Skunk Works, old IBM campaigns, the idealism that brought landmarks like the Empire State Building, Golden Gate Bridge, and Chrysler Building to life, and grainy photos of engineers in ties standing around drafting tables solving problems, the sort of world that felt part architectural archive, part Mad Men episode.
The retro feeling reflects the philosophy behind how the SwitchPoint team approaches its work: the tools may shift, but the principles required to build great things are timeless.
Building what matters
“Venture building,” as Ray describes it, isn’t about launching cool tech startups.
He doesn’t talk like a typical tech exec. Instead of startup theatre, with its emphasis on words like “disruption,” “scale,” or “market capture,” Ray looks behind the scenes, where tougher words, like “operational friction,” “systems thinking,” and “execution,” sit. He’s deliberate and contemplative, prone to stop in the middle of a thought, reconsider the premise entirely, and circle back twenty minutes later with a better articulation of the idea.
Likewise, the SwitchPoint team as a whole doesn’t project the hyper-curated certainty that has become standard in the tech world. The team talks openly about false starts, difficult pivots, dead ends, changing assumptions, and projects that took far longer than expected. They’ve learned from hard experience that meaningful innovation rarely happens in a straight line.
“It’s easy to think of innovation as the big idea. The breakthrough moment,” Ray observed. “But most of the time, it’s actually disciplined iteration. Testing, refining, failing, starting over.”
This isn’t how most of the tech world thinks.
Everybody’s in AI. Moving fast, revolutionizing their sector, and redefining the future of something. Spend fifteen minutes on LinkedIn and you’ll encounter enough “industry disruption” to last several lifetimes.
Ray sees it. “I can’t tell you how many people tell me they’re in AI now,” he laughed. “And they don’t even really know what that means.”
He isn’t dismissive of AI itself. What bothers him is the tendency to confuse acceleration with understanding. Most companies, in his view, stop thinking too early.
He brought up the “five whys” framework during our discussion, the exercise where you keep asking why until you uncover the real structural problem underneath the visible one. Most organizations never get that far. They identify the obvious pain point, build a product around it, and start selling.
Sometimes that works; other times, it just creates a more efficient version of a broken system. That distinction has shaped nearly every company SwitchPoint has built.
Years ago, before predictive healthcare AI became all the rage, the team started working on sepsis prediction. At the time, many people believed sepsis couldn’t be predicted. They went after it anyway.
Later came Polaris. At the time, conventional wisdom in healthcare was that staffing shortages were the existential problem facing hospital systems. But as the team looked deeper, they uncovered the real challenge. The issue wasn’t simply a shortage of workers; it’s that the workers themselves were often being deployed inefficiently. Hospitals were compensating by overstaffing in some places, understaffing in others, and relying heavily on temporary labor to fill gaps at enormous cost.
Everybody else was solving for hiring because hiring was the visible symptom. SwitchPoint kept digging until they found the structural issue underneath it. The work evolved into Polaris, which was later acquired.
Ray returns to that idea constantly. At one point he asked a question that probably explains the company better than any mission statement could:
“Are people actually innovating? Or are they just scaling broken processes and calling that progress?”
Building only 1 of every 50 ideas
One of the most revealing moments came when Ray started talking about Winnow, a company SwitchPoint sold to Aya Healthcare a couple of years ago. Most founders would probably tell that story as a triumph of speed and execution. Ray framed it differently, pointing out that the idea spent close to four years in various stages of exploration before the company was seriously built.
Those years were not spent polishing pitch decks or trying to manufacture momentum around an idea that was still half-formed. They were spent talking with operators, questioning assumptions, discarding approaches that initially sounded promising, and repeatedly circling back to a more uncomfortable question: was this even a venture-scale problem to begin with?
In many venture environments, a timeline like that would probably be viewed as drift, indecision, or pipeline failure. At SwitchPoint, Ray described it as part of the work. The goal was never to maximize the number of companies launched, it was to build companies that actually deserved to exist.
“We’re not shortcutting it,” he said.
That’s because SwitchPoint is focused on endurance, not volume.
At one point Ray mentioned that most of the ideas they explore internally never become companies at all. “Forty-nine out of 50 ideas don’t survive,” he said, without batting an eye.
It’s an astonishing number that would make most VCs sweat.
In a more traditional venture environment, that kind of ratio would raise serious questions. But listening to Ray talk about it, you see how SwitchPoint views the math. Most ideas, once you spend enough time with them, reveal themselves to be smaller than they initially appeared, or operationally unrealistic, or simply not important enough to justify building an entire company around them. The filtering is not separate from the work; it is the work.
Again, not exactly how most startup decks are written.
No venture clock = freedom to innovate
Part of what allows SwitchPoint to operate this way is structural. The company doesn’t run a venture fund, which means there’s no LP deployment clock, no pressure to put capital to work on a fixed timeline, and no incentive to manufacture momentum for the sake of optics. That creates a very different operating environment, one where patience is not just philosophical posture but an actual operational reality.
As Ray sees it, that freedom is essential for their ability to innovate.
Take, for example, a project the team is currently developing. Most startups at that stage would already be talking aggressively about pricing strategy, TAM calculations, and monetization models. Ray said they still hadn’t fully figured out pricing yet because they were still trying to understand what the thing actually wanted to become.
That answer would not go over well in certain venture capital circles, but it’s revealing about how SwitchPoint operates.
“If we start with the model first,” Ray said, “it’s going to hamstring the way we look at the world.”
Most companies start with market size and work backward toward a product. SwitchPoint seems to start with a different question entirely: does this deserve to exist?
You might call it radical intellectual honesty.
As Ray sees it, venture building isn’t about the endless chase for whatever category happens to be trendy, but about building things carefully enough, patiently enough, and rigorously enough that they still matter after the excitement fades.
As we were wrapping up, Ray mentioned that another founder had recently told him, “You guys are winning.”
Ray shrugged. “I don’t know that winning is the right word. We’re just playing a different game.”
It was the best description of SwitchPoint Ventures that I’d heard all day.