From Michigan to Silicon Valley: A Conversation With Mohamad Yassine
Back in the year 2000, just after we completed our state of Michigan Y2K remediation, all eyes in the government technology world turned to new Internet opportunities, building websites and web applications, and moving government services like driver’s license renewals and campground reservations online.
In the Michigan Department of Management and Budget (DMB) at that time, we built a “rapid web development team” to bolster our new Internet efforts, which eventually got rolled into our e-Michigan team. Over time, we launched michigan.gov, which was the first state (or local) government “dot-gov” portal.
Fortunately, Ric Tombelli, who was our outstanding webmaster at the time and who worked for me when I was the Michigan DMB CIO, hired Moe Yassine straight out of Michigan State University. Moe was — and is — an energetic entrepreneur, who went on to start his own company before joining Ford Motor Company managing next-generation diagnostics. From there, Moe became the engineering manager over diagnostics for Tesla before moving to Google to become their head of North America OEM engagements for the Android Automotive Partner Engineering and later the head of systems engineering for that same Google organizational unit.
And in January 2024, he became the co-founder and CEO of a new AI company called Predictive Horizons. The firm is now doing groundbreaking AI work that has been highlighted by Inc. Magazine and many others.
I reached out to Moe and asked him for an interview, and he agreed. What follows is that conversation.
Dan Lohrmann (DL): What excites you about applying AI to the auto sector?
Mohamad Yassine (MY): The most exciting thing is really helping the industry as a whole thrive by giving the people on the front lines the tools they need to remove their biggest roadblocks. I grew up only a few blocks away from Ford World Headquarters, and I often tell people that these companies, whether in the United States or overseas, are the lifeblood of so many communities. And one of the most gratifying aspects of this industry is in service engineering and operations. You get to touch a part of the auto sector where you deal directly with real people, whether that be a service engineer, a service technician, a service adviser or the customer themselves. It’s a very humbling experience to be able to contribute to the part of the industry that’s not only underserved but has a massive impact on the overall health of the industry and its ability to innovate.
DL: Where did the idea for Predictive Horizons come from? Why is this company needed now?
MY: Our ultimate goal is to develop a “minority report” for cars that essentially can know what’s wrong with a vehicle before the customer even experiences the symptoms. Between myself, Jason and Dave (my co-founders), we all had different experiences within the auto industry. Jason was one of the early pioneers of using data science to solve these vehicle problems. I remember once he presented a neural network he trained over a weekend to diagnose issues at Tesla, and we all just sat there amazed as he was showing us the accuracy compared to what the technicians were doing, just dumbfounded that he could do something like that over the weekend. Dave is our resident “car guy” who grew up loving cars all his life and built out the service organizations of both Tesla and Rivian. If he’s not with us or his kids, you’ll probably find him under one of his nine cars or figuring something out for a family member’s car. And for me it was really the software engineering and problem solving that I loved about this, the sheer scale of the systems and how best to present information to users, the closeness to the customer experience that just made me love this space. And we all felt like we had unfinished work, and for whatever reason the stars seemed to align when we decided to do this.
In any industry, as innovation accelerates you tend to see quality take the biggest hit; it’s sort of the natural order that this will happen. The auto industry specifically has been going through a lot of innovation lately, but we even see it in other industries. And we see many companies out there who are on the other side of the quality equation, building products and services that ultimately serve product development with the philosophy that if you catch problems earlier in the design process it costs you much less to handle it. This philosophy, however, misses some really important factors: First, that when you are moving to innovate as fast as you can, you will ultimately break things. Second, if you are constantly moving fast, and aren’t able to contain these problems early or effectively, your post-sale costs skyrocket not only hitting your bottom line but also severely impacting your brand loyalty and the customers’ experience with your product. And we see that only accelerating as competition heats up, we only see that automakers will need to “step on the gas” even more.
DL: How is AI risky or overblown in 2026? Are companies overpromising?
MY: I believe AI presents us with an unbelievable opportunity, one that could change the course of history on par with the Industrial Revolution. But I do think we need to also move with caution. To me it’s clear we’re operating in an AI bubble where companies are overpromising near-term capabilities to keep up with the hype. And while everyone is eager to discuss the immediate impacts of AI, there isn’t nearly enough talk about the risks it poses when used irresponsibly even in its current iteration.
I think that many companies are acting irresponsibly in having an overreliance on AI, especially in cases where it could be a life or death scenario or have major impacts on people. We’re seeing AI implemented in areas, such as mass surveillance, where we really need to take a step back and exercise extreme caution. We have to remember that foundational models, the backbone of every AI pipeline, still have a high tendency to hallucinate and lose their grounding in facts. Leaving critical decisions entirely to an AI is a recipe for disaster now and in the future. Ultimately, I believe the best and safest use of AI today is keeping humans in the loop and focusing on gaining productivity. This philosophy not only allows us to protect for peoples’ well-being and jobs, but also gives us a path to accelerate innovation with safety across the board in industries such as mobility.
DL: Can small public or private organizations compete with the big tech companies? Should they?
MY: Today they absolutely can and should compete with large companies. I grew up when the Internet came of age and then the big data revolution and these massive platforms grew. But I can’t remember a time as exciting as today. I see a world where some big tech companies will sort of become the backbone of AI. I think companies like Google, OpenAI, Anthropic, Nvidia will become something like the Internet backbone companies but for AI. And then I think there is a massive opportunity to leverage that backbone to build all sorts of interesting things.
If I were founding another company today, there would be a few types of companies I would personally stay away from. I do think that sort of generic SaaS companies, things like test tracking tools that don’t require any specialization to build, will struggle in the future. I think the old “buy-before-build” mindset is something that is going to significantly shift in these cases because the cost to build and maintain has significantly gone down. I would also not go into building generic “AI tools,” because I think that as the models get better and better there will be a lot of generic use cases that will become foundational. But I think vertical SaaS companies like ours where there’s not a lot of information or expertise out there on a topic, where it’s something that you specialize in, like deep, predictive automotive diagnostics, I think these are the most exciting things to be working on.
And if you have the right talent, without the bureaucracy that big companies have to deal with, you can produce at a blinding pace and build something meaningful.
DL: You have an amazing background at Ford, Google and Tesla, and now as CEO at an exciting AI startup. Is there a secret to your success?
MY: Yes, it was actually something that one of my best friends told me. Early in my career I had a chip on my shoulder. I always wanted to feel like the smartest nerd in the room. But it was something that a good friend told me in my 20s that really changed the trajectory of my career. He said that he always looked to work with people that were smarter than him. And that he gravitated toward that so that he could learn from others.
Over the course of my career I’ve been extremely fortunate to work under amazing leaders and with great people. Every single place that I’ve had the opportunity to work I’ve found people that I could learn from. And from the inception of Predictive Horizons, I sought people out that I thought we could learn from.
And the interesting thing is you don’t always know where that learning will come from. I often tell people about the story of Dave V. (a different Dave), who was my teammate at Ford. I had come on as a solution architect to work on their service information portal. We were both there as contractors, but he and I clashed when I first was there, because I wanted to rewrite and change everything and I was doing it blindly. In hindsight I was really arrogant coming into that role. Later I came to find out that he was sort of the father of Ford’s diagnostic philosophy. And I quite literally learned everything I know today about diagnostics, technicians and service from him.
DL: How should IT and security leaders approach innovation in AI?
MY: I think IT and security leaders should approach AI with optimism and caution. On the one hand I think that because AI can take in vast amounts of information, such as whole code bases, and really synthesize what that code base is doing for you. So in many ways I see that a lot of the basic back and forth that happens between the security organization and the software engineering teams can be significantly improved, making both sides happier and allowing focus on bigger, more challenging problems.
On the other hand I think AI poses major risks for CISOs. In 2025 between 68 and 80 percent of all successful breaches were either an attack on the application layer or through successful social engineering attacks. From a percentage standpoint this falls in line with what we’ve been seeing for years, however, because AI has this ability to automate attacks on a mass scale, the sheer number of successful data breaches and attacks I think will significantly rise if you don’t approach AI with caution.
Additionally, this idea of “cowboy vibe coding” that’s being pushed by some of the big tech players which give the impression that “product builders” are launching things that aren’t reviewed by engineers — I think this is a recipe for disaster. We’ve seen, in multiple instances, where an AI not only doesn’t generate optimal code, but generates specifically insecure code.
So if I were a CISO today I would be looking at where I can leverage AI to improve processes and productivity, but also look at opportunities to shift focus to the areas where we expect AI could have a high impact at introducing insecurity into what’s being built. The tendency going forward is going to be to move faster and faster, and so CISOs today really need to get ahead of the things that are going to be impacting their organizations.
DL: You began your career at the state of Michigan as a web developer. Did you ever think you would be where you are today? How so?
MY: Since high school I always wanted to prove to myself that I could make it with the folks out here in Silicon Valley. Maybe that was for my own ego more than anything. I always loved building things and solving technical problems, so I knew I wanted to end up here. However, I often tell people that my path in life has been full of setbacks along the way and not a straight line in any sense of the word.
For example, not long after I joined Google, I was handed one of the most ambitious tasks possible: launch GM. A few months later, COVID hit, and everything was thrown up in the air. And despite that, my manager, leadership, teammates and our counterparts at GM found a way to push through and launch the program on time starting with the Hummer EV.
When people see that car, they see a Detroit-made EV with crazy capabilities, and we were all incredibly proud of it. But when I look at it, I see the day and night phone calls with the program manager and director at GM. I see myself sleeping on the couch listening to the Google Assistant on a loop, and phone calls with my BD and Partner Engineering colleagues where they had to talk me down from the crazy pressure. I remember fixing benches for colleagues in the garage while trying to maintain distance, driving around with my director performance testing maps. Literally everything was stacked against the Google and GM teams on that program, and I was terrified that if the program failed, it would be my own personal failure.
But as I was driving into work one day after the lockdown lifted, it hit me. There was a massive group of people there to support not just the launch but me personally: product managers who would do whatever was needed to get things working for users but also take the time to just chat; my partners in Partner Engineering (too many to list) who would jump in to troubleshoot anything I needed help with; and engineering managers willing to do whatever it took to make things work well. I realized they had my back no matter what.
So, looking back at my journey from the state of Michigan to startup life, I’d say it only happened because even in the most bleak situations I’ve had people around me who have encouraged me to keep going. The saying goes that it takes a village. I’ve been very lucky in that sense because I’ve always had a city.
DL: What is your vision for where the auto industry can be in five years?
MY: I see an industry that needs to get much closer to their customers than they are today. For many years the auto industry essentially outsourced their customer service to their dealership franchises and that has not always been a model that has worked for them. And to be clear I do believe that the dealership franchise model is important to maintain in the long run, but it needs some tweaks to ensure that the goals are aligned between the automakers and their franchisees. We see some of these new entrants into the marketplace struggle to build up their service networks by owning their sales and service centers directly, trying to replicate the Apple model, but it’s a model that I don’t believe scales in automotive. So while it may work for a very small company, inevitably you see the scalability break down as the automaker’s footprint becomes larger and they start having to rely on third parties for their service. Having said that, automakers really need to find a balance between the scalability of the dealership franchise model and the close customer touchpoints that you get from direct customer interactions.
The challenge in achieving that is really about the scale and complexity of these auto companies and their products and I think that’s where AI and services like ours come into play. I think diagnostic and prognostic software services, over-the-air updates, more direct digital services into your post-sale strategy, as well as more direct contact between the automaker and their customers will not only help get the automakers closer to their customers, but it also significantly improves the fast product development cycles that are inevitably needed to compete with smaller, faster moving entrants into the marketplace.
DL: Where do you want to be in 10 years (professionally or personally)?
MY: Oh, I’m not sure, but I do hope one day to write a book about my experiences and provide my family with a few years of some calm time without them having to worry about Dad’s new hairbrained idea.
About Author
