In the early 1990s, I wrote a high school essay arguing that artificial intelligence would fundamentally transform how we protect people and assets. My teacher gave me a decent grade and politely suggested I consider more practical career paths.
The vision I outlined autonomous devices performing human work everywhere was considered fringe. Barely coherent, even. The technology didn't exist to make it real, and the market had no language for it. I kept going anyway.
Why Conviction Matters More Than Timing
The failure mode most founders make is not being wrong about the vision it's abandoning it when the timing is off. I spent decades being early. That is categorically different from being wrong. The question was never whether autonomous security would arrive. The question was whether we'd be the ones building it when it did.
We are.
Today AITX has over 1,000 autonomous units deployed at sites across North America. The technology I described in that high school essay is now running 24/7 at convenience stores, university campuses, corporate headquarters, and distribution centers. The vision didn't change. The world caught up to it.
What the Mainstream Gets Wrong
Now that autonomous security is mainstream, I notice something irritating: everyone claims they saw it coming. The conferences that barely acknowledged us five years ago now dedicate keynote stages to our category. The investors who passed on our early rounds now position themselves as AI-first security specialists.
That's fine. The market validating the thesis is the point. But there's a lesson buried in all of this that the newly-converted tend to miss: the reason this took 30 years isn't that the idea was wrong. It's that being right without resources, without investors, without industry validation and staying right is genuinely hard.
Most people who had the right idea in the 90s didn't make it to 2024. The conviction compounds. The setbacks erode it. We chose not to let them.
What Comes Next
The autonomous security category is real now. But the market is still in early innings. The current deployment model robots at fixed sites, operating on human-monitored loops will look primitive in a decade. Fully agentic systems, integrated into building infrastructure, operating without human oversight for routine events, are the next frontier.
We've been building toward that since 2016. We're still early. And we're still here.