Most "lessons from founding a startup" essays are written by people looking back at the work from the other side of a big exit. This is the version from someone still in the middle of it.
I didn't start this company because I saw a hole in a market. I started it because I got tired of watching the game played badly — technology aimed at whoever would click the loudest, built by leaders who confused extraction for innovation. We experienced a tragedy, and out of it came Tactical Talk and other companies. That's the ceiling I want to put on that part of the story. The lesson under it is the one I care about: if you're going to spend the best years of your life building something, build something that leaves the field better than you found it.
Here's what that actually looks like in practice.
Lesson 1: Decide what you won't build before you decide what you will
Every product decision is also an ethics decision, whether you name it that or not. In an AI conversation product, the default temptation is to build the most powerful manipulation engine you can ship — "win every argument," "close every deal," "never lose a fight." That sells. It also poisons the thing you're supposed to be improving.
We wrote down what we would refuse to build before we got too far into what we would build. No manipulation-as-a-service. No dark-pattern coaching. Detection over dominance. That list is more load-bearing than any feature roadmap I've written.
Lesson 2: A partner who thinks differently than you is worth more than any hire
I'm a builder. I see a problem and start making things. My partner Tavish is an operator — she sees the chaos a builder creates and turns it into a system that other people can actually use. Every founding pair needs both modes. If you're both builders, nothing ships on time. If you're both operators, nothing gets built.
What took me longer to learn is that the real unlock isn't just division of labor — it's having someone in the room whose instincts disagree with yours. Every meaningful call we've made got better after a disagreement got aired out instead of skipped.
Lesson 3: Lead by example, not by memo
"Build with ethics" is the kind of line that lives on an About page and dies the moment an incentive points the other direction. The only way it survives contact with real decisions is if the founders behave like it's non-negotiable when no one is watching.
That means shipping the safer default even when the aggressive default would test better. It means owning the mistake in public instead of laundering it through a PR line. It means paying attention to the invisible things — what data we store, how consent is asked for, what the product nudges users toward — because those invisible things are the product, whether customers ever see them or not.
Lesson 4: Users don't care about your tech stack
You can spend a week optimizing something deep in the system and have nobody notice. Then you make the AI's voice one notch more natural and people email you. Build what users feel, not what engineers appreciate. The stack exists to serve the experience, not the other way around.
Lesson 5: Ship, then listen — don't listen, then ship
Perfection is a trap. It's an especially elegant trap because it looks like conscientiousness. The fastest way to learn what your product actually is, is to put it in front of real people and watch where they get stuck, confused, or surprised. Every week you delay a launch is a week you're guessing instead of learning.
The discipline isn't "ship anything." It's "ship something real enough to get honest feedback, then change it." An imperfect thing in front of users beats a perfect thing in your head every time.
Lesson 6: The hard part isn't building — it's deciding what not to build
Every feature you add is a feature you maintain, document, support, explain, and eventually untangle. The most valuable word in a founder's vocabulary is "no." Every "no" protects the one or two things the product is actually supposed to be great at.
We keep asking one question: does this make the core experience — real-time conversation coaching done ethically — meaningfully better? If the answer is "kind of," that's a no.
Lesson 7: You are building the culture whether you mean to or not
Every decision, every Slack reply, every customer email, every rushed tradeoff — you're teaching the people around you what's normal. If you want a company that leads by example, the example starts with you, on a Tuesday, when nobody's watching.
The honest truth
Building a company is long, unglamorous, and full of days where the most important thing you do is not quit. The reason to keep going isn't the exit. The reason is that for once you get to point technology at something better than widening the gap. If you're considering it: go. Ship something real. Listen harder than you talk. And decide early what you refuse to be, because the market will tell you every day to be it anyway.
— Michael Moosbrugger, Founder & CEO, Tactical Talk