Serval - Operating Principles- Serval - AI Agents for IT
Serval’s Three Operating Principles
Jun 1, 2026
Jake Stauch and Alex McLeod
We don't have a culture deck. We have three operating principles.
Most companies write a culture deck and then stop. Some pay exorbitant amounts of money to outside consultants to help them set values and then forget what they are a week later. We have three principles and apply them tirelessly. They're how we make real decisions on a Tuesday afternoon when something is on fire. And, we cannot promise they won’t change - because if we don’t iterate, then we aren’t thinking hard enough about this.
1. We are negative maintenance
This one was something Varun and Clay taught us. You can see his original explanation here.
Every system you touch, whether a team, a process, a doc, a meeting, a line of code, or a hiring loop, should leave the company lighter, clearer, and more reusable than you found it. The default in most companies is “low maintenance”. You keep things running with minimal overhead. Our default is negative maintenance. Leave it better than how you found it.
That cashes out in four habits. Remove drag means we simplify, automate, document, and fix root causes, not symptoms. Protect trust means we prioritize customer over team over individual, and we say what we actually think. Candor is kindness. Vague feedback isn't. Kill process means we keep only the rules that obviously pay for themselves. Scale clarity means context beats control. The job of leadership is to make people less dependent on us, not more.
2. We pull timelines forward
The default question at most companies is "when can this be done?" Our default question is "what would have to be true to do it sooner?" Those two questions produce different organizations.
The first one ends in a Gantt chart. The second one usually ends with someone realizing the blocker is one stakeholder, one missing piece of data, or one decision nobody wants to make. We compress loops by questioning delay by default. We choose momentum by starting smaller, shipping the first 80%, and learning before polishing. We use autonomy wisely, which means moving fast on recoverable decisions and slowing down on the ones we can't undo. And we coordinate directly. We align on goals and constraints, not committees and theater.
When in doubt, ship the smaller version to a few trusted partners and let them teach us what we missed because learning is faster than guessing.
3. We raise the bar
Every meaningful hire and every meaningful deliverable should leave the company stronger than it was. Raising the bar isn't a slogan but is a forcing function.
Talent density is the operating reality: every hire makes the company better, or it doesn't, and there's no in-between. The best perk we offer is the team you get to work with. Performance clarity is the second piece. Nobody at Serval should be surprised by what we think of their work, ever. If you're surprised, we failed, not you. Pay for impact means compensation reflects value, scarcity, and retention. It isn't a formula and it isn't a function of tenure. And grow through challenge means scope grows the same way comp grows: by raising the bar, not by waiting your turn.
This is the part that gets the most pushback. People want predictability in compensation, and we understand the instinct because that is how we’ve been conditioned to operate. The companies we want to build aren't the ones where the comp curve is the same shape as the calendar. They're the ones where the best people are stunned at how fast they grew, and the rest of the world is stunned at what we shipped.
How these stack
Negative maintenance compounds into leverage: every system you touch gets lighter. Pulling timelines forward compounds into momentum: every loop you compress gets faster. Raising the bar compounds into talent density: every hire makes the team stronger. We think those are the right principles today. And when they stop being right, we’ll write the next version.