Building an AI-Native Company: Onboard Everyone Like an Engineer

Written by Ludo Antonov and Casey Winters

At most companies, if you’re not an engineer, you’re frequently stuck waiting: for access, for tools, for dashboards, for decisions. In an AI-native startup, that gap shouldn’t exist. Building in the age of AI is different. The rules—and the tools—are changing fast and with that the day-to-day operations of the team. We take an AI-native approach to running our company to make us more efficient. One rule we’ve recently embraced is to “onboard everyone as if they are an engineer”.

Historically, you can judge the effectiveness of tech companies based on how well they onboard engineers. Meta famously tries to get people to ship changes in their first week during their “bootcamp” that teaches engineers how things work at Meta. At some companies, it can be months before an engineer effectively commits code. Many companies are now attempting to re-onboard their engineers into AI-native engineering, driving adoption of new tools like Cursor, Devin, or Windsurf.

At SuperMe, our Growth Ops person, Anton, joined us a few weeks ago. Instead of the usual maze of tools and docs, we decided to switch things around and onboard them as an engineer. We set up their Github account, downloaded the backend repository, and set them up with Cursor. Our thesis was that instead of building tools or having to go through engineering to get things done, we can instead use our codebase as our internal tools suite. We had a thesis that code-assisted onboarding would unlock a faster, more empowered workflow. So we tried it.

On Anton’s first day, he was able to leverage AI tools to write complicated scripts to retrieve data and get a sense of the current state of the environment. Within his first week, he was able to create a tool set that allowed him to validate and sync account information and discover gaps in both our tooling and setup and validate where things are working fine. At most companies, Ops people depend on engineering and are always begging for engineering capacity. At SuperMe, Ops is doing the engineering.

Technical barriers usually slow teams down or lock them out. Our bet was that AI could flatten those barriers. Instead of manually stitching together workflows with GSuite and having manual tasks take hours or days, many of Anton’s early tasks took minutes. When AI couldn’t get all the way there, it at least covered 80% of the way. That created a much faster ramp to productivity than we’re used to from prior companies. At previous companies, new non-technical hires would surface tooling gaps that kicked off months of work, like building dashboards, configuring Retool for new internal tools, or buying and integrating new analytics or martech stacks. With this approach, we’re no longer stuck stitching together tooling; we’re just shipping.

At SuperMe, we’ve made the decision to onboard everyone like they are an engineer moving forward. The expectation is that you are going to use code-assisted tools regardless of role to directly access the data and automate tasks. You also do not need to understand the full technical depth to get and create a lot of value. Code is leverage, and that leverage is now available to everyone at the company regardless of their engineering experience. In an AI-native world, code is no longer a moat. It’s a multiplier. And that means everyone needs to wield it.

This approach can trigger a lot of red flags in people’s heads: What about database access? This can be very unsafe! What if a mistake happens? Similar to internal tools, it’s important to have a stance on what is safe or unsafe. We do force the usage to happen through our API layer, and that, while not bullet proof, allows us to maintain a level of sanity within the rest of the system. So far we’ve realized that our APIs are our best internal tools, and we’re leaning into our APIs and using AI tooling to generate personalized, on-demand internal tools.

Code is no longer just for engineers. And onboarding like an engineer is no longer just for engineers either. It’s the fastest way to unlock your team’s potential in an AI-native world.

It’s working for us. We think it’ll work for you too.

Anton’s already talked about his experience on his profile if you want to learn more. We’re learning just as much from these new teammates as they are from the tools. And you can always ask Casey AI or Ludo AI our thoughts on building a new startup in general.

Currently listening to my Fauxld School playlist on Spotify.