Working at Atlan in the AI Era
What we expect. Why it matters. How it connects to how we perform.
Something fundamental has changed about what valuable work looks like. Output is no longer scarce. A competent skill can produce a first draft, a data analysis, a working prototype, an autonomous workflow. What remains scarce is the quality of judgment, the impact of the work, and the reach you create for yourself and the people around you.
We are not a SaaS company adding AI features. We are an AI-multiplied company: every role is expected to use AI to multiply what they do. This document makes our expectations explicit. They connect directly to how we assess performance. At the highest level, we need to be able to answer yes to these three questions:
Has AI expanded what I am capable of, not just what I produce, but the complexity and ambition of the problems I take on?
Has AI created leverage beyond myself: skills, systems, and knowledge that others build on?
Am I learning and rebuilding continuously, staying at the frontier of what is possible so I can keep elevating how I perform on the two questions above?
The baseline: what we expect from every Atlanian
Regardless of role, there is a set of expectations that are not optional.
Use AI in your daily work
You are using them to think, draft, analyze, automate, build, and improve — every day. If you are not, that is a gap, and it is visible.
Understand how AI tooling works well enough to assess the output
Usage without judgment is not the bar. You know when AI produces something good and when it produces something that is merely adequate. You push back. You refine. You do not ship the first output because it looks clean.
Be the human on the loop, not in it
Am I staying involved in this manually out of habit, or because my judgment is genuinely irreplaceable? AI handles the repetitive, high-volume, pattern-heavy work. You set direction, evaluate quality, handle exceptions, and make calls.
Continuously rebuild your own workflows
The tools change. Better approaches emerge. What was a strong workflow three months ago may not be the strongest one now. The expectation is not just to learn — it is to unlearn and rebuild, on a shortening cycle.
Share what you learn — and build on what others share
When you find something that works — a skill, a workflow, a new use case — you bring it back to the team. AI nativity compounds when it is collective. This runs both ways: share what you learn, and actively build on what others share.
How expectations scale with your role
The baseline above applies everywhere. What scales is the depth of transformation we expect.
Roles like engineers, product managers, designers
You build what Atlan sells and ships. Your work is the platform. The AI-native expectation here is deep: zero hand-crafted code as the standard, AI embedded in every stage of the SDLC, judgment about what to build and what to delegate to agents. You are not just using AI to go faster — you are changing what "building" means.
The progression
- Use AI tools fluently in your daily work
- Understand the output well enough to evaluate, challenge, and improve it
- Identify what in your domain should be augmented or automated — and say so
- Contribute context, judgment, and requirements when others are building for your domain
- At the high end: prototype, build, and drive adoption yourself
What quality means now
The failure mode we are guarding against isn't someone who ignores AI. It's someone who is confident, polished, and mediocre — someone who produces clean-looking output and never pushes past good enough to great. Quality has two dimensions:
Judgment: did you bring something only a human could bring?
High-quality work reflects genuine thinking where you challenged the assumptions, pushed past the first draft, applied real domain knowledge, and produced something that AI alone could not have shaped to be excellent.
Ask yourself
- Did I interrogate this output or accept it because it looked complete?
- Did I bring real judgment, domain knowledge, or insight that changed the result?
- Would a thoughtful person in this domain find this genuinely great — or just adequate?
Impact: did it actually move something to a material degree?
Good judgment applied to the wrong problem, or to a problem AI could have handled entirely, is not high performance. Quality also means working on the right things, using AI's leverage to do more, reach further, and drive bigger outcomes.
Ask yourself
- Did this move something forward, or does it just exist?
- Did I use AI's leverage to expand what I could accomplish, or just to do the same faster?
- What changed — for the customer, the team, or the company — because of what I did?
Someone operating at the frontier has rebuilt their entire workflow as many times as needed, is working at a scope that would have required a team not long ago, and is actively pulling others forward — sharing what works, raising the floor for everyone around them.
The mindset underneath all of this: unlearn and rebuild
There is a specific disposition that makes everything above possible: the willingness to let go of what has already worked.
AI capability has a shortening half-life. The tools available today are not the tools we will be using in six to twelve months. The workflows that are strong now may not be strong then. Someone who built a genuinely excellent AI-native workflow last year and has not touched it since is likely already behind.
The expectation is not just to learn. It is to unlearn and rebuild, on a cycle that keeps shortening. That requires intellectual honesty about when your current approach is no longer the best one.
How this connects to performance
These expectations are not separate from how we assess performance at Atlan. They are part of it.
How you work with AI — the quality of your judgment, the impact of your work, the depth of your adoption, the degree to which you are continuously improving — is part of what it means to do your craft well at Atlan today. It shows up in craft mastery. It shows up in your outcomes. It shows up in whether you are making others better.
In Growth Conversations and Performance Reviews, your manager will be assessing not just whether you use AI tools, but whether AI is genuinely changing how you work. There is a meaningful difference between "AI is assisting me" and "I have deployed AI to radically multiply myself." The bar we are setting is the latter.
If you are unsure what this looks like in your specific role, that is a conversation to have with your manager or your People Business Partner. Feeling stretched or in a learning curve is expected; staying still is not.