Plane

AI-first ​

Something fundamental about how work gets done shifted in the last three years. We are going to act like it.

Plane is AI-first, and we mean it literally. Not AI-curious, not experimenting, not a couple of Cursor seats and a ChatGPT account. AI is the base layer everything here runs on: code, design, marketing, sales, support, ops, finance. The people at Plane who internalize that compound their output every month, and we want that to be every one of us.

If you are reading this as a candidate, take it at face value. This is not a trend we are following. It is how we work.

We are not chasing 20%, we are chasing 20x ​

Most companies use AI to do the work they already did, 20% faster. That is not what we are after.

Think scale rather than speed-ups. The real opportunity is to do things that were impossible eighteen months ago, in the time it used to take to write a single email. A few things people at Plane have already built:

  • Agents parked on our competitors' sites that ping us the moment they publish a new page.
  • A support setup that learns from our best answers and auto-drafts replies to new tickets.
  • Daily Slack runs that summarise what moved across the company and nudge the right person.

The upside is real, and most of it is still unclaimed. We would love everyone at Plane to be looking for their own 20x as part of how they work, not as a someday project.

What we expect you to be fluent in ​

If you are joining Plane, the baseline is this:

  • State-of-the-art models. You know the trade-offs between the frontier models: what each is best at, what they cost, where they break. You change models when something better lands, which is roughly every six weeks.
  • Prompt engineering. As a craft, not a buzzword. You can take a vague problem and turn it into a prompt that produces useful output reliably. You version your prompts. You build small evals so you know when one is better than another.
  • Orchestrating multiple agents. You can chain together two or three agents to do work no single model can do alone. You know when to reach for one large model and when to reach for five small ones in parallel.
  • Basic automation. You can wire up Zapier, n8n, MCP servers, custom scripts, whatever it takes, to remove repetitive work from your day. If you are doing the same thing twice a week by hand, that is a personal bug to fix.

These are not things you can pick up in your first year. They are the floor. We assume them on day one, and we expect them to keep growing.

Pick your stack. We pay for it. ​

Frontier AI moves too fast for us to standardize on a single vendor, so we don't. Pick the model that is best for the work in front of you, and switch whenever something better lands.

Every person at Plane gets a generous monthly AI budget. Use it on whatever moves your work forward: Claude, GPT, Gemini, Cursor, Cline, Linear AI, v0, ElevenLabs, Replicate, agent platforms, niche tools we have never heard of. If you can make a case that a tool will save you ten hours a month, the answer is yes. Your imagination is the constraint here, not the budget.

Our stack right now: a snapshot, not a standard. It changes as the frontier does, and so should you.

  • Coding: Claude Code, Cursor, Codex
  • Writing & research: Claude, ChatGPT, Perplexity
  • Design & media: Figma AI, v0, Midjourney, ElevenLabs
  • Automation & agents: n8n, Zapier, MCP servers, custom scripts
  • Internal: Plane itself. We run the company on Projects, Wiki, and Plane AI, and we dogfood Desk before it ships.

How this connects to the rest ​

If the speed-of-light quote on the values page sounded ambitious, this is how we close the gap to it. AI is what lets a small team move at the scale of a large one, and a team that stays hands-on, goes all in, and lives on the edge can move faster still. Use it that way.