AI proficiency in the workplace is no longer a nice-to-have. At Ramp, a $32 billion fintech startup, it’s becoming a condition of employment.

Geoff Charles, the company’s chief product officer, said it plainly on a recent episode of the “Behind the Craft” podcast. “If you’re not using Claude Code this year, no matter what your role is, you’re probably underperforming compared to others at the company.” He wasn’t talking about engineers. He was talking about everyone.

Ramp isn’t alone. Google told non-technical staff in February that AI use would factor into performance reviews. Atlassian cut 1,600 roles last week, citing AI as part of the reason. Block laid off nearly half its workforce last month. The pattern is hard to ignore.

So what does the ladder actually look like?

Charles frames AI proficiency in the workplace across four levels. At the bottom, level zero, are the people who “sometimes use ChatGPT.” They open it occasionally, perhaps to draft an email or look something up. However, Charles was direct about what happens to them. “The people who are still in L0 will most likely not be at the company.”

Level one is where things start to get serious. Specifically, these are employees who have built custom GPTs and have some hands-on experience with tools like Claude Code. In other words, they’re not just using AI. They’re starting to shape it around their work.

Level two is the vibe coders. These are people who can build apps that automate parts of their job, even without a traditional engineering background. As a result, at Ramp, 50% of the company’s code is already built by AI. Charles expects that number to hit 80% soon.

Finally, at the top, level three, are the systems builders. People who don’t just use AI tools or build with them, but instead design the workflows and infrastructur e that others operate inside of. These are, by Charles’ framing, the most valuable people in an AI-native organization.

Ramp’s framework might be the most explicit version of this conversation, but the expectation itself is spreading fast.

In February, Google told non-technical employees that incorporating AI into their daily workflows was no longer optional. Moreover, in some cases, managers made clear that AI usage would factor into performance reviews later this year. This wasn’t a message for engineers. It was for everyone.

Checkr, a background-check startup, took a different approach. Rather than issuing mandates, CEO Daniel Yanisse gave every employee a monthly stipend to explore AI tools, along with company-wide AI days and demos. The result was striking. “After one year, 95% of the employees use prompting daily,” Yanisse said.

Meanwhile, companies that haven’t moved fast enough are making harder calls. Block cut nearly half its workforce last month, citing AI advancements as a key factor. Atlassian laid off roughly 1,600 people, about 10% of its global staff, as it reorganizes around AI development. “It would be disingenuous to pretend AI doesn’t change the mix of skills we need or the number of roles required in certain areas,” CEO Mike Cannon-Brookes wrote to employees. “It does.”

The message across all of these companies is essentially the same. AI proficiency in the workplace isn’t a bonus skill anymore. It’s a baseline.

The question, then, isn’t whether AI proficiency in the workplace matters. It clearly does. The question is where you sit on the ladder right now and what moving up actually looks like in practice.

Level zero is the easiest place to leave. Start by picking one AI tool and using it daily, not occasionally. Claude, ChatGPT, Gemini — it doesn’t matter which one. What matters is that it becomes part of how you work, not something you open when you’re stuck.

From there, level one is about customization. Build a custom GPT around a task you do repeatedly. Set up a Claude Project with context about your role. The goal is to stop treating AI as a search engine and start treating it as a collaborator that knows your work.

Level two requires a shift in thinking. Vibe coding sounds technical, but at its core it’s just using AI to build things that save you time. A simple automation. A workflow. A tool that handles the repetitive parts of your job so you can focus on the parts that actually require you.

Level three, systems building, is where most people aren’t yet. Furthermore, it’s also where the most job security lives. People who can design AI-powered workflows for entire teams are, right now, among the hardest to replace.

Charles put it simply. “If you’re not a self-starter and you don’t have that growth mindset, it’s going to be very, very hard to train.” That’s not a threat. It’s just where things are headed. The ladder exists whether or not you decide to climb it.