A marketer who has never built an agent is now a junior marketer. Even with ten years on their resume.
That sounds harsh. I’ve thought about it a lot, and I still believe it.
At Foundation we’ve added an AI fluency screen to every marketing hire. I want to share the exact bar here, because I think every marketing leader is going to need one in the next 12 months… and most of the hiring scorecards I see still look like they were written in 2021.
Why I added a screen
Our clients hire us to get them cited by ChatGPT, Perplexity, Claude and Google’s AI surfaces. The work runs on tools that didn’t exist three years ago. The people doing it have to be comfortable inside those tools from day one.
I kept seeing the same pattern in interviews. Strong resumes. Great portfolios. And then a blank look when the conversation turned to how they actually use AI in their workflow beyond writing a first draft.
You can teach someone a client’s industry. Teaching curiosity is a lot harder. And right now, AI fluency is the clearest signal of curiosity I have.
The bar
Here’s what I look for. A candidate doesn’t need every one of these. But the more they hit, the better.
1. They know what an MCP is and how to use one
Model Context Protocol is how AI assistants connect to the tools you already use. Your CRM. Your analytics. Your docs. A marketer who has connected Claude to their own stack thinks differently about what’s possible.
They can skip the theory. Tokenization and model training aren’t on the test.
2. They’ve used an AI visibility platform
Tools like Profound, AirOps and LetAIDo are how our industry measures whether a brand shows up in AI answers. If a candidate has logged in, run a prompt set and looked at citation data, they understand the job we do.
3. They’ve invested in learning it
An AI marketing course. A certification. A paid cohort. The specific program matters less than the signal. They put their own time or money on the line to get better.
4. They’ve built an agent
This one separates people fast. Someone who has built even a simple agent, one that researches a topic, drafts a brief or monitors a subreddit, has crossed from user to builder.
That’s the person I want on the team.
How the bar changes by seniority
The same four signals apply at every level. What changes is how deep I expect someone to go.
1. Early career
For someone two years into their career, curiosity is the whole test. Have they tried the tools? Have they built anything at all, even something small and messy? I’ll take a coordinator with one scrappy agent over one with a perfect resume and zero experiments.
2. Mid-level
A strategist or manager should be using AI inside real client work. I want to hear how it changed the way they research, plan or report. They should also have opinions about which AI visibility tools are worth paying for.
3. Senior and leadership
A director or VP should be able to explain how AI changes the economics of their team. Which work gets faster. Which roles change. Where human judgment still matters most. If a senior leader can’t connect AI to how the team runs, they’ll struggle to lead people who can.
The questions I ask
These go into our interview scorecards for marketing roles.
- Walk me through the last thing you built with AI that saved you real time. What did it do?
- Show me an AI workflow you run every week. What’s connected to it, and what does it pull from?
- Show me how you’d check whether a brand shows up when someone asks ChatGPT for a recommendation in its category.
- What AI course or program have you taken, and what’s one thing you changed after it?
- Tell me about an AI tool you tried and dropped. Why did you drop it?
The first one is my favourite. Anybody can list tools. The people who’ve built something that saved them real time are the ones who’ve actually done the work.
What great answers sound like
The questions are simple. The answers tell you almost everything.
On the first question, a weak answer describes using ChatGPT to write a first draft. A great answer describes a workflow. “I built a prompt chain that pulls our sales call notes, finds the objections that come up most and drafts a FAQ page every two weeks.” Specific input. Specific output. Real time saved.
On the workflow question, I’m listening for whether they’ve actually connected something. “I hooked Claude up to our analytics and I ask it questions about last week’s traffic” beats a textbook definition every time.
On the visibility question, the best candidates start running prompts. They pick a category, ask ChatGPT and Perplexity for recommendations and notice the differences between the two. Bonus points if they ask which sources each one cited.
On the dropped tool question, I want a real opinion. “It was slow, the output needed too much editing and our team stopped opening it” tells me they judge tools by results.
The risk I’m watching
There’s a real trade-off here and I want to name it.
I grew up in Preston, Nova Scotia. Nobody was handing me a playbook, and I couldn’t have afforded a fancy course early in my career. Curiosity and internet access were the whole toolkit.
So I don’t want this screen to become another credential wall. That’s why the bar is built around what someone has done, and why free paths count. Free tiers. Open-source agent frameworks. YouTube tutorials. If you built it on a laptop at your kitchen table, it counts.
Rolling this out on your team
Here’s how I’d add an AI fluency screen if you’re starting from zero.
1. Put it on the scorecard
If it isn’t scored, it gets skipped. Daria, who runs recruitment at Foundation, is folding this bar into our screening questions and interview scorecards so every interviewer measures the same things.
2. Train the interviewers first
Your interviewers need to know what a good answer looks like. An interviewer who has never built an agent can’t tell a real one from a buzzword answer.
3. Ask for a demo
A five-minute screen share beats a 30-minute conversation about AI. Ask the candidate to show you something they built. Then ask them to change one thing about it live.
4. Hold the current team to the same bar
A hiring bar that only applies to new people creates two classes of employee. If AI fluency matters for new hires, it matters for everyone. Give the people already on your team the time and budget to get there.
Credentials are nice. But curiosity is what actually moves the needle.
If you’re hiring marketers right now
Add one AI question to your next interview. Ask the candidate to show you something they built.
You’ll learn more in five minutes than you will from the whole resume.