Everyone's telling you that you need AI in your content marketing. So you open a req for an AI content marketer, or an AEO specialist, or whatever the title is this quarter. The problem isn't that candidates will lie to you in the interview. The problem is that this role is new enough that your version of the job and their version of the job might genuinely be different things, described with the same words. Part of your job in this interview is learning the space well enough to notice the gap.
There's a specific candidate profile worth watching for. They're genuinely skilled with prompts, sometimes impressively so. They can produce a lot of content, fast. But when you look closer, all of it traces back to the same well: your old blog posts, your existing website copy, whatever's already sitting in a Google Drive folder, or a general web search on the topic. Nothing new is entering the system. It reads like content, but it's recombination, not reporting.
This isn't usually a character problem. It's that AI has been doing the heavy lifting for this person for so long that they've never had to build a process for finding new material. And AI is only as good as what you feed it. A prompt fed only your past content will keep producing variations of your past content. It won't tell you what your sales team is hearing on calls this week, or what a prospect actually asked in a demo yesterday. Only a person can get you that.
The mechanical side of this job is genuinely easy now. Getting your ICP documented in a format AI can actually parse, structured fields and ranked lists instead of paragraphs, mapping topic clusters, feeding that structure into a tool like SEMrush or HubSpot so it tells you what to write for SEO and AEO, none of that requires much specialized skill anymore. We built our own AI-Ready ICP / Persona Template for exactly this, and honestly, a candidate could prompt their way to something similar in an afternoon. A candidate who's fluent in that part isn't wrong for the role, but that part was never the hard part.
The hard part hasn't changed, and AI hasn't solved it. Where does new content actually come from? It has to come from humans: your sales team, your leadership, whoever is closest to what your buyers are actually asking right now. A good AI content marketer's calendar should be full of short conversations, a call with a rep about what objections came up this week, ten minutes with your CEO about why you built something the way you did, a recorded meeting turned into a transcript that becomes the raw material for a blog post or a script.
That transcript matters more than people expect. AI defaults to generic, sanitized language when it has nothing specific to work from. Feed it a real conversation, in your team's actual words, and it starts sounding like your brand instead of every other company's AI-generated blog. The starting material determines the output more than the prompt does.
Listen for nuance. If the answer is that AI basically solves the whole problem, that's a candidate who hasn't run into the limits yet, or isn't being straight with you about them.
This is the core question. You want to hear about ongoing conversations with real people inside the business, not a workflow that starts and ends with your existing Drive folder or a general web search.
A strong answer sounds like a standing cadence, recurring time with specific people, not something that happens once during onboarding and never again.
Good candidates can describe their discovery questions specifically: what came up on calls this week, what a prospect got confused about, what a customer said that surprised you. Vague answers here usually mean thin follow-through in practice.
You're checking whether they have a real pipeline: recording, transcript, structuring, drafting, brand-voice pass, versus a vague gesture at "then I write it up."
Listen for an answer rooted in using real source material and your team's actual language, not just a claim that they'll "adjust the tone" after the fact.
This checks basic fluency without over-indexing on it. It's a floor, not a differentiator. If this is the most confident part of their answer, keep probing the earlier questions.
A scenario question surfaces everything above at once, whether they'd go talk to sales first, what they'd ask, and whether the resulting content plan sounds specific to your business or generic to the category.
The uncomfortable truth for anyone hiring this role is that AI didn't remove the oldest problem in content marketing, it just changed what solving it looks like. You still need someone who can get the right people talking, capture what they say, and shape it into something worth reading. The candidates worth hiring are the ones whose process starts with a person, not a prompt.
This pairs with a broader question worth asking before you hire any AI-labeled GTM role. If you're also evaluating hires outside content specifically, 7 Questions to Ask Before You Hire Your Next Go-to-Market AI Hire covers how to vet real AI experience versus recently-learned tooling across sales and RevOps roles more broadly.