What My Job Search Taught Me About How Companies See AI

Happy Wednesday! This week:

Let’s get started!

As some of you know, I'm knee deep in a job search, looking for my next role as a communications leader. This fall, I've had a dozen or so interviews — and I'm not counting the first-round screens, just the conversations with hiring managers and leaders. My focus is on digital channel strategy and brand and corporate narrative work, so you'd think AI would come up a lot. It... well, it doesn't. At least not in depth.

That surprised me enough that I went back through my notes from each conversation (with an assist from my AI chief of staff, naturally) to see what actually happened when it did come up. Here are the common themes.

When the role itself was about AI adoption, AI was the whole interview — and those were some of the best conversations I've had all fall. Those interviewers didn't want to hear that I use AI. They wanted to know whether I build with it, what I'd try next, and where I thought the comms function could go. One team started from the premise that most of the barriers to AI transformation aren't technical at all — they're human. Our conversation discussed the work we do generally and how AI acts as an accelerant.

That was the conversation I expected to have — at least in part — across other interviews, whether the thrust of the job is traditional communications or marketing work. But it didn’t.

For these roles, AI came up briefly, if it came up at all. I'd briefly describe how I use it, they'd nod, and we'd move on. That’s somewhat expected in a 30-to-60-minute interview — there's a lot of ground to cover, and a deep dive into my AI workflows can read as overkill when what they need to hear about is the job they're actually hiring for. But when you’re talking to senior candidates like myself, all of whom who can do the actual job, learning about how candidates approach new ways of working is a differentiator, not a footnote.

And that’s what is surprising to me. In these conversations, AI wasn’t seen as a differentiator. AI literacy is almost always mentioned in job descriptions, so — particularly given my background — I expected a little more curiosity and discussion. Many people seemed to see AI as a diversion, hobby, or maybe a side hustle rather than part of the core toolset available to use as communicators. And more importantly, across the board, AI use and literacy felt outside of the scope of what they needed. It isn’t being seen as transformative (and, to be clear, I’m increasing convinced that it is), and deep AI use may even be seen as a negative.

One CEO showed a deep curiosity when AI came up, asking me where are the opportunities, and what are the pitfalls? And how do you get the benefit when your data people want everything automated, without getting led off on a tangent? It was the sharpest AI question I got outside the AI-focused roles. My answer was that the upside is real — audience research and persona building at scale, fast creative prototyping — but so are the risks. AI tends to validate your opinion rather than challenge it, and you can't see that bias happening. Full automation means mistakes compound at machine speed. My rule: AI is the assistant, and the human stays at the center. My own setup won't take an action without my explicit approval.

Thinking back, I really love this question. It gets everything right about AI use — AI can’t replace the judgment or wisdom we bring to the job, but it can help a lot with the work that is more rote, data-heavy, and iterative. Instead of focusing on how I use AI, this conversation — because of the interviewer — focused on when and why I use it. On my part, it’s a stronger way to tell my story; on their part, it’s a better way of finding whether there’s a good fit.

I'm one candidate with one sample, so take this with a grain of salt. But here's my guess: many of the people doing the hiring aren't using AI to transform how their teams work — not yet, anyway. They're using it to clean up edits or run a quick web search. So the market expects candidates to be at roughly that same level, and if you can describe your use that way, you've checked the box. Anything else, though, is a deviation from the actual job and comes off as a net negative.

There's a line I keep coming back to from a PRWeek roundtable back in August sponsored by Meltwater. StubHub's Jill Gonzalez predicted that GEO skills will "soon go the way of putting Microsoft Word on a resume" — table stakes, in other words. She was talking about GEO, but inside the interview room, AI fluency is already being treated that way. Long-term, I agree — we’re all going to be expected to be well-versed in how to use AI, just like we are all the other tools to our disposal. But we’re not there, yet, and that’s concerning. If AI fluency has already become a resume line item, and my experience as an interviewee is any guide, then the industry is glossing over AI adoption. Nobody asks how you use Word because everybody actually knows how. I'm not sure we can say the same about AI yet.

Instead of next steps this week, I'd love your help — I'm one data point, and I want more.

Hit reply and tell me. I read every response.

You can have thousands of people following your work and still need an algorithm’s permission to reach them. Creators know the drill: you post, you pray, you check the numbers, you try again tomorrow.

A newsletter changes that relationship. On beehiiv, you can build a subscriber list you own, reach readers directly, and turn their attention into a sustainable business with subscriptions, sponsors, a global ad network, and more.

Get started for free.

A few weeks ago, I wrote that your next reader might be someone's AI agent. Now someone has tested what that agent actually sees — and the answer is a different email than the one your human reader sees.

Matt Howland, Cordial's chief product and engineering officer, built a test around a fictional home brand: 37 marketing emails, each hiding the same member discount in a different place — the subject line, the body, a banner image, a linked page, an attachment, text shrunk to zero size, custom email headers. Then he connected four AI assistants to Gmail — Meta's Muse, Grok, Claude and ChatGPT — and asked them to find the deal. Across 35 placements, the most thorough assistant found 32. A typical human, Cordial estimates, would reliably catch about 8.

But the assistants didn't agree with each other, either. "Only four placements reached all four agents," Howland writes. Two of the four never saw an offer that lived in a banner image. One skipped linked offer pages entirely. And text hidden off-screen — which no person would ever see — was fair game. Cordial sells email marketing software, so this is a vendor running its own test, and Howland flags it himself: "this isn't a benchmark," and a miss can be "a plumbing problem as much as a reading problem."

For communicators, the part to sit with isn't the score. The same email now has two readers who see two different things — people skim the top and the images, software reads everything, including what people never see — and unless you write for both, you've effectively sent two different messages. That applies to the all-staff note and the press release as much as the retail promo. It also raises a fair question about what we bury. Howland's advice is plain: "Talk to the agent the way you talk to the customer: out loud." Don't "whisper to the agent in hidden text."

Why it matters: As more people hand their inbox to an assistant, what the software picks up may matter as much as what the person sees — and right now no two assistants read the same way.

Your next step: Take your last important email and check where the key ask actually lives. If it's only in an image, a link or the footer, move it into plain text near the top — and into the subject line if it fits. That works for both readers — the person you want to reach and their AI intermediary.

Claude can be your analyst, editor, and strategist.

But most professionals are using it to fix grammar.

These 1,000+ Claude prompts take it from grammar tool to your most powerful AI work assistant.

Sign up for Superhuman AI and get:

Claim your free prompts

*Sponsored: I may earn a commission if you sign up through this link.

Looking for a role at the intersection of communications and AI? Here are some opportunities to check out:

Was this forwarded to you by a friend or colleague? Want to get this in your inbox next week? Click here to subscribe!

Quick question: how can I help?

What workflows are you struggling with? Where does AI still feel mysterious or overwhelming? What has worked that you’d like to share with others?

I’m a builder, and I’d love to help you and the rest of The Comms Stack community find great new ways to use AI.

Reply and tell me.
I read every response.

Even a one-sentence reply helps. For example:
“I wish AI could help me with ______.”

Until next Wednesday,

Dan

Advertise | Subscribe