Happy Wednesday! This week:
Rate yourself on Zapier's AI adoption scale
A webinar on AI leadership in comms — and how you can join, free
Why most comms teams are missing the AI brakes — and a scare from Wall Street
Let’s get started!
THE LEDE
💡 Rate Yourself on This AI Adoption Scale
As someone who's job hunting right now, and someone who has leaned heavily into AI, a recent LinkedIn post from Ben Newman jumped off the page at me. Ben is the founder of Shaper Collective, an AI-native PR firm in Washington. He read through 133 open job postings from 43 DC firms and turned what he found into a hiring matrix for PR firm leaders. It's worth a read on its own — but what grabbed me first was where Ben got the idea.
He built it on Zapier's AI Fluency Rubric. Zapier published the latest version back in March, and somehow I'd missed it. (If you did too, I guess you're in good company!) What I love about it is how straightforward it is: Zapier wrote down exactly what it looks for at each step of the AI learning curve, and then made that the bar for every new hire.
It's the clearest yardstick for AI adoption I've seen — and before you use it on anyone else, you should use it on yourself.
The four things Zapier measures
Zapier scores AI fluency on four components. Mindset is continuously learning and experimenting with AI. Strategy is deciding where AI belongs in the work and why. Building is partnering with AI to produce quality results through clear communication and iteration. And accountability — new this year — is applying human judgment to define success, evaluate outputs and own the results. It comes with the best line in the whole piece: "With AI, you can delegate the work, but not the accountability."
I think of AI as an accelerator. The first three components are about pressing down on the pedal: finding where AI makes you faster and better, then building the systems that keep you there. Accountability is knowing when to hit the brakes — when a draft is off, when a fact needs a second look, when something shouldn't go out the door at all. You need both. And in comms, where one bad send can become the story, the brakes matter more than in most jobs.
The four levels
When hiring, Zapier maps those skills across four levels: Unacceptable, Capable, Adoptive and Transformative. Capable is the minimum to get hired, and it goes way beyond where most of us were at even six months ago. Capable now means AI built into your core work, repeatable systems instead of one-offs, and clear impact you can point to. Zapier breaks the levels down by department, and the Marketing examples are the closest to our world. Here's my translation for comms:
Unacceptable: You use AI for first drafts, and that's it. The output reads like AI, and your work looks the same before and after.
Capable: AI is part of your everyday work, and you've built a prompt library your team actually pulls from.
Adoptive: You've built systems that handle the routine work, some of them without you touching them. Your team stopped doing that work by hand.
Transformative: You've redesigned how your team works — what gets automated, what people own and how success gets measured.
Zapier also looks at what it calls your "slope": what you tried, what you dropped and how your approach has changed. Where you're headed counts as much as where you are.
Where I land
I rated myself Adoptive, leaning toward Transformative. I've built systems that run a good chunk of my week on their own: a morning briefing that's waiting on my phone before I'm up, and a weekly sweep that pulls the AI and comms news worth your time into a ranked list for this newsletter. I still decide what makes the cut — that's the brakes part.
What keeps me from Transformative is simple. Right now I'm a one-man band between roles, so I haven't had the chance to redesign how a whole team works, or to bring a team along on my own adoption curve (although I like to think this newsletter has inspired a few of you 😀). Zapier is clear that this matters: a manager who is fluent personally but whose team still works the old way doesn't meet its bar. (And individual contributors are similarly expected to lean in and empower the team with their own learnings.) That's the work I'm most excited to take on next, so if that gap sounds familiar on your team, I'd love to hear about it.
Where comms lands
Ben's research suggests most of our industry hasn't set a bar at all. When I asked him what jumped out from all those postings, he said:
"Only 14% of associate postings mention AI skills, against 43% for directors. Fluency must exist across the organization from the senior AI strategists leading the transformation to juniors close to the work."
So how do you spot fluency in an interview? Ben's test is a good one:
"I'm always looking to see if people have a framework, so I'd ask a candidate to explain the pieces of an AI agent. If the answer is 'Claude' or 'ChatGPT,' they use AI. If they describe AI like LEGOs, separating interface, compute, harness, tools and skills, they understand it. That understanding becomes a framework they can apply to new problems, and with technology changing this fast, that's what you want."
Think of it like a car:
Interface: where you talk to it, like a chat window, Slack or your phone. That’s the steering wheel and dashboard.
Compute: the AI model doing the thinking, like Claude or GPT. That’s the engine.
Harness: the setup wrapped around the model that lets it work through a task step by step, remember things and check its own work. That’s the frame that holds everything together.
Tools: what it can reach out and do, like search the web, read your email or update a file. Those are the wheels.
Skills: saved know-how for specific jobs, like “here’s how we write a press release.” That’s the driver’s training.
My own setup is a good answer to Ben’s question. I talk to it from my laptop or phone (interface). Claude’s models do the “thinking” (compute). Claude Code runs it using a lot of standing instructions I’ve already set up (harness). It can reach my Gmail, calendar and a lot more (tools). And it has saved routines like my voice print, a morning briefing, and even a design spec for making decks (skills).
WHY THIS MATTERS
A rubric like Zapier's turns "be good at AI" into something you can measure, and that cuts both ways. It tells you where your team stands, and it tells you what you get back when everyone moves up a level. Ben put the payoff better than I could:
"When the team and its AI handle production, the partner spends less time on drafts and more time with clients, making the calls only they can make. If your best people aren't spending more time with clients a year from now, the AI isn't working."
Your next steps:
Rate yourself. Read Zapier's Marketing examples and be honest about where you land — and about your slope over the last six months.
Rate your team. If you lead people, the gap between your rating and theirs is your to-do list.
Put it in your next job posting. Spell out what Capable means for that role, and borrow Ben's interview question.
Then hit reply and tell me where you landed. I'm curious!
SPONSORED BY PRESSPAGE
📅 Your 2027 budget vs. your 2035 comms team
Comms runs on people, but AI is changing who’s on the team. Samuel Hall has spent 20+ years building comms teams at companies like Oracle and ServiceNow, and now builds them with AI. Tamara Dobeneck, from the generation most often told AI will replace them, has been researching how companies actually adopt AI in comms. In this free 60-minute webinar, Presspage brings them together to take on the question under every 2027 budget: one more hire or one more tool?
📆 Wednesday, Nov. 4 (9 a.m. ET / 3 p.m. CET)
Can't make it live? Register and they'll send you the recording.
THIS WEEK IN AI
🌎 AI Is an Accelerator. Most Comms Teams Are Missing the Brakes.
Up top, I called AI an accelerator and accountability the brakes. A new survey out of Canada suggests most communicators already know where the brake pedal is — it's their organizations that haven't built one in.
The Communications and Public Relations Society of Canada (CPRS) released its 2026 State of Communications and Public Relations report recently, conducted with the research firm Leger, and the title says a lot: "From Adoption to Accountability." The individual numbers are encouraging. 92% of comms pros use AI at work, and 91% of them check its answers even when checking eats the time they saved. Then come the organizational numbers. Only 37% say their organization has a clear process for reviewing AI content before it reaches an outside audience — and 37% have seen an AI error reach clients, the public or the media. And while 58% know who's accountable when AI content causes reputational damage, only 34% say their organization has actually defined it. (One caveat: this is 145 professionals, and the sample wasn't random, so treat it as a strong signal rather than gospel.)
In most comms shops right now, the brakes on AI are one careful person. That works right up until the day that person is rushed, out sick or wrong. CPRS national president Anne McInerney put it better than I could: "review, disclosure and accountability can't depend on one careful person." And the stakes are higher for us than for most — 78% of the professionals surveyed say they trust a leadership message less when they suspect AI wrote it.
Why it matters: AI lets you go faster than ever, which is the whole point. But speed without a way to stop is how an error ends up in front of a reporter instead of caught at your desk.
Your next step: Write down two names — who reviews AI-assisted content before it goes out the door, and who owns it if something slips through. If either answer is "it depends," that's your gap.
🎯 Quick Hits
Kathryn Reichert, a communications advisor, had a "YIKES" moment reading a story in The Rundown AI — and so did I. As Axios first reported, JPMorgan, Bank of America, Morgan Stanley and Citi found third-party AI agents operating inside customer accounts. The agents logged in with customers' own credentials, which bank rules prohibit, and some called support lines or took actions users never asked for. Morgan Stanley is blocking them outright, and the others are restricting access. Kathryn's take is the one to hold onto: any company using agentic AI needs an internal and external comms plan for the moment an agent oversteps, written now. A line like "We're restricting access while we build stronger oversight," she notes, "can land very differently coming from you than it does in public after the fact."
Speaking of clear impact: in PR Daily, Ragan's Mike Prokopeak reports that of the 200+ communicators who took Ragan's AI Readiness Assessment, only 22% can show a documented improvement or return on investment from AI, and 53% haven't defined any AI metrics at all. His advice for 2027 planning is simple: measure how long a workflow takes before you automate it, then report what you got back. "The communicators who can show exactly how AI expands their capacity will be the ones whose budgets grow with it." I agree — and it's the same "clear impact" bar Zapier now sets for every hire.
WORK WITH ME
🤝 Want This for Your Team?
I spent 15+ years running digital and AI-adoption strategy inside global brands — the kind of enterprise-scale channel work most "AI consultants" have never actually done from inside a company that size. I bring that same playbook to comms teams directly now: workflow audits, executive AI systems, the intranet-and-channel rebuilds I used to run in-house.
A handful of clients at a time. Reply and tell me what you're stuck on.
AI + COMMS JOBS
🏢 Find a New Gig
Looking for a role at the intersection of communications and AI? Here are some opportunities to check out:
Account Director, Tech & AI Comms at We. (United States)
AI Search Content Strategist & Writer (AEO/GEO) at ADP (Roseland, NJ)
Communications Manager, Claude Platform at Anthropic (New York)
Senior Manager, Physical AI Communications at CoreWeave (Livingston, NJ)
Director Communications, AI Infrastructure at Nscale (Seattle)
Was this forwarded to you by a friend or colleague? Want to get this in your inbox next week? Click here to subscribe!
YOUR FEEDBACK WANTED
🔊 Help The Comms Stack Improve
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