Happy Wednesday! This week:

  • Learn to build in Claude Code — join tomorrow’s webinar

  • My path from being a curious-about-AI user to where I am today — and how you can follow the same path

  • Will comms own the AI workflows? I’m skeptical, but a recent essay I read has me thinking.

Let’s get started!

THE LEDE
💡 I Started This Newsletter in January as a User. I’m Ending the Summer as a Builder.

I’ve been finding AI very valuable, I want you to get the same value out of it, and building my own tools and workflows is how I got there. And I want to get you there, too. The webinar — details below — is a great way to start, as is my new course site, The Build.

I’m only one data point, but I think my journey has been instructive. What follows is the path I actually walked between then and now — not because it was special, but because it wasn't. Every rung was available to any comms person reading this. Here's the order it happened in, and what you can do to climb the same one.

1. First, I had to believe it was worth my time.
Last September, I was very reluctant to use AI tools — at work, I had access to Copilot, which wasn’t doing much for me, and at home, I was using ChatGPT as a magical search box. It wasn’t helping me with my work. My colleagues created a set of prompts for all sorts of purposes and I gave a few of them a try — and that helped a lot, because I was seeing outputs that made sense and were faster (but not better) than what I could do on my own. For low-value work, it helped — and that was the whole unlock. Nothing after it would have happened without that first "oh, this is different" moment on something I actually cared about. Here’s a path you can follow:

  • Download my prompt playbook and run three prompts from it this week on live work — a draft, a media list, a recap — never a throwaway test doc.

  • Point the very first one at the task you dread most. This lands hardest where the pain is worst.

  • When a prompt works, save your tweaked version somewhere so it starts becoming yours, not mine.

2. Then I got honest about why I wanted it.
I learned a lot more about how to prompt better and I began to realize that AI use is a skill and one I needed to learn. (The good news is, so did everyone else!) I really wanted it to do everything really well from the get-go because that’s what was promised to us, but it’s not magic — it’s just another tool that gets more useful the better you become with it.

But to start building that skill, I needed to want this thing to work for me. Personally, I like being innovative, so I started to try things that I hadn’t seen many others try, and finally built a custom GPT. I showed it to everyone who would listen. While I no longer use or share that GPT, it sparked the joy I needed to keep learning. It's the same thing I now tell every leader who asks how to move their team: figure out your why before your what.

Most people aren’t me, though — you’ll need to see a more tangible result if you want to really build that desire to do more. My suggestion:

  • Write down the one thing you'd reclaim if the grunt work did itself — an evening, a Friday, the 4pm scramble.

  • Track your real hours on one repeating deliverable for a week, so you know exactly what's on the table.

  • Name the version of your role you're aiming at, and keep it somewhere you'll see it when the reps get boring.

3. I stopped prompting and started briefing.
The real skill turned out to be one comms people already have. I quit hunting for clever prompts and started treating AI like a sharp junior hire — context, the goal, the constraints, an example of my voice. The day I built a reusable voice profile and led with it every time was the day the output stopped sounding like a robot and started sounding like me. Honestly, it’s still not great and I find myself discarding that copy more often than I use it. (A quick aside: “Honestly” is a dead giveaway of AI-drafted copy, but — honestly 😀 — that was 100% human written.) But the larger prompts that brought context to the effort were a big step up. I wasn’t just tossing things at an LLM and hoping it would give me something useful, I was — effectively — mentoring it on a path toward success.

To follow the trend of this newsletter, here are three things you can do to experience the same:

  • Treat every ask like briefing a freelancer: context, the goal, the constraints, what "good" looks like, an example of your voice.

  • Build one reusable voice profile — paste in three things you've written — and lead with it every time instead of starting cold.

  • Iterate, iterate, iterate. When an output misses, fix the brief, not the output. Rerun it and watch what one added line of context changes.

4. I learned to diagnose, not just do.
I’ll admit — I’ve yelled at Claude and ChatGPT a few times. Not verbally, that’d be super weird, but the “WHY CAN’T YOU DO THIS??!?” all-caps thing? Yep, I’m guilty as charged.

After a few of those, though, I treated that rhetorical question as an honest one — why can’t the LLMs figure out what I’m after? I had no idea. But something clicked — I realized I could just ask the chatbot what’s going on. And I then asked myself the same — where am I taking shortcuts, asking the chatbots to infer info I haven’t provided, for example. The important part is the shift — asking the chatbots and yourself to diagnose, not just do. This is where I think you level up your AI use — it’s like the mushroom or fire flower for Mario, but real.

No three-part path for you in this one — if you’re at the step, it’s straightforward: pressure-test what you’re doing and ask the LLMs to do the same about what they’re doing.

5. And then I started building my own tools.
This is the part I didn't see coming in January. Using a plain-English coding tool — Claude Code — I stopped borrowing other people's software and started building my own: a media-intelligence system that does journalist research the way I'd do it, a scanner that audits how AI models "see" a brand, a chief-of-staff setup that drafts my own recurring work on a schedule while I sleep. No developer. Just me describing what I needed in English.

Most of what I’ve created isn’t “software” in the typical sense. It’s just a lot of complicated prompts that Claude Code has access to and reads without me copy/pasting. I didn’t write them — Claude did for me, as part of my conversations with it. I’ve never even looked at them — there’s no need. I just worked with the AI chatbot to create something that I needed, it created it, and now it uses that creation to help me do my work.

This is the hardest part to get started on, which is what tomorrow’s webinar and The Build are there for — I want to help you take those first few steps. I hope you’ll join me for at least one, if not both.

But either way, that's the whole arc: believe it matters, get honest about why, learn to brief, learn to diagnose, then build. I climbed it in seven months, in public, one issue at a time — and I'm now convinced any communicator reading this can climb it faster than I did, because I've watched enough of you do exactly that.

ABOUT THAT WEBINAR
💻 Let’s Build a Comms Tool in Claude Code

Nearly fifty of you have now sat through this one with me: an hour, seven prompts, and a real working tool at the end of it. No coding experience required.

What you'll build is The Comms Gut Check, a web app where you paste in a draft and get structured feedback on clarity, jargon, tone, and where your audience is likely to push back. (I ran the Declaration of Independence through mine. The loyalists did not care for Jefferson's prose.) But the tool isn't really the point — getting comfortable enough with Claude Code to build the next one on your own is.

The next session is tomorrow, August 6 at 11 a.m. ET. Same build as last time, so if you couldn't make the earlier ones, this is your chance. Save your seat.

💡 Can't make it? Click here to build The Comms Gut Check on your own — same course materials, but on your own time, and also free.

👍 Planning on attending? Here’s a bit of setup before we start, so we can get through the whole build in the hour:

  • Claude Code set up inside the app — it needs a free tool called Git, and Claude itself will walk you through that in a few minutes if you ask

  • A free account at platform.claude.com with a few dollars added ($5 is plenty). You'll need an API key, but we'll set that up together during the webinar.

If you'd rather see the full walkthrough of what we build and why, here's the issue where I broke it all down.

And if you want me to run a guided build for your team, get in touch.

SPONSORED BY

2 Free AI Courses. No Credit Card Needed.

5,000+ professionals use Skill Leap to get ahead with AI. Right now, two of their best courses are completely free - Claude 101 and the 14-Day AI Boot Camp.

Claude 101 covers prompting frameworks, Artifacts, file analysis, and real-world workflows in 19 lessons.

The Boot Camp covers ChatGPT, Gemini, Midjourney, and prompt engineering in 16 lessons. Downloadable workbooks. LinkedIn certificate.

Zero cost, no credit card, no catch.

THIS WEEK IN AI
🌎 Will We Have a Chief Workflow Officer in the Future?

I came across this essay by Bob Pearson, "The Chief Workflow Officer," and it put a title on something I've been circling for months. Pearson's argument is simple and, for us, a little thrilling: major technology shifts change how work happens before they ever change the org chart, and the person best positioned to redesign that work in the AI era isn't in IT — it's the Chief Communications Officer. His logic is that comms leaders already live in the places AI strains hardest: how information flows, where it breaks down, how stakeholders react, and how quickly trust erodes when the machine gets it wrong. So instead of waiting for a technical function to hand us a rollout plan, he says we should be the ones architecting the new workflows — building the governance, the training, the human-escalation rules, and the institutional memory that keep an AI-accelerated organization from flying apart. Rename the role if you like; the point is that redesigning how work moves is becoming a communications job.

What makes this land for comms pros specifically is that it reframes AI from a productivity toy into a mandate we're uniquely qualified to own. If content creation is collapsing from days to minutes and research is nearly instant, the scarce, premium skills become the ones AI can't fake — judgment, strategy, creativity, and the ability to lead people through change. Those have always been our skills; we just rarely got to apply them to the operating system of the company itself. That's the opportunity hiding in Pearson's title. The communicators who thrive in the next few years won't be the ones with the cleverest prompts — they'll be the ones who stopped thinking of AI as something to use and started treating it as something to build workflows around, on behalf of the whole organization. Whether or not "Chief Workflow Officer" ever shows up on a business card, the work is already ours to claim.

🎯 Quick Hits

  • Get ready for what marketer Emily Kramer calls “multiplayer AI” — that moment when your whole team is working on a shared AI-powered workflow. Read the whole newsletter if you have time, because the progression is real. Her 12-step level ladder has been very similar to my experience — in fact, I built a team repo (rung #10) and now have it updating continuously (#11). We’re all going to get there, and Emily’s ahead of the curve.

  • Julie Averill, former Chief Information Officer at Lululemon, wrote an essay for the New York Times arguing that when it comes to the value provided by AI, employees expect proof, not promises. (And to be clear, she thinks employers are failing there, noting that companies are overhyping AI — and that gap between rhetoric and reality is causing employees to distrust both leadership and the technology.) For communications leaders, that means resisting the urge to overstate AI progress and instead communicating clearly about what's actually changing, what's still being tested, and how AI is improving work in measurable ways.

  • Sharing the title only here because I know it’s a must-click: “A PR Firm Is Using Fake Publicists With AI-Generated Headshots to Spam Journalists With Pitches for Its Clients.” Yikes, don’t do that!

I’M HERE TO HELP
🤝 Your Comms Team — Leveled Up

Whether you're trying to build systematic AI adoption across your team or focused on AI search visibility for your brand or organization, I’m here to help.

I have three slots open for late August/Septemberlearn more and book an intro call.

COOL AI TOOLS
🔨 Tools To Try This Week

These links are sponsored; trying these tools will help support this newsletter. Thanks!

  • Granola — I swear by this AI notetaker. I’ve been using it since January and just upgraded to the monthly plan this week, and now have it integrated into Claude Code. I can’t recommend it highly enough — it’s that good.

  • Wispr Flow — AI-powered dictation tool that a bunch of people swear by. I’m on the free plan and it’s really great, capturing nuance and tone with rare misses. I use it on my phone a ton, running into usage caps quickly. I’ve used it on my laptop a few times, too, but I like to type, so we’ll see how it goes.

  • RankedAI — A platform to help your work appear on AI search results. I was really impressed by how they have you ask ChatGPT to tell you what they do — it’s a “proof is in the pudding” idea.

  • Viktor — They position this as “an AI employee that lives in Slack and Teams,” and as someone working autonomously for now, I haven’t had the chance to try it. But it looks pretty neat.

AI + COMMS JOBS
🏢 Find a New Gig

#CommsJobs started with a problem nobody was solving.

Daniel Méndez Aróstica had just moved countries and was rebuilding his network from scratch. His feed began filling up with comms roles, so he started sharing them with his network. That network is now 25k strong.

Every Friday, the #CommsJobs newsletter drops the week's comms roles straight to your inbox. Hand-curated. No spam, no filler, and never a charge to the communicator looking for work. 

Your next role is hopefully in this week's list.

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