I Didn't Think I Had a Backup. AI Saved the Day.

Happy Wednesday! This week:

Let’s get started!

This isn't a comms story -- at least, not immediately one. It's a personal one and a strong indicator of how AI can fundamentally change how we work for the better.

You'll need a bit of quick context if you only know me from this newsletter: I also write another newsletter, called Now I Know. It's a daily "huh, I didn't know that" email that's been running for sixteen years and reaches about 50,000 people every morning. It has nothing to do with AI or comms. It's just stories -- and I've written thousands of them, going back to June 2010.

Three weeks ago, someone broke into the web host running it, got in through an old admin account I'd forgotten even existed. I had no idea -- until early last week, when I woke up and noticed that someone had deleted everything. Not "some posts are missing." Everything -- sixteen years of articles, on a server with no backup service of any kind.

I used the word catastrophic more than once that week, and I meant it every time. If you know me personally, you know I rarely if ever panic.

Last week? I panicked.

All the content: gone. The server wasn't backed up. Sixteen years of work -- scattered into oblivion. Sure, I had copies of just about everything I wrote in my email inbox, but that would take forever to piece back together, and it all came with broken links and old designs. It was an absolute mess and I didn't think I'd be able to cover.

Claude saved the day. That, and my AI literacy.

I'd set up an AI chief of staff about three months earlier -- mostly to run my calendar, my inbox, my job search, my consulting work, and to keep me on track on other projects (like The Comms Stack). It already had access to a lot of things most people never bother connecting. And because I'd spent that time getting comfortable with how AI actually works -- not just prompting it, but understanding what it's good at and where it needs a push -- I could sit down in the middle of a disaster and work out an actual plan with Claude. It helped me write emails to my hosting company tech support to find databases that were left untouched by the hacker. It scanned through probably thousands of emails to see if there were clues about other backups -- especially ones that could have the content in a structured way, which could help with a rebuild. It was able to dig through a similar number of links on Internet Archive to find a few personal photos I had also saved on that server which I thought were gone forever. With Claude's help, I was able to recover almost 100% of the written content and the majority of the images.

But it was a huge lift. The recovered database held 3,409 published stories going back to 2010. Nearly 3,000 of them needed to be rebuilt from scratch outside the archive they used to live in. Almost 11,000 short links, scattered across hundreds of those stories, needed to be individually checked and, where the destination was dead, pointed at an Internet Archive copy instead. Images had to be matched back from four different graveyards -- an old backup, an email marketing export, my own inbox, and the Internet Archive -- and in one batch of test posts, only three images out of nineteen came back empty. I didn't count any of this by hand. I couldn't have. The tally itself only exists because the tool doing the rebuilding was also the one keeping score.

And recovery was just step one, restoration was step two -- but I still need to republish everything. That's a work in progress, but it's moving forward. I am working with beehiiv, my newsletter host, on back-populating the archives on their servers. Again, my AI chief of staff was able to help me turn some of the more technical concepts into things I understood, drafting replies back to the fantastic beehiiv engineer that jumped in to help. I connected Claude to beehiiv and it was able to analyze my recovered archives, find about a dozen different ways that I had approached that newsletter over the 16 years of writing, select a test version from the archives, and upload a draft right to beehiiv for my review.

About ten days after my catastrophe, I'm well on my way to recovery. And -- as I've said to whomever would listen over the last ten days -- this is something I simply couldn't have done three months ago, before I started building this AI chief of staff. The amount of work is massive and my technical knowledge, while strong for a comms pro, wasn't quite enough to get me past a few hurdles. And frankly, having a dispassionate robot who knows how I work was the perfect thought partner for the moment -- an expert who can solve problems without the overwhelming feeling of dread that you can only imagine I was experiencing.

That's what I learned about the role of AI and a crisis: the value wasn't the labor alone. It was having a thinking partner and a workforce in the same conversation.

I wanted to share this story today because the saving grace was the work I had already put in. There was no magic prompt that got Claude to help me here. The thing that actually saved sixteen years of work was the effort I had put in previously. I'd already built the fluency and the system, for reasons that had nothing to do with this, long before I needed either one. And if there's one key takeaway I want you have from my experience, it's simple: get building now.

Because really, the story above isn't just about my other newsletter. A kind of loss that used to just be permanent -- a one-person operation's entire archive, gone, no IT department, no enterprise backup contract -- doesn't have to be permanent anymore. And there are countless examples of other crises, not limited to the one-man-band situation I'm in, that this applies to. Think about what your "oh [explicative]" work moment would be right now -- and then ask yourself, could your AI habits and systems help you through it? I don't think most people, including most comms teams, have asked themselves whether they have that same thing in place before they need it. But you should.

Your next step: don't wait for your own version of this. Open a conversation with whatever AI you actually use and ask it plainly -- if this disappeared tomorrow, could you recover me? What's actually gone for good? You'll either get a plan worth having, or you'll find out exactly what you're exposed to while you still have time to fix it. Don't wait for your own catastrophe to hit -- trust me, it's not fun.

Shift your workforce into hyper-productivity mode with Dell Pro laptops powered by Intel® Core™ Ultra with Intel vPro®. Shop on Dell Premier.

Explore More

Axios HQ -- the internal communications platform built by Axios -- surveyed more than 1,200 U.S. executives and employees for its 2026 State of Internal Communications report, and the headline finding is one most of us have felt without having a number for it: in the AI era, comms volume is up, and most executives agree clarity is down. The gap shows up everywhere else in the data too. Only 1 in 6 employees feel fully aligned with their organization's goals, and that misalignment drove nearly double the missed deadlines in 2026 that it did the year before. Seventy-nine percent of employees say the quality of communication they get from leaders directly affects how well they understand where the organization is headed.

The AI-specific numbers aren't great, either. Just 26% of staff say their organization has a clear, enforced AI-use policy -- and yet 61% of leaders believe their employees are already regular or power users of the tools. In many cases, most of the organization is improvising with AI while leadership assumes someone, somewhere, wrote the rules down. And maybe they did, maybe they didn't, but those rules aren't being followed regardless.

More AI-generated content is not the same thing as more clarity, and leadership (or the comms team) that mistakes the first for the second is about to find out the hard way. AI makes it surprisingly easy to produce another update, another summary, another all-hands recap. None of that fixes the thing that actually drives alignment, which is whether the message was clear enough and engaging enough for someone to read it, internalize it, and act on it -- and without guessing.

Why it matters: The 26%-vs-61% gap means most organizations right now have employees experimenting with AI on their own terms, with no shared sense of what's allowed, while leadership assumes the opposite is true.

Your next step: Before you greenlight one more AI-assisted update this month, ask whether your org actually has a written AI-use policy employees have seen -- not one you assume exists. If the answer's no, that's the gap to close before the next volume-boosting tool goes live. And one follow-up step -- run some surveys to see how those AI-assisted updates are actually landing. There's a good chance they aren't.

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.

You’ve read the AI breakdowns. Watched the million AI explainers. Nodded along to the hot takes.

And then opened a blank doc and had no idea what to actually do.

The Shift is a newsletter built to help you in moments like this. Every tool covered works in the real world, and every prompt in the 1000+ library solves something you’ll actually hit. 

3,000+ vetted tools. 1000+ tested prompts. Daily newsletter.

And right now, 3 subscribers win a free 1-year Claude Pro subscription. One click to enter.

Subscribe to The Shift

*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