This issue is built differently than usual. Instead of walking you through something I built this week, I pulled the sharpest findings out of the research I already run every morning for the podcast and turned it into three things: a shift worth knowing about, a framework you can use this week, and one thing from the show worth the extra five minutes. If this format works better for you than the old build log, tell me. That's the whole point of testing it.

Two numbers landed this week that change how you should think about being found. Seer Interactive tracked 25 million impressions and found organic click-through on AI Overview queries dropped from 1.76% to 0.61% between June 2024 and September 2025, a 65% drop. Separately, the overlap between top-10 Google rankings and AI Overview citations fell from 75% to 17-38% by early 2026. Ranking #1 no longer means the AI answer mentions you.

Here's the mechanism: AI search doesn't hold a fixed rank for your business, it calculates a rolling probability based on whether the wider web agrees with what your own site says about you. If your site claims one thing and your reviews or forum mentions say another, the AI's confidence in citing you drops, even if your SEO is fine.

What to check this week: type your own conversational questions ("who handles X for a company like mine") into ChatGPT and Perplexity, not just Google. If you don't show up, that's a real gap now, not a hypothetical one. The traffic that does arrive through an AI answer converts at unusually high rates, so this is worth the twenty minutes.

Most people get worse results from Claude or ChatGPT than they should because they're prompting instead of contracting. A prompt contract is five sentences, written before you hit enter on anything that matters.

Pair this with a claude.md file in whatever project you're running: three to five rules the agent reads at the start of every session, with instructions to append anything it got wrong to the bottom after each run. That's the difference between re-explaining your business every time and an assistant that actually gets sharper.

Meta shipped its first real coding agent this week, Muse Code, and undercut Claude Code and OpenAI's Codex on price by up to 250 times. The catch: the cheap tier is opt-out, not opt-in, so Meta trains on your code by default unless you switch it off. If you're evaluating it for anything client-facing or under NDA, that's not a footnote, it's the whole decision. Full breakdown, pricing math, and benchmarks in this week's episode.

🎧 Full episode on YouTube

I'm a dad of three teenagers, and on top of everything else I do with AI, I've been trying to keep up with what all the apps my kids use can actually do now. It's a moving target. School's starting back up in a lot of states this week, which means phones are back in the mix too.

I built an AI screen time tracker to make sense of it instead of guessing. Check it out here, and if you're a parent dealing with the same thing, I'd genuinely like to know if it's useful to you.

A month or two back I ran an AEO pass on Listaza's site, wrote the content so AI tools would actually cite it, not just rank it for Google. When I dug into the analytics, it wasn't a fluke. Claude was referencing the site directly and serving it to people searching for something similar.

I did the same thing on the AstraNex Ventures site, and I've seen more hits show up in Google Analytics since making those changes.

One thing I want to be clear on: traditional SEO didn't go away. It's not AEO instead of SEO, it's both working in tandem now. If your site doesn't rank in the first place, the AI has nothing to pull from when it's deciding who to cite.

n8n — The automation tool I use to connect apps, trigger workflows, and stop doing things manually. If there's a repetitive process in your business, this is where you start fixing it.

VoiceInk — A local AI dictation tool for Mac that transcribes your voice with near-perfect accuracy and runs entirely on your device, meaning nothing you say ever touches a cloud server.

Blotato — Handles the full content distribution side of your business: drop in a topic and it generates platform-specific posts, or feed it existing content and it repurposes it across formats. TikTok videos become tweets, podcasts become blog posts. Includes a scheduling calendar, visual creation tools for carousels and infographics, and publishes natively to 9 platforms with no per-post fees.

Beehiiv — What you're reading right now is published on Beehiiv. If you're thinking about starting a newsletter or moving off a clunky platform, this is the one I'd recommend. 20% off your first 3 months with my link.

Google Workspace — Beyond email and Docs, a Business Standard plan includes Gemini Pro built into every app, NotebookLM Plus, and access to the enterprise versions of the whole suite. Better value than a standalone Gemini subscription when you're already paying for Google anyway. 14-day trial and 10% off your first year.

Descript — Video and podcast editing that works like a text document. You edit the transcript and the media follows. Cuts filler words, cleans up audio, and handles captions automatically. 50% off your first two months on the Creator Plan.

The thread running through all three stories this week: AI stopped being something you check in on and became something you have to actively manage, whether that's how your business shows up in an AI answer, how you talk to Claude day to day, or what's actually happening on your kid's phone. None of that gets easier by ignoring it. Pick one thing from this issue, the audit, the prompt contract, or the tracker, and actually try it before next week's issue.

More free resources here.

Cheers,
Joe

PS: If this format landed better for you than the usual build log, reply and tell me. I'm testing it.