The newest model won't fix your business. I checked.
Everyone spent this week losing their minds over a new AI model. I actually used it. It's called Fable 5, Anthropic's latest. Every headline said the same thing, "the most powerful AI ever built."
My honest take? It's a little better. It caught some context the last model would have missed. A real improvement. But nowhere near what the headlines promised, at least not for the day-to-day work of running a business.
Then a few days later, the government pulled it offline over export concerns.
The most hyped model of the year, gone almost as fast as it showed up.
Takeaway: You can't run your business on whoever is winning the benchmark race this week.
The models change. They get better. Sometimes they vanish overnight.
The people who actually win with this stuff aren't the ones with the best model. They're the ones with the best implementation skills.
Knowing which job to hand AI. How to give it the right context. How to turn an answer into hours saved or money kept. That's the part that pays off, and it's the one thing nobody can take offline.
That's what we care about here. Not the newest tool every week, but getting good with the ones you already have. Let's get into it.
Learn how to actually use AI in your business.
I used to start every week staring at a wall of meetings, hoping it would work out. Now I let AI read my calendar and tell me where the week is about to break before it does. It's the first thing I'm walking through live in the new Applied AI Series.
Here's the version you can run today.
Most assistants now connect straight to Google Calendar. In ChatGPT it's Settings, then Connectors, then Google Calendar. Claude has it under Google Workspace connectors, and Gemini has it built in. (It only needs permission to read your calendar for this.)
Then paste this on Monday morning:
You can see my calendar this week. Act as my chief of staff. Where is my week overloaded, and where do I have open focus time? Flag any meeting without a clear purpose so I can cut it. Recommend two 90-minute focus blocks and tell me why you picked them. Name the one thing I should protect this week.
The result: A read on your week in thirty seconds, the meetings worth questioning, and where your real work will actually happen.
In the upcoming live session, I will walk through this connection, prompt, and then show you how to add in Email & Your Tasks so you can create the ultimate gameplan for your week. Join the next Applied AI session — it's 25 minutes, every other week, and you walk away with the exact prompt.
Attend the Next SessionHere are the best insights our team found from across the web this week.
There is more AI news than any operator has time for. So we read it so you do not have to.
Our take: We always benchmark the before/after effect of deploying our agents — the 'before' response times we have seen are pretty wild (in many cases, leads were not being responded to for multiple days.) That's for hot leads who are ready to buy — so you can imagine the review response times at the same businesses. When someone engages with your business and they don't hear back for days, they've already forgotten about you. AI or automated response — make sure it sounds human.
Our take: This is like the review responses, but for hiring. Lead or candidate, make sure you get back to people. Everyone hates getting left on read.
The agent answered: "Yes, we are located about 0.1 miles from TJ Maxx. If you have any other questions or need details about our classes, feel free to ask!"
Technically correct. Completely robotic. Nobody texts "0.1 miles," and nobody signs off a text with "feel free to ask."
This is exactly why you can't ship AI straight out of the box. The model isn't wrong, it just doesn't sound like you. A few lines of customization turned that same answer into something a real person would send: "Yep! We're right across from TJ Maxx, super easy to find. Want me to get you booked for a class?"
Same question. Same facts. One sounds like a call center, the other sounds like your front desk.
See How We Build It →Our take: This is the good kind of pressure. The same AI that writes the ad can now tell you whether the ad actually made money, which means fewer reports that look impressive and say nothing. If you get a report, drop it into your favorite AI and ask for a strong critique.
That's it for today. See you soon.
Brian, Murphy, John, Beth, and Jake — some of the humans behind AI Front Desk.