From the show
Another Orange Morning Recap: July 24, 2026
It was a Friday, and it was the end of a week that went almost entirely to AI. Beach plans on one side, a client go-live on the other, and in the middle of it a HubSpot update board that finally started filling up with the July 24 releases we had been waiting on all morning. Casey came in off a trip to Academy Labs in Houston, where the first half of the day was Breeze Assistant building ICP segments and the second half was AEO, which she admitted landed better in the room than it does in her head.
Here is what we got into: Agent Hub arriving and quietly rearranging where your Breeze context lives, the new describe-it-and-Breeze-builds-it flow inside Agent Builder and what happened when Casey pointed it at a real client workflow, and the redesigned AEO citation tracking, which turned into the argument that ate the back half of the show.
Agent Hub Is Here, and It Moved Your Furniture
The headline for the week is Agent Hub, now in public beta. If you select Breeze and think you are looking at Breeze Studio, look again. It is more than a rename. The place you used to reach your ICP context straight from the nav is now inside Agent Hub, on its own tab, and if you go looking for it in the old spot you will not find it. On the show we watched a CSM at Casey's table in Houston hunt for exactly that, walk to another table to ask, and land on the same answer: it lives in Agent Hub now.
Why it matters: Agent Hub gathers every agent into one place, organized by the outcome it drives, right as a lot of HubSpot's agents stop being free and start metering credits. The furniture moved, so budget a few minutes to relearn the room before you need something mid-task. This one is worth pulling up today.
For more details: read our full take on Agent Hub.
Describe It, and Breeze Builds It in Agent Builder
Inside that same wave is the piece Casey and Trisha dug into on Thursday and kept testing on their afternoon show: a box in Agent Builder where you describe what you want in plain language and Breeze Assistant configures a workflow or an agent for you, using your portal's context instead of a blank canvas. You say what you are trying to do, and it decides whether the answer is an agent or an automation and starts building.
Then Casey took it to a real client workflow that night, and the recap has to be honest about the result. It asked good scoping questions, and the thing it built was very not right, with multiple issues. The lesson was not that the feature is broken. It was that one-shotting a build is the wrong instinct. The people who get clean results from AI have already done their planning in a separate conversation first.
Why it matters: The on-air ask to HubSpot was a Plan with Breeze button, sitting right next to Create, so the first move a builder makes is thinking, not generating. Until that exists, treat the describe-it box as a fast way to skip the blank canvas, not a fast way to skip the plan. Describe the plan, not just the outcome, and expect to iterate.
For more details: read our full take on Breeze Assistant in Agent Builder.
AEO Citation Tracking, and the Argument Underneath It
HubSpot rebuilt the AEO Citations table. It is full-screen now, with two new columns, Content Type and Brand Mentioned, plus search, so you can finally see which sources are citing you and in what context. The old version was buried, thin on context, and impossible to filter. Casey opened hers live on the show and the difference was obvious. As a piece of tooling, this is a real upgrade, and if you are running AEO you should go look at it.
Then the show found its argument. The question Chris kept pressing was whether anyone actually opens an LLM and types a branded best-of prompt in the first place. To make the point, he moved his own portal's citation score from zero to 0.11 percent and its brand sentiment from zero to 80 percent in real time, just by pasting prompts into the tool that then measures the citations those prompts generate. That is the vanity-metric loop in one screen share. Madelyn made the fair counter, that AEO is not that far from SEO, you are targeting topics instead of keywords, and the discipline underneath is one we should have been practicing the whole time.
Why it matters: Track citations as a diagnostic if you want, but do not let the score become the goal. The work that actually earns a citation is the same work it always was, which is making useful content and being a good human. The risk Chris named is real: optimizing your day around recommendations for a world of prompts nobody runs is how you end up with no time left for the content that would have earned the citation anyway.
For more details: read our full take on AEO Citation Tracking.
The Rest of the Day's Updates
Plenty more shipped around the edges of the AI wave: a sandbox deploy wizard that now shows deployable and non-deployable assets with the reason why, a Payment Links API moving to the shared PaymentMethod enum, auto-association of activities extending to contracts, orders, and marketing events, workflow templates arriving in the marketplace, and a handful of sandbox and CRM index improvements. We did not pull each of those forward on air, but they are all logged with our read on the feed. The full set from July 24 lives on the HubSpot updates feed.
The Sign-Off
You probably already own the value you are looking for in HubSpot. You just need to wake up to it. Enjoy the weekend, and we will see you Monday morning.
Produced by Value-First Media