Define It Before You Configure It

Another Orange Morning Recap: Tuesday, October 27, 2026

Casey Hawkins is still on vacation, so Tuesday was Chris Carolan and Ryan Ginsberg of Unified Support Solutions, the Tuesday co-host, working through a board with four fresh rows and a fifth that showed up since yesterday. Chris opened by saying he needed Casey back to keep track of what day it was. The updates did the rest of the talking: nearly every one hands context to a person or to AI, and every conversation came back to the same step. Someone has to define what the thing means before a tool can act on it.

Here is what we got into: a new navigation on anotherorangemorning.com and a calendar full of agent and Office Hours sessions, Customer States and Follow-up Actions, Image Display for File Properties, Premium Engines in the AEO Tool, Handoff Summaries, the Reporting Empty State Redesign, and a Will It Breeze run through the service lens.

Before the Updates: A New Nav and a Busy Calendar

Chris rebuilt the navigation on anotherorangemorning.com overnight. The site now splits by hub and by host, and events sit up front, because he wants community-led and HubSpot-led events together in one place. The events calendar has plenty in it: the HubSpot admin user group session on Nurture Agent with the product managers who build it, Hubsessed Office Hours on loop marketing and AEO at 11 a.m. Central, and on Thursday a virtual Agent Hub workshop from the Houston HUG with Becca Stamp, Davis Mastin and Masuma Zakir.

Chris wants that workshop to start with context, because "Super easy to say, 'Hey Breeze, I want to make a new agent.' And then you get a new agent. And then it's like, okay, how do I explain this to anybody else? How do I explain it to myself?" It ends with a working agent and an understanding of what it costs to run.

Customer States and Follow-up Actions

Ryan picked this one first. Customer States and Follow-up Actions lets customer success managers track states such as onboarding, upcoming renewal, declining usage or cancellation indicated right on the customer record, then trigger the same follow-up every time a customer enters one. The states come in three groups: lifecycle, risk and opportunity. A follow-up can send an email that Breeze drafts, create a task, enroll the customer in a sequence, or notify a team in Slack or Microsoft Teams. It is in public beta for Service Hub Professional and Enterprise, and you can turn states on for every customer or only for specific segments.

Chris read it as an answer to a gap. "I'm gonna argue upfront before I even read anything. This is likely because there's nothing in lifecycle stages after the customer." Ryan came at it from the work behind it. Lifecycle stage never captured the whole story, and teams ended up building extra properties and workflows to fill the gap. He asked where the usage signal comes from, and his answer was the one the segment kept returning to: "what does it mean when they're in onboarding? And what does it mean when we're declining usage? So all of those things need to be defined first."

Ryan also wondered whether each state could have its own custom agent trained on your processes. Chris agreed the custom event layer makes it easy to feed these states into HubSpot, but "you have to design it. You have to architect it and then build it, and then make sure it's working." A one-person success team should pick one segment and start there. What a state is underneath, an object, a record or a set of properties, stayed an open question for a closer look inside the beta.

Image Display for File Properties

File properties can now show image thumbnails on a CRM card, a record and index pages, and the images work as personalization tokens in marketing emails and CMS content. Image Display for File Properties is in public beta for all hubs and tiers. It is opt-in per property: file properties created during the beta have it on, and existing ones get a "Display uploaded images" checkbox in the Rules section.

Both hosts had a project in mind. Chris asked Ryan whether he remembered how they had to build all of this, and Ryan went straight to a project they were on: a custom UI extension so a client could upload an image, edit its name and toggle through file properties in a carousel. Ryan has a client whose invoice PDFs land on company records from another system, where a quick action could not put images in an email body; now a property token can.

The part that mattered most to Chris is about AI. "I am always thinking about AI being able to see what's on the page and literally see it. … those kind of context switches are points of risk to it happening at all." If Breeze has to open a file in a new tab or download it, the image is not on the page with everything else.

Premium Engines in the AEO Tool

Premium Engines in the AEO Tool is now in public beta. The tool already tracked prompts through ChatGPT, Perplexity and Gemini. The new piece adds Claude, Google AI Mode and Google AI Overviews, and results from all six count toward your brand visibility score. The settings screen notes that tracking across the new engines needs added capacity.

Chris took this one on his own, and used it to explain something he had noticed. ChatGPT, Gemini and Perplexity carry search in their DNA, which is why they came first. Claude was one of the last to join the search game and still does not default to it unless you ask. "[Claude] is not naturally a search mechanism," he said, so approach it like one and it will likely not have the results for you. The caution is on the Google side: "you're not in charge of Google AI mode and Google AI overviews at all." Google decides on its own when to show an overview, so put your energy into what you control instead of writing ten list articles to land in one.

Handoff Summaries: How Many Before You Trust One

When Customer Agent hands a ticket to a person, Handoff Summaries posts a note on the ticket with a headline on the customer's core issue, the reason for the handoff, a recap of what the customer asked and what Customer Agent already tried, the unresolved issues, and any links or files the customer shared. Like other notes, they can be edited or deleted. It is in public beta for Service Hub Professional and Enterprise, and Chris reminded everyone that Customer Agent itself does not require Service Hub.

Ryan read the why and went straight to the phone tree: how many times have you been passed between three people and had to explain yourself again, and are we done with that? Chris answered with a harder question. How many of these summaries does a rep need to see before they stop scrolling back through the whole ticket to check? The trouble is the wording. "Sharp, brief, concise, clear, those are subjective words, folks. And the only way you're gonna learn to trust these is by actually seeing it and doing it."

Ryan's counter is the benchmark. Human reps typed notes in the old system and people trusted those, so the question is what the AI summary has to beat. In an hour-long conversation, he said, "the AI will be able to better summarize that than any human could," and a team under pressure to cut resolution time is not going to read the hour. Chris landed on the consequence: "it's gonna force folks to start trusting these summaries probably before they're ready." The unified view on the ticket is the safety net, since you can check the record as you pick it up.

The Reporting Empty State Redesign

One more showed up since yesterday. Reporting Empty State Redesign is live for all hubs and tiers. Where a report card used to show a plain "no data" message, it now shows a placeholder shaped like the report that is meant to appear, a table or a pie chart. The message underneath is the same, and HubSpot says empty states keep functioning as normal.

Chris called it a small, small update, then found the card that made it interesting. On "Open deals not contacted recently," the new copy says to keep the list empty by following up. That is a report where empty is the goal, a case of "here's the queue, you actually want it to be empty." Whether empty states can be configured in your own reports stayed open; the hosts did not know and said so. Wednesday's Office Hours on reporting and dashboards, at 11 a.m. Central, is the place to ask.

Will It Breeze: Rank What Deserves Attention First

Will It Breeze ran through the service lens, which on the Will It Breeze page means the questions a service lead would ask. The wheel came up Conversations, Tickets and rank what deserves attention first, and the prompt read "As our service lead, using Tickets, can you rank what deserves attention first across Conversations?" Before Breeze answered, Chris noted that the pipeline on his ticket screen was not the one he expected. In his words, "this is where AI amplifies whatever is already happening."

Breeze ranked by impact instead of age. Handle a cancellation first, then the access defect rather than the already-resolved password request, then recover the conversations where the AI failed to help. It named its own limits: the ranking covers the current support ticket view and the linked communications it retrieved, not every inbox, and the open tickets had no priority or SLA values filled in. Chris had been about to explain that scope himself, and Breeze said it first.

His message for anyone holding back: "start talking to [Breeze], folks. … You don't need to know everything about everything or how AI works or how great your HubSpot data is or who's doing what when to get value out of starting conversations with [Breeze]." Breeze wins again. Watch the episode for the full run.

The Rest of Tuesday's Updates

Two rows from Monday's board were still on screen. Chris gave a nod to yesterday's Microsoft Dynamics 365 Custom Lookup Sync, which we covered Monday, and Help Desk Channel Send-From Permissions went unopened. Everything else that shipped is in the HubSpot updates feed.

The Sign-Off

Thanks to Ryan for a Tuesday that kept coming back to definitions. Have a great rest of your Tuesday, and we will see you tomorrow.