From the show
Twelve Columns, and the Update Joshua Oakes Waited All Year For
The last day of August, and Joshua Oakes marked the start of fall the way he marks it every year: by walking face first into a four foot spider web on the morning dog walk. Ninety degrees in Nashville, ninety degrees in Casey's corner of Maryland, and neither of them willing to argue with the spiders about what season it is.
Here is what we got into: the twelve column dashboard grid that turned out to be Joshua's favorite update of the year, conditional property logic finally learning to hold two thoughts at once, the new Email Creation Agent, a help desk feature that landed better than anyone expected, Google Ads recommendations showing up inside HubSpot, and a Will It Breeze run that went sideways in the most useful way.
Two Segments That Set the Tone
Dashboard Grid Layout Update
Chris made Joshua guess before revealing it, and Joshua guessed wrong twice before landing on the answer he had been waiting all year for. Dashboards now use a twelve column grid instead of six. On paper that is a layout tweak. In practice it is the difference between a dashboard you can read and one you scroll past.
Joshua's case had nothing to do with fitting twelve things in a row. It is about sizing a tile to its job. Under the six column grid, a single scorecard number had to take up a third of the dashboard, while a trailing twelve week bar chart got compressed until the data was unreadable. His verdict on the old grid was chicken and waffles: not one thing or the other enough to actually work.
What he wants is a dashboard you can read in the five minutes before a weekly meeting starts. Three to ten pulse metrics you look at often enough to build an intuition about what good and bad look like, a trailing bar graph for the quarter, maybe a pie chart. All above the fold on a laptop. Twelve columns gets him closer to that.
Why it matters: The flexibility does not fix the problem underneath it, and Joshua was the first to say so. A dashboard should have one job. Not "give me a 360 degree view of my business," which produces something so overwhelming nobody knows where to look. More control over layout makes a glanceable dashboard easier to build, and it makes a monstrosity easier to build too. Just because you can does not mean you should.
For more details: the full take on the dashboard grid layout update.
Conditional Property Logic with Combination Operators
Conditional property logic could only ever look at one property at a time. Now it can combine conditions, so a dependent property appears when multiple criteria are met rather than just one.
Both use cases that came back from the room turned out to be the same use case: pipeline conditionals. Show or require a set of properties when the deal is in this stage and the product is that, then cascade the prompt as a rep advances to the next stage. That is where conditionals actually get used day to day, and it is exactly where the one property limit hurt.
Why it matters: There was a fair objection to HubSpot's own example, which requires a reason for churn field only on high priority closed lost deals. You would want that on all of them. And the deeper data problem this does not touch is the one reason problem, where every closed lost comes back "price" or "features," and the very same reason shows up on the wins.
For more details: the full take on conditional property logic with combination operators.
Casey Takes the Email Creation Agent
The Email Creation Agent gives marketers a complete on brand email draft through conversation instead of a form. You describe the goal, pick a starting point, refine in chat, and see a live preview before the email object is ever created. Brand voice comes from the HubSpot brand kit automatically.
Casey's read is that this is not a new behavior. It is a relocation. This is already what most email marketers are doing, just not inside HubSpot. Moving the motion to where the email actually gets built means the draft lands as a real asset, and possibly the audience with it, instead of getting pasted in from wherever it was written. Keep your tools where you work is one of her AI adoption guidelines, and this is that guideline showing up in a product.
Why it matters: Two caveats worth carrying. It is Marketing Hub Professional and Enterprise, and it was not appearing in any of the qualifying portals we have access to while we were on air, so treat it as a phased rollout rather than assuming you have configured something wrong. And it has "agent" in the name, which almost certainly means credits at some point. Even then, probably better value than spending tokens outside HubSpot to do the same job with less context about your business.
For more details: the full take on the Email Creation Agent.
The Help Desk Feature Nobody Expected to Like This Much
While working a ticket in help desk, you will automatically see resolved tickets similar to the one in front of you, so you can reference how past issues were handled without leaving the ticket.
This landed better than expected, and for a specific reason: the room had assumed it worked the other way around. The guess was that closing a ticket would surface other open tickets on a similar theme. What HubSpot built is more useful to the person actually in the seat, though the open ticket version is still wanted. There is a real triage moment where three people report the same broken process and you need to bring them together without flattening the fact that it came from three separate people.
Why it matters: This is in development, so the usual forward looking caveat applies. The bigger point is that this whole class of feature is only ever as good as the data already in your system. AI can excel at trend analysis. It cannot excel over a help desk nobody has been keeping clean.
For more details: the full take on automatic similar ticket suggestions.
Google Ads Recommendations, and the Real Reason to Care
Google's optimization recommendations for your ad campaigns now render inside HubSpot, with a count per campaign, estimated impact, and apply or dismiss in place. Mechanically that is a small change. It took the longest stretch of the show anyway, because the interesting part is not the recommendations.
Nobody should be one click accepting these. They are tuned to make Google money, not necessarily to make you money, and "increase your daily budget by a thousand dollars" is a suggestion, not advice. The reason this matters is that once the data is inside HubSpot, it is in front of Breeze alongside CRM data. Ask what the number one reason is to want something in HubSpot and the answer is customer data. Ask what the number two reason is and the answer is still customer data, this time for Breeze.
The teams this actually helps are small ones. A marketing department of four with no agency running the ad buys, doing lifecycle and reactivation and recapture work where CRM context is the whole game. Agencies and large ad accounts are not working in HubSpot for this and statistically are not about to start. It also helps the newly arrived marketing leader who should not have to chase down separate Google Ads access just to get a read on how things are going.
Why it matters: The question the segment kept circling is the one worth sitting with. Where are people actually working a year from now? A meaningful share of the market is going to expect to ask an AI a question like that and want very badly to trust the answer, whether or not they should. Better that they ask Breeze, which is plugged into the data, than that they ask an admin who spends hours building a report nobody acts on.
For more details: the full take on Google Ads recommendations.
Will It Breeze: What Lands This Month
The wheel landed on tickets and data enrichment, and the question became "using data enrichment, can you tell me what lands this month across tickets?" The word "lands" was chosen on purpose. It is not HubSpot vernacular, and it is exactly the kind of word a stakeholder uses when they have not memorized the platform's vocabulary.
Breeze worked visibly. It hunted for a ticket property matching "lands," tried several mappings, then correctly inferred from the page context that the question was about product updates landing this month for tickets. Which was a fair read, and also not what was meant.
That turned into the useful part of the segment. Breeze weights the page you are standing on very heavily, which is a strength when it saves you a sentence of setup and a problem when you are seven turns deep in a workflow investigation and it starts talking about the Sam Smith record. The initial frame of an AI conversation is what everything after it inherits. When it misreads you, the move is to correct the frame rather than rephrase the question, the same way you would with a collaborative partner who is human.
Chris also showed Breeze answering why the dashboard update is Joshua's favorite, grounded in call recordings and summaries already in the portal, with citations. It took about a minute. Joshua conceded it articulated his own preference better than he would have, then noted he could take next Monday off. He was talked out of it.
Also Worth Knowing
The HubSpot admin hug is tomorrow, Tuesday September 1, on attribution reporting. Kyle Jepson is back after a month off, and if you also took August off, we are putting together a catch up page with the month's updates and clips on the Another Orange Morning site.
The Rest of Monday's Updates
Sandbox creation copying permission sets from production for easier UAT and the Notetaker with smart deal progression both got named on air without a deep dive. Three more shipped that we did not get to: the rebuilt CRM index page and its flexible views, the Node.js 18 and 20 deprecation for chatflows custom code snippets, and bulk unmark spam in help desk. Everything else that shipped is in 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. We will see you tomorrow morning.
Produced by Value-First Media