Show recap

Show recap — Friday, September 25, 2026

Friday, September 25, 2026 · with Casey Hawkins, Madelyn Donovan

From the show

Another Orange Morning Recap: September 25, 2026

Friday's show was a two-person room, one week after UNBOUND. Casey Hawkins held the host seat and Madelyn Donovan was the Friend of the day, once her audio decided to work. Chris Carolan was still away at Moonshot, an invitation-only AI gathering on the West Coast, which left Casey running the screen share for the first time and both hosts admitting the show is nothing without its bearded friend. The update volume, meanwhile, is back to normal.

Here is what we got into: the UNBOUND updates nobody was talking about, and the case for quality-of-life improvements over agents; the Nurture Agent and what it is really running on; HubSpot Work on its second day in public beta; Email Quality Checks and the custom checks hidden inside them; Campaign Agent and the argument for doing your planning inside HubSpot; a Libra-season HubSpot horoscope; and targeted pages for the account-based marketers who have wanted them for years.

The UNBOUND Updates Nobody Was Talking About

The show opened on a Kyle Jepson post from the day before, a multi-page rundown of things you might have missed at UNBOUND. The loudest announcements were the agents: Campaign Agent, Content Agent, Nurture Agent, Agent Builder, agent workflows, Revenue Agent. Kyle's point, and the reason it landed with this room, is that the loudest ones did not resonate with all of the HubSpot nerds, and the HubSpot nerds are this show's bread and butter. He had already presented the quieter list at a HUG.

Casey's personal favorites off that list were cross-object filtering on index pages, cell selection on the CRM index, and property documentation. The filtering one is personal. A few weeks ago she met with the product manager for filtering and said the thing every admin has felt: filters act differently depending on where you are, on segments, on index pages, inside workflows. Any expansion there is welcome. Kyle's own pick was the twelve-column reporting grid.

Asked how she came out of UNBOUND, Madelyn said curious, and a little overwhelmed. If you have not had a need or an interest in the agent side of HubSpot before, a week like that can make you feel behind, and a lot of people in her circle feel the same way. The things she is most excited about are the small quality-of-life improvements. That morning she noticed she could select multiple cards on a board view, which may or may not be the same underlying update as cell selection.

Why it matters: the room landed on a useful distinction. Agents fundamentally change how you use HubSpot, while quality-of-life updates improve the way people are already using and navigating the tool, and the nerds get excited about the before and after they can see. With the agents, the agents are the nerds. Casey had posted on LinkedIn the day before about where AI works best for her, and the uncomfortable part is that having AI figure out a workflow for her, while more efficient, removes some of the ownership. Madelyn's word for it was agency. The flip side, which both agreed is the real gain, is that people who could never build a workflow now can.

The Nurture Agent, and What It Is Really Running On

Nurture Agent

Madelyn wanted this one, and the show had never covered it. The Nurture Agent personalizes your automated nurture emails for every recipient at send time, drawing on CRM data, engagement signals and live web research, and it rewrites the subject line, preview text, body copy and call to action individually for each contact. The pitch is that marketing teams talk about one-to-one nurture and almost nobody does it, because it is so hard to execute at scale. It is in public beta for Marketing Hub Professional and Enterprise, priced at ten credits, which works out to about ten cents.

Madelyn's first reaction, as a marketer who has been burned before, was somewhere between stress and a red-pen essay feedback flashback. The thing that settled her was reading the setup screen. On the left-hand side it is functioning off personalization tokens, which is exactly how you would build this by hand today, with properties as tokens, smart rules, or a cloned email per segment. The agent is doing that logic for you once you pick the tokens, and you can choose which activity it pulls from. There is a compare tab to see the before and after, though not the manual control Madelyn was hoping for.

Why it matters: the fields you pick are the whole thing. A clean job title property, which is not best practice but exists in a lot of portals, an industry field a client fills by hand because they only sell into three segments, a company name on the contact that somebody actually populated. Those are the tokens this agent should be reading. As with anything AI, it is as good as your data, and both hosts know a lot of not great data sets. Data cleanliness, fail as old as time.

HubSpot Work, Second Day Running

HubSpot Work

HubSpot Work is Madelyn's number one, and everyone she talks to is curious about it, so the show ran it again after Wednesday's deep dive. It moved from private to public beta around UNBOUND. Once you are enrolled it may take a few minutes to appear, and it lives in the left navigation under Work Management, the little star with dots around it, possibly under More until you pin it. The update card links straight to it, for anyone like Casey who was asking where it was.

The what is AI work management natively connected to your CRM. Describe what you need in plain language and it builds projects, tables, documents, forms and agent automation from a single prompt, replacing the spreadsheets and project tools where the work around the CRM has always lived. The why is the question Casey got tripped up on Wednesday too: why do this in HubSpot rather than a project management tool. HubSpot's answer is that when operational work and customer data live in different places, three things break. The person running onboarding cannot see the deal history without switching tools, nobody shares a clear view of what needs doing and who owns it, and it becomes nearly impossible to build agents that do meaningful work because they do not have the full picture. The connected CRM part is what Madelyn is most excited about.

Why it matters: Casey got a question about use cases, and marketing is the cleanest one. If you are running a webinar or any HubSpot campaign, it will help you build it, and agents can step in on publishing content or creating landing pages. The HubSpot Work team also pointed out to her that it works for client projects, because you can associate tasks with a specific client and track their progress. Madelyn's example is a client whose feature requests currently create Jira tickets so they know which customer to alert when the feature ships; tracking that here, associated with the customer directly, is a good fit. It also sidesteps the debate people have had about the projects object and whether to tie it to tasks, since this is its own thing, and marketing content is the least ingrained place to start.

Email Quality Checks and the Custom Checks Inside Them

Email Quality Checks

Casey saved this one specifically for Madelyn. Email Quality Checks are a built-in AI review in the email editor that scans before you send: spelling and grammar, spam triggers, broken links, missing token fallbacks, data formatting, a weak first sentence, off-brand voice and more. Today that review lives across test sends, browser extensions and Slack pings to a teammate. This brings it into the editor you are already in.

Broken links was the check that stood out to Madelyn immediately, ahead of spam triggers. She has started seeing this in the portals she works in, and as a self-described terrible speller, spelling help is one of her favorite things AI does. Both hosts have used the Grammarly plugin and both have found it intrusive on other sites, so having it built in is the win. The hidden gem is that you can create custom checks.

Why it matters: custom checks turn tribal knowledge into a guardrail. Casey had a client who was particular about orphan lines, the single word stranded on the last line of a paragraph, and while it is not clear the checks get that granular, it is the kind of formatting rule this could hold. Madelyn's version is the product marketer's: you tell the sales team to always call it this and never describe it like that, and whoever is doing your emails that week may not know. A check that catches the wording is the thing that keeps the brand consistent without another Slack message. Marketing Hub Starter, Professional and Enterprise, public beta.

Campaign Agent, and the Case for Doing It in HubSpot

Campaign Agent

Madelyn asked whether the show had looked at the Campaign Agent yet, and Casey searched rather than scrolled. You can now chat directly with Breeze Assistant to turn a goal or an insight into a complete campaign plan, with channels, audience, messaging and assets all connected to the same objective. Building a campaign in Marketing Studio used to require a campaign brief up front, even when you did not have the details. Now the agent asks clarifying questions, suggests assets aligned with your goals and brand, and builds the campaign automation so more of the time goes to getting the campaign live. Marketing Hub Professional and Enterprise, public beta.

Why it matters: this is the overarching thing Casey is going to keep saying. It will be really helpful for teams once they can change behaviors, and behavior is the hard part. She already does this kind of planning with AI, and if you have used some other AI tool for years, switching is hard because the old one already works and the new one will get things wrong the first time. But if you do it in HubSpot, you are giving HubSpot more data, it can look at how past campaigns performed, your team gets more visibility, and it keeps getting better the more you use it, all in one place instead of across tools. Madelyn's framing closed it: switching any software costs you an adjustment, but it is different when it is already your CRM. You are not going to regret having the data in the system you use for everything else.

Horoscopes: Libra Season

Madelyn's standing Friday segment moved into a new season this week. Libras are fairness-obsessed, relationship-minded and weighing both sides; a planet here will not move until the scale looks even. The HubSpot horoscope for Libra asked who decides which version someone sees, and lined up three betas along that scale. Personalized variants let one landing page, website page or CTA carry versions for specific companies, contacts or segments on a single URL, and HubSpot chooses. Targeted pages build a tailored page per company from a template, and a rep chooses. The Nurture Agent rewrites each email per recipient at send time. Machine judgment on one side, human judgment on the other, and you will want both weighed before anything goes out.

Neither host knew about personalized variants, which is the point of the page. Each reading shows which updates are informing it, so when a horoscope gets you excited you can click through to the specific update. If the volume of updates feels overwhelming, Madelyn's advice is to read only the ones relevant to your sign. Casey, an Aries who always feels like every other sign, is going to be a Libra for a while.

Targeted Pages, for the People Who Have Wanted Them

Targeted Pages: Personalized Landing Pages at Scale

The horoscope pointed straight at the last update of the morning. Targeted pages, in private beta for Sales Hub and Marketing Hub Professional and Enterprise, generate tailored landing pages for specific companies or contacts from a reusable template, each with its own URL. Breeze builds the pages from your HubSpot data and your personalization instructions, and a rep or marketer chooses which page and link to send. Unlike personalized variants, HubSpot does not decide who sees which page.

Why it matters: Madelyn has always wanted to do this and never has, because by hand it is a real pain. Realistically you can only pick about five companies, they have to be big enough to be worth a dedicated page, and the moment a rep decides they want to target a sixth, they have to go back to marketing and ask. Some tools do part of it. A scaled version is the one she would use. It is built for account-based campaigns, sales prospecting and one-to-one outreach, worth requesting access if that is a real workflow for you, and a skip if your page strategy is broad audience. Casey's takeaway going into the weekend was the same thread: targeting, personalizing, and balancing the human with the AI.

The Rest of Thursday's Updates

Five updates shipped on September 24 and none of them made it to air. There is now an API tester for customer agent actions, so admins can send a test request and see the full request, response and errors from the action configuration screen before an action goes live. WhatsApp voice notes are in public beta for Service and Marketing Hub Professional and Enterprise, sent and received natively from HubSpot conversations. The knowledge base agent picked up custom instructions, so you can set a target audience and guardrails for the articles it drafts. Customer mapping for QuickBooks Online and Xero creates a lasting link between HubSpot records and your accounting customers that survives name and email changes. And composing emails on HubSpot mobile with Breeze Assistant is live for every hub and tier, which the show covered in depth on Thursday. Everything else that shipped is in the HubSpot updates feed.

The Sign-Off

The show is back Monday at 9:15 Eastern, 8:15 Central. Chris returns from Moonshot and Joshua Oakes joins for the Monday seat.

You probably already own the value you are looking for in HubSpot. You just need to wake up to it. We will see you Monday morning.

Produced by Value-First Media