There was a line from Aziz Ansari’s (hilarious) set Thursday night that kinda nailed the backdrop for UNBOUND this year.
The joke was basically, “AI leaders are saying chatbots are going to end humanity in the next decade and you guys are all like ‘Okay, great but how do I optimize my CRM?!’”
Exciting times! Since we still have a few years before the apocalypse, let’s look at what HubSpot announced at UNBOUND this year so we can all optimize our CRMs… before it’s too late.
As always, there was a lot – and HubSpot’s full rundown is here. From our perspective (Charlie and Jill), this is the roundup of stuff that’s worth your attention.
Growth Context: HubSpot's admission that AI is a data problem
As a firm whose focus is Data Trust, this conference was especially validating to hear that the problem we're focused on solving is so ubiquitous. Trust was the word of the week.
HubSpot's own framing for this cycle is blunt: "Growth Context is everything HubSpot knows about your customers, your team, and your business. The more of it you have, the better your results." It's an intuitive idea, but it's definitely a gap for most businesses we talk to.
Growth Context is the harness sitting underneath Breeze, the agents, and pretty much everything AI-flavored in the portal now. Practically, it shows up as Context Home, a new place to see and manage what HubSpot knows about your business in one spot, so every team and every agent is pulling from the same source instead of each one guessing independently.
The catch is a point we've been making to clients since before "AI readiness" was a category: this only works if what's in there is trustworthy. Growth Context doesn't fix dirty properties, duplicate records, or three different definitions of "qualified lead" living in three different teams' heads. It just means those problems now show up in your AI outputs faster and more visibly than they used to show up in your reports. If your data hygiene was a background problem before, it's a front-and-center one now.
If you're going to use Context Home, treat it like a practice, not a setup task: audit what's already in there, configure it against how the business runs today, and maintain it as that changes — because it will. Before you hand HubSpot the keys, it's worth answering two questions internally: What's your current process for catching data errors before they compound? And who owns AI context settings when your product, ICP, or positioning changes? If nobody on your team can answer either one, that's the actual starting point — not Context Home itself.
Also worth noting, as we all rush to update our context and provide robust, accurate data about our businesses, processes and ICPs — what an incredible gift to HubSpot! How they decide to use all of this data, I guess we'll find out.
Self-updating CRM = less data entry, more to govern
The self-updating CRM logs calls, emails, and meetings automatically, so records stay current without someone manually updating a deal stage or logging a call after the fact. For any RevOps team that's fought the "reps won't update the CRM" battle, this is a real, welcome attempt to fix a real problem.
But also, this is one to be cautiously optimistic about. Automatic capture means more data flowing into the CRM with less human review at the point of entry. That's great for volume but a possible disaster for trust if nobody's watching what's landing in those fields. Self-updating is a data quality accelerant in both directions — it'll amplify whatever data discipline (or lack of it) already exists in your portal.
As with most updates, I expect this will reduce some friction while also creating new work to ensure the data can be trusted.
Breeze Assistant, Agent Hub, and Agent Builder
The bigger structural change is that HubSpot rebuilt Agent Hub and Agent Builder around a simple premise: "every agent your team needs, including the ones you build yourself." Breeze Assistant is positioned as the go-to-market layer — draft a campaign plan, write a nurture email, spin up a landing page — while Agent Builder is where ops and marketing people, not just engineers, can now assemble custom agents for the specific, unique ways their teams prefer to work.
Agent building has been a developer's game inside most platforms up to now. If HubSpot has genuinely lowered that bar for non-technical admins, that's a real shift in who gets to automate what — and it puts more pressure back on the same question: whoever builds the agent needs to truly understand what data it's working from, or it'll automate the wrong thing very efficiently.
Notable updates for fellow agencies
Two additional updates that didn't make the keynote highlight reel but are probably interesting if you're running an agency.
- Microsoft Advertising integration is now on par with Google's for conversion tracking, attribution, and targeting — worth a look for any paid demand gen manager running this channel.
- The HubSpot connector for Claude also picked up enhanced build and query actions. This is good to know as we see more prospects and clients doing real work in Claude alongside HubSpot instead of living entirely inside Breeze. We also sat in on a workshop that showed how the CLI can be used to map and visualize HubSpot data much more efficiently than using the MCP.
The big takeaway
Big picture, what many of us have been predicting is now happening: the build part of GTM operations is now fast and easy to the point that anyone can do it in a fraction of the time it used to take. And as we've seen, much of that has resulted in an endless tsunami of AI slop that largely gets ignored. But it has also resulted in some really innovative ways businesses have been able to transform themselves and reduce often soul-crushing manual work better left to the bots.
The businesses that will succeed in this moment are the ones who invest in more strategy, more training, and (of course) better, trustworthy data. At a minimum, this is a great time to take a beat and really think: what outcomes do we want to drive with AI?
If you're trying to figure out what Growth Context or the self-updating CRM mean for your specific portal, that's the exact type of conversation we have with clients every week.
