HubSpot's refounding means the CRM stops being the screen your reps live in and becomes the record and context layer that AI agents work from, whether that agent is Breeze, Claude or ChatGPT. For revenue leaders, the real question is not whether to switch CRMs. It is whether your data, your workflows and your people are ready to work through agents.
I heard the news the way a lot of HubSpot partners probably did: in the middle of a call about something else. I was working through a prospecting agent rollout with my HubSpot partner team when one of them asked if I had seen that the CEO was out and Brian Halligan was back. I had not. By the end of the call we had stopped talking about the deal and started talking about what this means for every client I have on HubSpot, and every client I have stuck on Salesforce.
Here is my read, as a HubSpot Gold Partner and an Anthropic partner who deploys both every week. Weigh it accordingly.
The backdrop is not subtle. HubSpot's stock was down about 50% for 2026 heading into the announcement. Net revenue retention fell to 101% in Q3 2026, Q4 guidance implies growth near 12%, and the company cut roughly 660 roles in October. A founder does not come back to run the same play.
HubSpot has not defined it yet, so here is how I read it.
In AI, the harness is everything wrapped around a model that turns it into something useful: the data it can reach, the tools it can call, the permissions it runs under, and the rules it follows. Two products built on the same model behave completely differently because their harnesses are different.
A customer relationship harness is that layer for go-to-market. Your contacts, companies, deals, conversations and history, plus the actions an agent is allowed to take on them. "Headless" means the head, the user interface, becomes optional. The record stays. The screen is one way in among several.
This was not invented on Thursday. The pieces have been shipping all year:
What changed on November 12 is not the direction. It is who is driving and how fast.
Not the stock price. This one, from HubSpot's Q3 2026 earnings call:
Read that as an operator, not an investor. HubSpot's own data says the teams that work through an AI front end stay longer and get more value. That matches what I see in the field. When a rep can ask Claude what is at risk in their pipeline and get an answer built from the actual record, the CRM stops being an administrative tax and starts being the thing that makes them smarter before a call.
| Classic: reps work in HubSpot | Embedded: Breeze agents inside HubSpot | Headless: Claude or ChatGPT on top of HubSpot | |
|---|---|---|---|
| Where the work happens | HubSpot screens and the mobile app | HubSpot, with agents doing prospecting, scoring and service work | The AI workspace, with HubSpot as the record behind it |
| What you pay for | Seats and contact tiers | Seats plus HubSpot Credits | Seats, credits and an AI subscription |
| What breaks first | Adoption. Reps treat it as admin. | Data quality. The day you launch a prospecting agent is the day every data integrity problem shows up. | Permissions and governance, if nobody owns them |
| Best fit | Teams still putting a basic sales process in place | Teams with clean data and a defined ICP | Teams with real AI fluency who already work in an AI workspace |
Most of my clients will run all three at once for a while. That is fine. The mistake is assuming the classic column is where you will stay.
People. One of my HubSpot colleagues said it best after hearing the new direction: it is the right vision, and it is not where the market is today.
I run AI fluency programs for sales organizations. Most people I meet think they are good at AI because they use Copilot to look things up. That is a search engine with better manners. Working through agents means delegating judgment-heavy work, checking it, and building it into a daily routine. That is a behavior change, and behavior does not change because the architecture did.
There is an upside here. HubSpot still looks and works like a CRM people recognize. During the transition, that familiarity is an asset, not a liability. You can move a team from the classic column to the headless column one workflow at a time, without asking anyone to abandon the tool they know on day one.
No sale has been announced. Brian Halligan told Bloomberg on November 12, 2026 that he does not want to see HubSpot swallowed up by a private equity firm or shrink, and that he came back to rebuild and reimagine the company. The board says it is united behind his refounding.
Nothing new was announced with the CEO change. HubSpot was already moving to hybrid pricing, combining seats with HubSpot Credits that agents consume, and moved two Breeze agents to pay-per-result pricing in April 2026. Expect more of that mix, and ask your account team before any 2027 renewal.
Reps will spend less time clicking through records and more time asking an AI assistant to prepare, summarize and act on what is in the CRM. HubSpot remains the system of record. The change is in where the work gets done, and that only pays off if reps are trained to delegate to agents well.
No. Cleaning CRM data, defining personas and segments, and piloting an AI connector with a small group pay off under any version of the plan. HubSpot's Q3 2026 data already shows connector users retain better and expand faster. Waiting just means starting later.
Yes. A headless CRM lowers the cost of switching front ends, because reps increasingly work in an AI workspace rather than in the CRM screens. That makes a phased approach realistic: keep Salesforce as the record, add HubSpot where it adds value, and migrate over time instead of all at once.
Steve Cadley is Managing Partner of SalesXecution in Alpharetta, Georgia. A HubSpot Gold Partner and Anthropic partner with master certifications across Miller Heiman Strategic Selling, SPIN, MEDDIC, Challenger and Sandler, he works with mid-market and private-equity-backed organizations to build the AI fluency that makes their CRM, methodology and agent investments operational.