HubSpot’s AI can now write, build, and publish a landing page on its own, and that is genuinely good news for marketing teams that have spent years waiting on production queues. The question worth asking is not whether the capability works. It is who signs off on what goes live, and how you keep the speed without losing the standards.
The June 2026 release quietly moved HubSpot past assistance and into execution. Until recently, Breeze and its cousins suggested copy, drafted emails, and summarised records. A human still assembled the asset and pressed publish. That line has now shifted, and it is worth understanding exactly what changed and what it asks of the teams using it.
What did HubSpot actually ship in June 2026?
In its June 2026 developer rollout, HubSpot added Landing Page Creation to its Remote MCP server. In plain terms, an AI assistant connected to a HubSpot account can now create, edit, and publish landing pages. It writes headlines and body copy, adds and reorders sections, adjusts styling, manages the form embeds that capture leads, and pushes the finished page live. A confirmation step sits in front of publishing, and every page stays fully editable inside HubSpot afterwards, so nothing the assistant produces is locked away.
It did not arrive alone. The same month brought broader access to the Prospecting Agent, a redesigned Breeze Projects workspace, richer reporting, and a new visual theme that becomes the default for every account on 31 August. Read together, these updates point in one direction: HubSpot wants the platform to do more of the doing, not just the advising. Landing pages are simply the most visible example, because a landing page is a public, branded asset that a prospect actually sees.
Why this is a real gain for marketing teams
Start with the upside, because it is substantial. Production has always been the slowest and least celebrated part of campaign work. A good idea can sit for a fortnight waiting on a designer, a developer, or an available slot in the web team’s backlog. An assistant that can stand up a working, on-template page in minutes removes a bottleneck that has frustrated marketers for as long as landing pages have existed.
The leverage lands hardest where it is needed most. A two-person marketing function at a manufacturer or an energy firm can now run the kind of campaign cadence that used to require an agency retainer. Testing becomes cheaper, because spinning up three variants of a page is no longer a day’s work. Momentum improves, because the gap between deciding to run something and having it live shrinks from days to minutes. Handled well, this is a genuine expansion of what small teams can achieve, and it is worth being optimistic about.
It also changes the shape of the marketer’s day in a way most people will welcome. The hours that used to go into wrestling a template into shape, chasing a developer for a small fix, or rebuilding a page for the third time can now go into the parts of the job that machines cannot do: understanding the audience, sharpening the offer, and deciding what is actually worth saying. The tool does the assembly so the person can do the thinking. That is a healthier division of labour than the industry has managed for a long time, and teams that lean into it will feel the difference within a quarter.
Can HubSpot AI really create and publish landing pages?
Yes. Through the Remote MCP server introduced in June 2026, an AI assistant linked to a HubSpot account can build a landing page end to end: drafting the headline and copy, choosing and arranging sections, applying styling, embedding a lead-capture form, and publishing the page live. HubSpot requires an explicit confirmation before anything is published, and the finished page remains editable in the normal HubSpot editor. So the automation is real, but it is designed to keep a human in the loop at the moment of going live rather than removing people from the process entirely. The practical difference is that the assistant now does the assembly, and the marketer moves from builder to reviewer.
Who should sign off on what the AI publishes?
This is where the interesting work begins. A confirmation step is only as good as the person clicking it, and the risk in a fast tool is that the confirmation becomes a reflex rather than a review. If the same individual briefs the assistant, reviews the output, and approves it, there is no independent check at all. The page reads well, the form works, and it goes live, which is exactly how a claim that is slightly off, a price that is out of date, or a compliance line that is missing slips onto a public URL.
The answer is not to slow everything down with heavy sign-off chains that defeat the point of the tool. It is to be deliberate about which pages carry real consequence. A gated-content landing page for a webinar can sit comfortably under a single reviewer. A page that quotes pricing, makes a regulated claim, or represents the brand to a new market deserves a second set of eyes before publishing, whether or not an AI drafted it. The governance question is not human versus machine. It is matching the level of review to the stakes of the page.
There is a useful precedent here. Most organisations already accept that not every email needs legal review and not every contract can be signed by anyone, and nobody finds that controversial. Publishing has simply never had the same clarity, because building a page was slow enough that the friction acted as an accidental control. Now that the friction is gone, the control has to be intentional. The good news is that this is a solved problem in other parts of the business, and the same instinct, more care where the stakes are higher, transfers cleanly to what marketing now publishes at speed.
How do you keep brand consistency when AI writes the copy?
Brand consistency with AI-generated copy comes from what you give the assistant to work from, not from correcting it after the fact. Teams that get reliable output feed the model a clear, current source of truth: an approved messaging framework, a tone-of-voice guide, product and pricing facts that are known to be accurate, and examples of pages that represent the brand well. They keep that reference material in one governed place rather than scattered across inboxes and slide decks. The assistant then drafts within known boundaries, and review becomes a check against a standard rather than a rescue mission. Consistency, in other words, is an input discipline. The organisations that struggle are usually the ones that never wrote the standard down in the first place.
What does good governance for AI-published assets look like?
A workable model rests on four ideas, and none of them require slowing the team down. First, ownership: someone remains accountable for every published page, and the tool does not dilute that just because it did the typing. Second, tiered approval, so low-stakes pages move fast and high-stakes pages get the scrutiny they warrant. Third, guardrails at the source, meaning accurate, current reference material the assistant is expected to draw on, plus clear rules on what it must never invent, such as prices, claims, or customer names.
Fourth, and most often skipped, measurement. Because pages can now be created so cheaply, the real risk is not a single bad page but a slow drift into volume without quality: dozens of near-identical URLs that dilute your message and confuse your analytics. A light monthly review of what the assistant has published, how those pages performed, and whether any should be merged or retired keeps the speed honest. Set up this way, governance is not a brake on the AI. It is the thing that lets you trust it enough to use it properly, which is the whole point.
The Sirocco perspective
We see HubSpot’s move as a clear net positive, and we would encourage marketing teams to embrace it rather than wait. The production bottleneck was real, and removing it frees people to spend their time on the parts of the work that actually need judgement: the strategy, the offer, and the message. In our experience across HubSpot, Salesforce, and Dynamics 365, the teams that get the most from tools like this are the ones that decide their governance before they scale their output, not after something has gone live that should not have. That usually means a short, written messaging standard, a simple rule for which pages need a second reviewer, and one owner who watches the whole thing. It is a modest amount of upfront work, and it turns a fast tool into a dependable one.
If you are rolling out HubSpot’s AI publishing and want to put sensible guardrails around it before your team scales up, we are happy to help you design them. You can schedule a consultation whenever it suits.
Get in Touch
If your team is weighing how much of its marketing production to hand to HubSpot’s AI, and where the approval gates should sit, we would be glad to think it through with you.
