Updates blog — Tuesday, May 19, 2026

HubSpot Updates: May 19, 2026

Two updates landed Tuesday, and both are graduations. The Agentic Automation Builder moves from in-development status into private beta with portals already running custom agents getting access automatically. Personalization tokens in rules-based chatflow welcome messages move from private beta into public beta. Different surfaces, same direction: features the show has been talking about for weeks are now in real hands.

If you watched the show today, you already heard the conversation around both, including a Will It Workflow that put the new builder through a live invoice-follow-up scenario. If you did not: Watch the recap to hear our thoughts on these updates and more.

Platform Updates Detailed

Agentic Automation Builder

The Agentic Automation Builder is a new application inside HubSpot that brings workflows, AI agents, and data from across the business onto one modern canvas. Triggers come from HubSpot CRM data, scheduled times, webhooks, third-party integrations, and other events. From there, you can add traditional workflow actions, run AI agents as steps, branch on conditions, and reuse common processes. The framing HubSpot uses is that you no longer choose between a workflow and an agent. You build both in the same place. This is in private beta across the Professional and Enterprise tiers of Marketing Hub, Sales Hub, Service Hub, Data Hub, and Smart CRM, and access is automatically enabled for portals already enrolled in custom agents.

Why it matters: The architectural decision is the part worth flagging. HubSpot is shipping this as a new application rather than evolving the existing workflows tool, which signals the underlying infrastructure was rebuilt, not resurfaced. The new builder treats non-CRM third-party events as first-class triggers, which means a row added to a Google Sheet, a file landing in Office 365, or an external webhook can start an automation directly without a custom integration in the middle. The other piece worth tracking is credit consumption. Triggers off buyer intent signals like mergers and acquisitions burn credits per company tracked, and the new builder does not yet make credit usage visible inside the canvas itself. Until it does, the cost of running these automations at scale is something to model before you turn them on.

For more details: Agentic Automation Builder private beta.

Personalization Token Support in Rules-Based Chatflow Welcome Messages

You can now customize a rules-based chatbot's welcome message using CRM data. Insert personalization tokens like first name, company name, or recent activity directly into the initial greeting. When a visitor is recognized via tracking cookies, the token renders automatically. When they are not, configured default values keep the greeting natural. This is now in public beta across all hubs and tiers, a step up from the private beta the feature launched in.

Why it matters: Personalization tokens in the first message of a chat are a sharper tool than they look. Used on a logged-in customer portal or a customer success room, they feel like the company finally caught up with what it already knows. Used on a stranger's first visit to a marketing page, they cross into uncanny territory and undermine trust. The other thing the feature requires is clean data. If the first name field is full of all-caps entries, email addresses, or the wrong name entirely, a personalized welcome makes the data problem visible to the customer in the rudest possible way. The right deployment is post-consent surfaces with clean data. That is a narrower deployment than the feature description suggests, and it is the deployment that actually works.

For more details: Personalization token support in rules-based chatflow welcome messages public beta.

Key Takeaways and Next Steps

Treat the new builder like the beta it is: The Agentic Automation Builder is the most significant automation release HubSpot has shipped in a long time, and not everything works yet. Breeze is the precedent worth holding in mind. It was rough at launch, has since matured, and a handful of bad early experiences still steer people away from it for things it now does well. Open the builder, test a real use case, give feedback, and check back regularly. Do not write it off if the first attempt does not land, and do not move production workflows onto it yet.

Credit transparency is the missing piece: The new builder makes it easier than ever to trigger automations off buyer intent signals and run AI agents as workflow steps. Both of those cost credits. The builder does not yet surface where credits are consumed inside the canvas, and credit-consuming properties are not visually distinct on record views either. Until that gap closes, model credit usage before you scale anything, especially on triggers that fire frequently or run on every contact in a list.

Personalization belongs where consent has already happened: The chatflow welcome token update is powerful, but the question is not whether to use it. The question is where. Authenticated customer portals, customer success rooms, and post-form-fill chat interactions are the right places. Anonymous first-visit pages are not. Build the rules around recognition state, not around the technical capability.