HubSpot’s UNBOUND 26 releases move the platform beyond individual AI features toward a more connected agentic customer platform. Businesses can build and manage AI agents, give those agents shared business context, use AI across marketing, sales, and service workflows, and connect those actions to CRM data and customer intent. Taken together, the releases point toward a different way of operating in HubSpot, one where AI doesn’t just help users complete isolated tasks, but increasingly participates in the workflows that move customers through the lifecycle.

For teams evaluating what to do next, you must understand where AI can improve an existing customer or revenue process, make sure the context and data behind that process are strong, and then introduce the right agents and automation around it.

Agent Builder, Agent Hub, Context Home, Marketing Studio 2.0, Content Agent, Nurture Agent, Prospecting Agent, Deal Progression, Customer Agent, Revenue Agent, Buyer Intent, Data Agent, and the supporting CRM capabilities are increasingly connected by the same underlying idea: Customer-facing teams should be able to understand what is happening, decide what needs to happen next, and execute more of that work from the same customer platform.

In marketing, that changes how campaigns are planned and optimized. Sales, it changes how prospects are researched and deals are advanced. Service, it changes how customer conversations are resolved and improved. And underneath all three, better data, intent signals, and shared AI context become much more important.

Here are the five transformations that make the UNBOUND 26 releases make sense as one platform story.

1. HubSpot is an operating layer for AI agents

The first major shift is architectural: HubSpot is creating a more centralized way to build, manage, contextualize, and run AI agents.

Agent Hub provides the home for that system, while Agent Builder and agentic workflows give teams a way to create custom agent-powered processes. Context Home gives HubSpot’s AI tools shared information about the business they are operating inside.

HubSpot’s updated Agent Hub workflow experience allows teams to create workflows that combine traditional automation with AI actions and custom agents. Workflows can begin with actions or triggers, incorporate agents directly into the flow, and work with information from CRM records as well as supported external sources. HubSpot also distinguishes these Agent Hub workflows from legacy workflows in several ways, including more flexible triggers and the ability to run custom agents as actions inside the workflow.

A traditional workflow is generally designed around rules: when X happens, do Y. Agentic workflows can introduce a reasoning step into that process. Instead of every possible outcome needing to be predetermined through branching logic, an agent can perform a defined job inside the workflow and use the information available to it to produce an output. And that is where Context Home becomes especially important.

HubSpot describes context as foundational information about a business that can be automatically applied across multiple AI tools. Teams can maintain information about the business, its products and services, customers, brand, market, and other organizational details so HubSpot’s AI features have a common foundation to work from.

The practical takeaway is that AI governance is no longer just about choosing which tool somebody can access. Teams also need to think about what their AI systems know, where that knowledge comes from, and whether it is current. For organizations beginning to adopt HubSpot’s AI agents, this is a smart place to start: establish the context and workflows first, then decide where an agent can improve the process. That foundation carries directly into the next four transformations.

2. Marketing is becoming an agentic, continuous loop

UNBOUND 26 also changes the marketing story inside HubSpot. Instead of treating campaign planning, content creation, distribution, nurturing, and measurement as separate activities, HubSpot’s new marketing capabilities point toward a more connected operating model: understand → create → distribute → nurture → measure → learn → repeat.

At the center of that shift is Marketing Studio 2.0, which HubSpot positions as a unified marketing workspace where insights feed campaigns, campaigns drive outreach, and campaign results inform what marketers do next. That is a meaningful evolution from simply adding generative AI into individual marketing tools.

Two featured agents help illustrate that change:

Content Agent is being rebuilt around research before creation. According to HubSpot’s Spotlight product positioning, it can incorporate information such as brand voice, audience data, and competitive context before generating assets including blog content, LinkedIn content, and landing pages.

Nurture Agent extends the same idea into email nurturing. HubSpot positions it around personalizing nurture emails using CRM data, engagement signals, and live research.

Content Agent supports the creation side of the marketing lifecycle. Nurture Agent supports the continuation of the buyer journey after engagement begins. Marketing Studio provides the operating environment that connects those activities back to the broader campaign.

Put together, they point toward a marketing system that can become more responsive as customer information and campaign results change. For marketing leaders, the question therefore becomes less “Where can we use AI to write?” and more “Where in our campaign operating model can AI help us understand, create, personalize, and respond faster?”

3. Sales AI is moving from productivity assistance to deal execution

The sales releases tell a similar story, HubSpot is connecting AI to more of the seller journey: find the right buyer → understand the buying committee → engage → advance the deal.

Two featured products sit at the center of that transformation: Prospecting Agent and Deal Progression.

Prospecting Agent can research contacts and target accounts and execute outreach strategies. HubSpot allows users to manually enroll contacts or automatically enroll them using defined rules, and the agent can also research target accounts and generate emails based on that research.

The UNBOUND 26 update pushes Prospecting Agent further. HubSpot’s Spotlight positioning highlights contact sourcing, emails that can improve based on rep edits, and the ability to align prospecting plays with a company’s brand voice. That shifts the role of AI earlier in the selling process. Instead of a rep beginning with a blank screen and manually collecting all the information needed for outreach, the system can take on more of that research and preparation.

Deal Progression addresses what happens after conversations begin. HubSpot’s Deal Progression card uses AI to analyze deal context, including logged calls, meetings, and emails, and recommend next actions. It can surface action items, suggest CRM property updates, and help users generate follow-up replies without leaving the deal record. When a sales methodology is configured, HubSpot can also use that methodology as additional context for its recommendations.

Prospecting Agent helps answer: Who should we engage, and how?

Deal Progression increasingly helps answer: Now that we are in a deal, what should happen next?

Buying Groups adds another layer by helping teams map stakeholders and understand the people involved in a purchase, connecting the individual contact experience to the broader buying committee. HubSpot’s Spotlight roadmap positions these capabilities together around reducing research and administrative work while giving sellers more support throughout the actual revenue motion.

For sales organizations, the implementation opportunity is not giving every rep another AI tool. It is identifying where sellers repeatedly lose time or momentum, account research, stakeholder mapping, follow-up, CRM updates, next-step identification, and deciding which portions of that process can be supported by HubSpot’s agents.

4. Customer Agent is an operating system for agentic service

Some of the most extensive UNBOUND 26 investment is happening around Customer Agent. That is why Customer Agent deserves to be understood as more than a chatbot update.

HubSpot’s Fall Spotlight roadmap expands Customer Agent across setup, routing, targeting, human handoff, actions, knowledge, testing, coaching, audience segmentation, performance reporting, channels, and voice.

At its core, Customer Agent uses HubSpot AI and existing business content to automatically respond to customer questions. Based on the situation and its confidence, it can provide a source-backed response, ask additional questions, or reassign a conversation to a human.

Teams can define guidelines for when conversations should be handed to human representatives. Customer Agent can be routed conversations when human agents are unavailable. Businesses can control which audiences Customer Agent engages with, personalize its behavior for different segments, test responses before deployment, analyze its performance, and use insights from prior support tickets to develop new knowledge for the agent. HubSpot is also extending Customer Agent into voice interactions.

The strategic direction here is: automating customer service is not just about answering more questions automatically. It requires systems for control, knowledge, testing, escalation, personalization, and improvement.

That is what makes many of the seemingly smaller Customer Agent updates important: audience segmentation gives teams more control over who receives which experience, testing gives teams a way to evaluate the agent before expanding its responsibilities, handoff guidelines create clearer boundaries between AI and human support, performance reporting makes it possible to evaluate what the system is actually doing, knowledge generation from ticket history creates a way for past customer interactions to contribute to future service, and together, these capabilities create a loop: deploy, observe, learn, improve.

For customer success and service leaders, that changes the implementation question. The question is not “Should we turn on Customer Agent?” A better question is: Which customer interactions can Customer Agent handle well, what knowledge does it need, when should a person take over, and how will we continuously evaluate the experience?

5. Intent, CRM data, and revenue intelligence are the connective tissue

The fifth transformation may be the least flashy, but it may be the one that determines whether the other four work. AI agents need information.

They need to know what happened with the customer, what the business knows about them, what signals they are producing, what the company sells, and what is happening across the broader customer lifecycle. That is why Buyer Intent, Data Agent, Revenue Agent, Context Home, Smart CRM, and Data Hub capabilities belong in the same UNBOUND story as Agent Hub, Prospecting Agent, and Customer Agent. The content strategy identifies this layer as the connective tissue behind agentic execution.

Buyer Intent is expanding significantly. HubSpot’s Buyer Intent tools can identify companies that meet configured intent signals and help teams determine which accounts may deserve attention. HubSpot can use website behavior such as visits and page views as well as broader company and research signals. That information can then be filtered, reviewed, added to CRM records, or used in workflows.

The Fall Spotlight roadmap adds more signal types, including company-change signals, custom Buyer Intent signals, churn-risk indicators, job changes, and Bombora Company Surge signals. HubSpot also introduces automatic tracking for custom signals.

That gives teams more ways to answer a fundamental revenue question: What is changing around this customer or prospect that should cause us to act?

Data Agent tackles a related problem from inside the CRM. HubSpot positions Data Agent CRM research around asking a question of CRM data without first having to determine which individual properties or activities to inspect. Smart properties can use Data Agent to generate structured insights for use cases such as pre-call research, ICP enrichment, and lead qualification.

Then there is Revenue Agent, which extends the agentic model further down the customer lifecycle. HubSpot’s Fall Spotlight positioning focuses Revenue Agent on personalized outreach for open invoices, adapting that outreach based on replies. That matters because the broader platform story does not stop when marketing generates a lead, sales closes a deal, or service resolves a ticket.

The underlying model is lifecycle-wide. Signals inform action. CRM information provides context. Agents perform defined jobs. Outcomes create more information for the system. That is what begins to turn a collection of AI features into a connected operating model.

What does HubSpot’s agentic customer platform actually mean?

In practical terms, HubSpot’s agentic customer platform means AI is becoming embedded in the systems that marketing, sales, service, and revenue teams already use to manage customer relationships.

The important shift is from AI as a standalone assistant to AI as part of a workflow. A marketer may use AI within a connected campaign process, a seller may use an agent to research a buyer before engagement and another AI capability to identify the next step once a deal is active, a service organization may allow Customer Agent to resolve defined conversations, hand others to humans, and use the results to improve its knowledge, a revenue team may use buyer signals and CRM intelligence to determine when something has changed and trigger the appropriate response, and because these capabilities increasingly share CRM data and business context, the value of one part of the system can support another.

That is the larger UNBOUND 26 story.

What should HubSpot teams do first?

The temptation after a release this large is to begin with the products, a better starting point is the process.

Identify one important customer or revenue workflow where your team currently spends too much time researching, creating, routing, updating, following up, or determining what should happen next. Then examine the foundation behind it.

Is the CRM data reliable enough for AI to use? Is HubSpot’s business context accurate? Are the relevant customer conversations or knowledge sources available? Are the boundaries between automated action and human review clear?

From there, choose the agent or capability that fits the job. That might mean starting with Agent Hub and agentic workflows around a repeatable internal process. It might mean using Prospecting Agent around a defined outbound motion. It could mean introducing Customer Agent for a specific class of support questions. Or it might mean improving Buyer Intent signals before introducing more automation downstream.

Starting with the operating problem rather than the product also creates a much cleaner way to measure whether AI is actually helping.

The bigger takeaway from HubSpot UNBOUND 26

The most important thing about UNBOUND 26 is not any one agent, it is that HubSpot is connecting more parts of the customer lifecycle to the same AI, CRM data, context, signals, and workflow infrastructure.

Agent Hub and Context Home establish the foundation. Marketing Studio, Content Agent, and Nurture Agent bring that model into marketing. Prospecting Agent and Deal Progression extend it through sales. Customer Agent applies it to service and customer experience. Buyer Intent, Data Agent, Revenue Agent, Smart CRM, and Data Hub help provide the information and intelligence those systems need to act.

That does not mean every business should deploy every agent. It means HubSpot teams now have a much broader opportunity to decide where work should remain human-led, where AI should assist, and where an agent can responsibly execute a defined part of the process. Because the future of AI in HubSpot is becoming less about asking an assistant to help you with a task and more about building a customer platform in which people, data, automation, and agents work as one connected system.