HubSpot’s July 2026 product updates show a clear direction of travel. AI is becoming more practical, more closely connected to CRM data and easier to build into everyday processes across marketing, sales and customer service.
But the value of these updates is not simply that HubSpot can do more. Used properly, they can help businesses identify opportunities earlier, move deals forward more consistently, respond to customers faster and maintain a healthier CRM.
We have selected ten of the most commercially relevant updates and explained what they could mean for your business.
Agent Hub gives businesses one central place to view, create and manage the AI agents operating across their HubSpot portal; agents are organised around the outcomes they support, including building demand, winning deals, helping customers and scaling operations.
Teams can also use Agent Hub to build new agents and connect the work happening between them.
As businesses begin using more AI tools, oversight becomes increasingly important. Without a clear structure, agents can quickly become disconnected from wider processes, duplicate work or operate without obvious ownership.
Agent Hub gives teams greater visibility over where AI is being used and what each agent is responsible for. This makes it easier to develop a coordinated AI strategy instead of introducing individual tools without considering how they work together.
A business might use one agent to research target accounts, another to prepare sales follow-ups and another to respond to common customer questions.
Agent Hub provides a central place to manage those agents around shared commercial outcomes. This can help the business create a more joined-up process from initial prospecting through to ongoing customer support.
The opportunity is significant, but governance still matters. Businesses need clear rules covering what agents can do independently, when human review is required and how performance will be measured.
Agent Builder allows businesses to create custom AI agents using their own instructions, tools, knowledge and CRM data.
Its new automation canvas connects agents with actions and triggers. Breeze Assistant can also help configure an agent from a plain-language description, making the building process more accessible to teams without specialist technical knowledge.
These agents can then be connected to workflows to analyse information, update records, follow up with contacts and support processes at scale.
Generic AI tools can provide useful answers, but they rarely understand how a particular business operates. Custom agents can be designed around your terminology, processes, qualification criteria and customer data.
This creates an opportunity to automate more complex work that traditionally required several workflow branches or regular manual intervention.
The strongest applications will not start with the question, “Where can we use AI?” They will start with a defined operational problem, a reliable source of data and a measurable outcome.
HubSpot’s Prospecting Agent is developing into a more capable outbound sales tool.
It can identify companies displaying relevant buying signals, including growth, recruitment activity, funding and intent. It can then source suitable contacts through CRM data or third-party integrations, generate personalised outreach and enrol contacts into email and calling activity.
Many sales teams spend too much time building lists and too little time speaking to the right prospects.
Buying signals can help teams focus their attention on businesses that may have a genuine reason to act. Contact sourcing then reduces the manual research required to find relevant decision-makers.
The real value is not simply higher activity. It is the potential to make prospecting more timely and relevant.
Customer Agent can now do more than answer questions. It can qualify prospects, create or update CRM records and book meetings with salespeople.
The new Actions Library also gives businesses access to a searchable catalogue of pre-built HubSpot CRM, workflow and Shopify actions. This allows Customer Agent to complete a wider range of tasks during a conversation.
A website enquiry does not always arrive during office hours. Even when a team is available, slow routing or inconsistent qualification can cause good opportunities to be missed.
Customer Agent can provide an immediate first response, gather the information sales teams need and offer an appropriate next step. That could reduce response times and make it easier for suitable prospects to move from an initial question to a booked conversation.
A visitor arrives on a software company’s website and asks about pricing.
Rather than providing a generic response, Customer Agent could ask about the visitor’s company size, requirements and expected timescale. If the answers meet the company’s qualification criteria, it could create the contact in HubSpot and book a meeting with the correct salesperson.
If the visitor is not yet ready to speak with sales, the agent could direct them towards relevant content or another appropriate route.
The experience should still feel helpful rather than interrogative. Qualification questions need to be proportionate to the visitor’s request and designed around a clear customer journey.
HubSpot is introducing several connected features designed to help salespeople understand what to do next on active deals.
These include:
Deals often stall because the next action is unclear, follow-ups are missed or important information remains buried in emails and meeting notes.
These updates bring more of that context into the salesperson’s working view. Reps can see recommended actions and prepare informed replies without repeatedly opening records, searching through activity histories or manually updating multiple properties.
For sales leaders, more structured methodology data can also make pipeline reviews more reliable.
HubSpot has expanded the information and actions available through its connector for Claude and applications built using HubSpot’s remote Model Context Protocol server.
Depending on the platform and account access, external AI tools can now work with a broader range of HubSpot data and processes, including:
HubSpot has also introduced verified domain restrictions to give Claude Enterprise administrators greater control over account access.
This update makes it easier for businesses to use HubSpot as the reliable source of customer context behind external AI tools.
A user could ask an AI assistant to analyse campaign performance, prepare an email draft or review active sales opportunities without manually exporting and reorganising the underlying data first.
That can reduce the gap between insight and action. It also reinforces the importance of permissions, data structure and governance, because connected AI systems can only work effectively when they receive accurate information and appropriate access.
7. Describe a custom app and deploy it to your HubSpot portal
HubSpot is testing a conversational, no-code tool that allows users to describe a custom application in plain language and deploy it directly within their portal.
The app can use the objects, properties and permissions already available to the user. It is designed for requirements that standard views, reports and dashboards cannot address clearly.
Businesses often adapt their processes to fit standard CRM screens, even when those screens do not give teams the information they need in the right format.
This update could make tailored operational tools much more accessible. Instead of commissioning a separate application for every specialist requirement, teams may be able to create focused HubSpot interfaces around a particular role or decision.
The new Data Model Health Check reviews a HubSpot account’s properties, objects, associations, pipelines and lifecycle stages.
It identifies potential issues and recommends actions that administrators can take directly from the data model overview.
CRM problems rarely begin with one major failure. They tend to develop gradually through duplicated properties, inconsistent stages, weak associations and processes that no longer reflect how the business operates.
These issues reduce reporting confidence, complicate automation and limit the value businesses can gain from AI. A health check gives administrators a clearer starting point for identifying and addressing structural problems.
A business may have several properties capturing slightly different versions of “customer type”. Different teams use different fields, making segmentation and reporting unreliable.
The health check could help surface that issue so the business can review how the data is collected, decide which property should be retained and update the processes that rely on it.
Recommendations should not be applied without context. A property that appears unused may still support an integration, workflow or historical report. CRM changes need proper impact assessment and controlled implementation.
Users can now filter records on CRM index pages using properties from associated objects.
For example, they can filter:
Customer information rarely sits within one object. A contact’s relevance may depend on the deals, company, tickets or other records associated with them.
Cross-object filtering allows teams to create more commercially useful views without relying on separate reports, exports or overly complicated workarounds. It helps users work with the relationship between records, rather than treating each object in isolation.
HubSpot has expanded and redesigned the citation table within its AEO Citations tab.
The full-screen table now includes:
These additions provide a clearer view of the content being referenced in AI-generated answers and whether the brand itself appears alongside the citation.
Being cited by an AI platform and being mentioned by name are not the same thing.
A business’s content might inform an answer without the brand receiving visible recognition. Alternatively, a business may be mentioned but linked to a page that does not support the next stage of the buyer’s journey.
The updated table can help marketers understand which content is earning visibility, where brand recognition is weaker and which subject areas may deserve further investment.
July’s releases bring AI further into the practical work of marketing, sales and service teams.
The biggest opportunity is not simply to automate more tasks. It is to create better-connected processes in which teams have the right information, clearer next steps and more time to focus on work that requires human judgement.
However, the foundations remain the same. Businesses will need:
AI will not correct an unclear process or an unhealthy CRM by itself. In many cases, it will expose those weaknesses more quickly.
The businesses that gain the most from these updates will be those that choose focused use cases, introduce them thoughtfully and measure whether they improve speed, efficiency, customer experience or revenue performance.
If you want help turning these updates into a real-world improvements inside your HubSpot portal, we're always here to talk through what's possible.