If your business runs on HubSpot, the CRM you log into next month may feel quite different. At its UNBOUND 2026 conference in Boston, held from 16 to 18 September, HubSpot unveiled its Fall 2026 Spotlight release and repositioned itself as an “agentic customer platform”. In plain English: AI agents that plan campaigns, write content, nurture leads, chase buying signals and even follow up unpaid invoices, all working from the data in your CRM.
Here’s a practical rundown of what was announced, and what a small or mid-sized business should do before switching anything on.
The big idea: “Growth Context”
The thread running through HubSpot’s announcements is something it calls Growth Context: the idea that AI agents only work when they understand your business, and that this understanding should be visible and editable inside HubSpot.
The practical expression of this is Context Home, a place where you can view and correct what HubSpot “knows” about your business, customers and processes. Coverage describes it as scoring how complete your data foundation is, and Penguin Strategies reports that it’s generally available on all plans.
Alongside it, HubSpot’s Smart CRM now automatically captures emails, calls and meetings, reducing the manual data entry that leaves so many CRMs half-empty.
The agents and assistants announced
- Breeze Assistant has been rebuilt to take a request and assign work across specialised agents, returning deliverables such as campaign plans and proposals. It also gains Projects (for ongoing context), voice on mobile and a Slack integration.
- Marketing Studio brings campaign planning, content creation and performance insights into one workspace, with three new marketing agents: a Campaign Agent, a Content Agent and a Nurture Agent. Coverage differs on its release status, with at least one report describing Marketing Studio as a public beta.
- Prospecting Agent monitors more than 40 buying signals, assembles buying groups and drafts personalised outreach.
- Revenue Agent automates collections with personalised invoice follow-ups.
- Customer Agent resolves customer enquiries across chat, email, voice and social, with handoff to a person when needed.
- Agent Hub and Agent Builder, available on Professional and Enterprise plans, with custom event triggers for building your own agents.
HubSpot also added advertising integrations, including what diginomica reports as the first CRM integration with ChatGPT Ads, plus Microsoft Advertising and Reddit Ads.
What it will cost
According to Penguin Strategies, many core features, including Context Home, Breeze Assistant and Agent Hub, are included without credits. Specialised agents are metered, consuming credits on a per-action basis, with portal-level controls and a simulation mode to estimate costs. Hacking Demand’s recap noted that full pricing details weren’t announced and advised watching credit consumption as you enable new agents. That’s good advice.
Our take: fix the data before you add the agents
HubSpot’s own framing makes the key point: agents are only as good as the context they work from. If your CRM has duplicate contacts, missing lifecycle stages, inconsistent deal pipelines or no record of where leads came from, an AI agent will simply automate the mess faster.
HubSpot has published some eye-catching multiples for customers using AI with high-quality context, such as several times more MQLs and deals. These are HubSpot’s own figures, they vary across its materials, and they likely reflect businesses that already had strong data and processes. Treat them as a sign of what’s possible, not a prediction for your account.
It’s also worth noting that HubSpot’s Spotlight page marks many of the new features as public or private beta, so not everything will be available in every account straight away.
What to do now
- Start with Context Home. Review what HubSpot thinks it knows about your business and correct anything wrong. It’s included, and it improves everything else.
- Clean your CRM. Merge duplicates, standardise lifecycle stages and deal stages, and make sure lead source tracking works.
- Pick one agent and one process. For example, use the Nurture Agent on a single follow-up sequence, or the Revenue Agent on overdue invoices. Measure the result before adding more.
- Set credit limits. Use the portal-level controls and simulation mode to estimate and cap spend before enabling metered agents widely.
- Keep a human in the loop. Review AI-written emails and campaign content, at least at first, to protect your brand voice and avoid errors.
- Connect ads to the CRM. If you run Google, Microsoft or other ads, make sure lead and revenue data flows back so you can see which campaigns produce real customers.
AI agents in the CRM can take a lot of repetitive work off a small team’s plate, but only if the foundations are right. If you’d like help cleaning up your CRM or setting up your first HubSpot automations, the CRM automation team at Big Digital would be glad to help.
Photo: KOBU Agency on Unsplash