HubSpot INBOUND Is Now UNBOUND: What the Agentic Customer Platform Means for Marketers

TL;DR: HubSpot buried the 15-year-old INBOUND brand on September 16–18, 2026, replacing it with UNBOUND — a conference name aligned with its new identity as an Agentic Customer Platform. Five products went generally available at the event: Context Home, a rebuilt Breeze Assistant (available on free plans), Agent Hub, Marketing Studio 2.0, and Deal Progression Agent. Four more tools remain in beta. If you use HubSpot, your workflows are about to look fundamentally different.
After 15 years, INBOUND is gone. What replaced it on September 16–18, 2026 in Boston was UNBOUND — a conference that drew more than 13,000 attendees and came with a clear message: HubSpot is no longer a marketing software company, it is an AI-first customer platform. The rebrand was announced in April 2026, but UNBOUND was where the product commitments behind that name became real and reviewable.
This is not a cosmetic pivot. The platform HubSpot unveiled at UNBOUND has a fundamentally different architecture from the one marketers have been using for the last decade. Understanding what launched, what is still being built, and what it means for your stack is the practical work this article covers. The short version: some of what is now GA will save your team hours per week immediately; the beta tools indicate where HubSpot’s roadmap is headed over the next 12 months.
Why HubSpot killed INBOUND after 15 years
INBOUND was built around a methodology: attract, engage, delight. That idea was revolutionary when it launched in 2009. By 2026, HubSpot had grown from a marketing tool into a platform covering sales, service, operations, and revenue management for more than 200,000 customers. The INBOUND name, rooted in the web-native era of content and organic search, no longer captured what the company does or where it is going.
The strategic reframe is significant. HubSpot is now describing itself as an Agentic Customer Platform rather than a CRM. That is a different claim than “AI-powered CRM.” An agentic platform implies that the software takes autonomous actions — queuing tasks, drafting sequences, processing invoices, following up on deals — rather than waiting for humans to initiate every operation. Whether HubSpot’s current AI capabilities match that ambition fully is debatable, but the architectural commitments the company made at UNBOUND point in a clear direction.
The three-layer architecture HubSpot announced
The new product structure has three named layers:
- Context Layer — Smart CRM: The shared data store. Every AI action draws on customer history, ICPs, call transcripts, and methodology data curated here. Context Home is the dashboard that lets revenue operations teams review and manage what the AI knows about the business.
- Action Layer — Hubs + AI Agents: The working surface. Marketing Hub, Sales Hub, Service Hub, and Operations Hub sit here, now joined by Breeze Assistant and the new Agent Hub where you deploy and govern autonomous agents.
- Coordination Layer — Human-AI Workspace: The oversight interface. This is where teams see what agents are doing, approve or override actions, and manage permissions. HubSpot Work (in private beta) is the first product designed specifically for this layer.
The architecture borrows from how enterprises think about AI governance: you need a place where context lives, a place where work gets done, and a place where humans stay in control. Getting all three into a single product is genuinely hard, and HubSpot is further along than most CRM vendors on the first two.
What went generally available at UNBOUND
Context Home (all plans)
Context Home is a RevOps dashboard that shows the AI what your business believes about its customers: ideal customer profiles, documented sales methodology, anonymized call transcripts, and historical performance data. The logic is that agents work better when they have reliable context, and that context needs to be curated by humans rather than auto-inferred from raw CRM records. Context Home is the interface for that curation job. It is available on all plans, including free accounts.
Breeze Assistant rebuilt with MCP connectors (all plans including free)
The original Breeze Assistant was essentially a chatbot layer on top of HubSpot data. The rebuilt version is meaningfully different. It is now role-aware — a sales rep and a marketing manager see different capabilities — and it connects to external data through Model Context Protocol (MCP) connectors, the same standard that major AI platforms including Anthropic and OpenAI have adopted for tool interoperability. Mobile voice commands are also now supported, making it the first HubSpot AI tool you can interact with hands-free.
The MCP integration matters because it means Breeze Assistant can read and write data outside HubSpot — in Google Sheets, Slack, or third-party databases — without requiring a custom integration build. That changes its scope from an in-app helper to something closer to a general business assistant anchored in CRM context. We covered a related capability shift in our look at agentic AI moving into CMS content workflows; HubSpot is applying the same pattern to the full customer operations stack.
Agent Hub + Agent Builder (Professional and Enterprise)
Agent Hub is the governance dashboard for AI agents: a single view of every deployed agent, what it is doing, what permissions it holds, and when it last acted. Agent Builder is the no-code creation interface that lets users define an agent’s goal in plain language and choose its triggers — a Slack message, a new Google Sheets row, a webhook from an external system.
This is the tooling that makes HubSpot’s “agentic” claim operational rather than aspirational. Before Agent Hub, deploying AI agents in HubSpot required either custom API work or acceptance of whatever Breeze did automatically. Now there is a surface where non-technical marketers can build workflows that were previously only possible for developers. The constraint is the plan tier: Agent Hub and Builder require Professional or Enterprise.
Marketing Studio 2.0 (Marketing Hub Professional+)
Marketing Studio 2.0 consolidates campaign creation, content scheduling, and performance reporting into a single workspace. The AI layer here handles initial asset drafts, variant generation for A/B testing, and flagging underperforming content against benchmarks pulled from the Context Layer. It replaces the scattered campaign management experience that grew organically across HubSpot over the last several years with a purpose-built interface.
Deal Progression Agent (Sales Hub and Service Hub Professional+)
Deal Progression Agent monitors active deals and initiates follow-up actions — drafting outreach emails, scheduling reminders, updating deal stage records — based on inactivity thresholds and deal health signals. It is the product most directly responsible for HubSpot’s cited statistic that teams using agentic workflows see 2.2x more leads and 3.4x more closed deals compared to those using disconnected tools. Third-party data from McKinsey cited at UNBOUND put the average sales cycle reduction at 30% for organizations with centralized AI orchestration.
What is still in beta
Four products were announced or promoted at UNBOUND but are not yet widely available:
| Tool | Status | Credits / Pricing Signal | Purpose |
|---|---|---|---|
| HubSpot Work | Private beta | N/A | AI-first task management workspace inside the CRM, replacing external tools like ClickUp or Notion for CRM-linked projects |
| Campaign Agent | Public beta | 500 credits per plan | Builds full campaigns from a goal description or a call transcript |
| Content Agent | Public beta | 1,000 credits per piece | AEO-focused long-form content creation anchored in CRM context and keyword data |
| Revenue Agent | Public beta | 500 credits per invoice | Accounts receivable automation: sends reminders, tracks payment status, escalates overdue accounts |
Nurture Agent is also in testing at 10 credits per email, handling automated lead nurturing sequences that adapt based on contact engagement signals. The credits-based pricing model is a notable design decision: it lets HubSpot charge incrementally for AI usage rather than bundling AI into flat-rate plan tiers, which aligns with how most enterprise AI tools are moving. Compare this to what we covered in HubSpot’s 2026-09 API, where the company tightened validation rules on CRM writes — another signal that they are hardening the platform for heavier programmatic usage.
New platform integrations announced at UNBOUND
Two integrations drew particular attention:
- ChatGPT Ads CRM integration: HubSpot contacts and deal data can now feed directly into ChatGPT Ads targeting, and ChatGPT Ads lead capture writes back into HubSpot. Given that AI search advertising is growing fast, this closes a loop that marketers previously had to bridge with manual exports or third-party middleware.
- Microsoft Advertising integration: Native sync of HubSpot audience lists and conversion events into Microsoft Advertising, including Copilot campaign signals. This is more significant for B2B marketers than consumer brands, given Microsoft’s stronger presence in enterprise buying journeys.
What this means for your marketing stack
The practical question coming out of UNBOUND is not “should I use HubSpot’s new AI tools” — it is “is my surrounding infrastructure ready for agentic workflows?” Agents that take autonomous actions depend on clean data, reliable webhook endpoints, and web infrastructure that exposes the signals the AI needs. A deal progression agent that sends follow-up emails needs accurate contact records. A campaign agent that builds from a goal needs a web presence structured to convert the traffic that campaign generates.
Teams that have cobbled together their web presence across legacy systems will find that the limiting factor for agentic marketing is not HubSpot’s capabilities — it is the fragmented infrastructure the agents are connecting to. Companies that built or rebuilt their digital stack with clean APIs and structured data are seeing this play out differently. If your current setup is a constraint, working with a custom software development team to modernize the underlying architecture before layering agentic tools on top is worth prioritizing. Infrastructure debt is the hidden ceiling on AI adoption.
Should you migrate to Agent Hub now?
For most teams on Professional or Enterprise plans, yes — but with caveats. Agent Hub is GA, which means it is stable and supported. However, the no-code agent builder currently handles a limited set of triggers and actions. If your workflows require complex branching logic or high-volume API calls, you will likely hit the builder’s limits quickly and need to fall back to the API or Zapier for more sophisticated automations.
The right sequence for most teams:
- Set up Context Home and audit what the AI currently knows about your business. Fix obvious gaps in ICP documentation and sales methodology records before deploying agents that will act on them.
- Enable the rebuilt Breeze Assistant and test its MCP connectors with one external data source — Google Sheets is the lowest-friction starting point.
- Deploy one Deal Progression Agent on a defined deal stage with a low-risk trigger (for example, no activity after 14 days). Measure the outcome for 30 days before expanding scope.
- Request Campaign Agent and Content Agent beta access if you are on Professional. These are the tools most likely to compress campaign production timelines in Q4 2026.
The HubSpot UI redesign that went mandatory on August 31 was a prerequisite for this architecture. If your team has been on the new interface for several weeks, the cognitive transition to Agent Hub will be smaller than it looks from the outside.
FAQ: HubSpot UNBOUND 2026
What is HubSpot UNBOUND 2026?
UNBOUND 2026 is HubSpot’s annual conference, formerly known as INBOUND, held September 16–18, 2026 in Boston with more than 13,000 attendees. The rename signals HubSpot’s shift from an inbound marketing company to an Agentic Customer Platform covering marketing, sales, service, and operations.
What is HubSpot’s Agentic Customer Platform?
It is HubSpot’s new product architecture with three layers: a Context Layer (Smart CRM with curated customer data), an Action Layer (Hubs plus AI agents and Breeze Assistant), and a Coordination Layer (human-AI workspace with unified governance and oversight).
Which HubSpot features went GA at UNBOUND 2026?
Five products went generally available: Context Home (all plans), Breeze Assistant with MCP connectors and mobile voice commands (all plans including Free), Agent Hub plus Agent Builder (Professional and Enterprise), Marketing Studio 2.0 (Marketing Hub Pro+), and Deal Progression Agent (Sales and Service Hub Pro+).
What HubSpot features are still in beta after UNBOUND 2026?
Four products remain in beta: HubSpot Work (private beta — AI-first task workspace), Campaign Agent (public beta, 500 credits), Content Agent (public beta, 1,000 credits per piece), and Revenue Agent (public beta, 500 credits per invoice). Nurture Agent is also available at 10 credits per email.
Do I need to upgrade my HubSpot plan to use the new AI agents?
Context Home and the rebuilt Breeze Assistant are available on all plans including free. Agent Hub, Agent Builder, Marketing Studio 2.0, and Deal Progression Agent require Professional or Enterprise. Campaign Agent and Content Agent credits are allocated by plan level. Check HubSpot’s official UNBOUND announcements for the current plan availability matrix as beta tools move to GA.