Blog | Unlimited Tech Solutions

HubSpot UNBOUND 2026: 8 Updates and What to Do Before You Turn Them On

Written by Thomas Richard | Sep 24, 2026, 12:40:59 PM

HubSpot's UNBOUND 2026 announcements could change how much work your CRM actually does for you.

 

The platform is moving beyond storing customer data and triggering predefined automation. HubSpot is building a CRM that can keep itself more up to date, understand more about your business, recommend what should happen next and use AI agents to take action across marketing, sales, service and revenue.

 

That could mean less manual CRM work, faster campaign execution, better sales follow-up and more personalized customer experiences.

 

But there is a catch: AI can scale a good process. It can also scale a bad one.

 

Duplicate records, unclear lifecycle stages, outdated workflows and disconnected systems don't disappear because an AI agent is involved.

 

So rather than walk through every Fall Spotlight announcement, we're focusing on eight updates that could have the biggest impact on how teams actually use HubSpot, plus what you should check before relying on them.

 

Before UNBOUND, we shared four things we were watching across AI, AEO, data and CPQ. Here's what stood out.

 

The short version

Update

Best for

First step

Self-Updating CRM

Everyone

Agree on lifecycle stage definitions

Breeze Assistant

Everyone

Test five questions your team asks every week

Marketing Studio

Marketing

Review brand, audience and campaign context

Agent Builder

RevOps

Audit existing workflows

Prospecting Agent + Deal Progression

Sales

Define your ICP and clean up your pipeline

Revenue Agent

Finance + RevOps

Confirm invoice and payment data reaches HubSpot

Customer Agent

Service

Clean up your knowledge base

ChatGPT Ads

Marketing

Fix attribution on your current channels

 

1. The Self-Updating CRM

HubSpot's Self-Updating Smart CRM is designed to capture more activity from calls, emails, meetings and other customer interactions automatically.

 

That information contributes to Growth Context, HubSpot's understanding of your customers, ICP, positioning, business and processes. The new Context Home gives teams a place to review that information, correct what's wrong and fill in what's missing.

 

The potential benefit is: less time spent manually keeping the CRM current and better context for the AI tools working from that data.

 

But automatic updates don't replace the decisions your company still needs to make about what the data actually means.

 

Before you rely on it:

  • Agree on lifecycle stage definitions across marketing and sales.
  • Merge duplicate contacts and companies.
  • Review your ICP, positioning and important business context.
  • Check that the properties driving your automation are consistently populated.

It can capture activity. It can't decide what an MQL means for your company.

 

2. Breeze Assistant, rebuilt

 

 

 

For example:

  • Which deals haven't moved in 30 days?
  • Which opportunities should close this month?
  • Which customers have high-priority open tickets?
  • Which campaigns influenced the pipeline last quarter?

If Breeze gives you a wrong answer, don't immediately dismiss the tool. Find out why.

 

A wrong answer may expose a missing association, inconsistent property, incorrect lifecycle stage or incomplete deal activity that was already affecting your reporting.

 

That's what makes Breeze one of the easier UNBOUND updates to start experimenting with now.

3. Marketing Studio

Marketing Studio brings more of campaign planning, creation, automation and optimization into one workspace.

 

HubSpot is pairing it with specialized agents including Campaign Agent for campaign planning, Content Agent for assets such as blogs and landing pages, and Nurture Agent for more personalized nurture emails.

 

The bigger opportunity isn't simply creating content faster. It's shortening the path from idea, campaign plan, content, nurture, execution.

 

That could be especially useful for small marketing teams managing multiple channels without adding more disconnected tools.

 

Before putting more of that process in the hands of AI, make sure:

  • HubSpot has accurate brand and audience context.
  • Your segmentation reflects the customers you actually want to reach.
  • Someone is responsible for reviewing AI-generated messaging.

Marketing Studio and its agents can speed things up, but Campaign Agent still needs to know who you're targeting, Content Agent still needs to understand your brand and Nurture Agent still needs reliable customer data.

 

4. Agent Builder

Agent Builder lets teams create AI agents around their own business processes instead of relying only on agents HubSpot has already built.

 

That opens up a lot of practical use cases. You could build an agent to summarize why deals were lost, check opportunities for missing information before a handoff, identify stalled deals or prepare a weekly pipeline summary.

 

But the first question shouldn't be, What can we automate?

 

It should be, Is this process worth automating in its current form?

 

Before building an agent, look at the process behind it:

  • Who owns it?
  • What starts it?
  • Are the criteria still correct?
  • What should happen when something goes wrong?
  • Which decisions still need a person?

A pre-handoff check for missing deal information or a weekly stalled-deal summary could save meaningful time without handing an entire process over to AI.

 

For more examples, see 7 revenue processes worth automating with Agent Builder.

 

5. Prospecting Agent and Deal Progression

HubSpot is also giving sales teams more AI support before and after an opportunity enters the pipeline.

 

 

The catch is that both depend heavily on how clearly you've defined what a good opportunity looks like.

 

Instead of an ICP that says: Mid-market companies that are a good fit.

 

Give HubSpot criteria it can actually evaluate, such as location, revenue range, industry, systems used or company type.

 

That means removing stages nobody uses, defining what must happen before a deal moves forward and checking that calls and emails are associated with the correct opportunities.

 

Good sales AI should help answer two questions: Why this account? Why now?

6. Revenue Agent

HubSpot's Revenue Agent brings AI into collections.

 

It can monitor open invoices, create personalized follow-up plans and adjust outreach as customers respond or invoice activity changes.

 

This matters because HubSpot continues to stretch beyond the traditional moment when a deal becomes Closed Won. Quotes, contracts, billing, payments, collections and renewals are becoming more connected to the same customer record.

 

But for many companies, the financial data Revenue Agent needs lives somewhere else.

If your ERP is the source of truth, answer a few questions before turning the agent loose:

  • Does invoice information reach HubSpot?
  • Does payment status sync back?
  • How frequently does that happen?
  • When should automated outreach stop and a person take over?

A nightly sync, for example, creates a window where a customer could pay in the morning but HubSpot doesn't know it yet.

 

A collections agent can only act on the information it can see.

 

If NetSuite is your billing source of truth, the quality and frequency of the NetSuite-to-HubSpot connection becomes part of your AI strategy.

 

Still deciding how quoting should fit into HubSpot? Start with Three Paths to Quoting, or watch our Revenue Hub podcast episode.

 

7. Customer Agent

Customer Agent continues to expand across customer service channels.

 

It can respond to questions using CRM and support context across channels such as chat, email, voice and social. When a conversation needs a person, the goal is to carry that context into the handoff instead of forcing the customer to start over.

 

Archive outdated knowledge base content, identify common questions you haven't documented and decide which types of requests should always go directly to a person.

 

One simple exercise is to review your 20 most common ticket types and separate them into:

  • AI can handle this.
  • AI can start this, but a person should review.
  • A person should handle this from the beginning.

Ticket deflection isn't the goal. Better resolution is.

 

Let AI handle the repeatable questions and give your team better context for everything else.

 

8. ChatGPT Ads

 

 

Before spending there, make sure you can already answer questions such as:

  • Which campaigns are generating pipeline?
  • Which channels are generating closed revenue?
  • Are UTMs consistently applied?
  • Are Original Source and attribution reporting reliable?

Then ask your newest customers something even simpler:

 

Where did you research vendors before contacting us?

 

Their answers may tell you more about whether ChatGPT belongs in your media plan than another prediction about where advertising is heading.

 

If you can't attribute LinkedIn spend today, you won't magically attribute ChatGPT spend tomorrow.

 

Fix the measurement first.

Your first 30 days

You don't need to activate every new UNBOUND feature.

 

Start with the foundation and one meaningful use case.

 

Week

Focus

Done when

1

Definitions

Marketing and sales agree on lifecycle stages and deal-stage exit criteria

2

Cleanup

Duplicates are addressed, unnecessary pipeline stages are removed and questionable workflows are flagged

3

Context

Your business, ICP and brand context are accurate, and Breeze passes your basic question test

4

First agent

One agent is running on one repetitive task with a named owner and review process

 

Revenue Agent and ChatGPT Ads can come later if the ERP connection, financial data and attribution needed to support them are ready.

 

Is your HubSpot portal ready for AI?

Before adding more agents, ask yourself:

  1. Do marketing and sales use the same lifecycle stage definitions?
  2. Does your pipeline match how your team actually sells?
  3. Are the properties driving your most important processes consistently populated?
  4. Do integrations sync without creating duplicate or conflicting data?
  5. Does the data you need from systems like your ERP reliably reach HubSpot?
  6. Are permissions and guardrails ready for tools that can take action?

If most of those answers are yes, you have a good foundation for a controlled AI pilot.

If several are no, fix the gaps that directly affect the feature you want to use first.

 

The biggest takeaway from UNBOUND 2026

The interesting part of HubSpot's UNBOUND announcements isn't simply that the platform has more AI.

 

It's that HubSpot is increasingly connecting: CRM data, business context, AI agents, automation

That changes what the CRM can do.

 

But it doesn't eliminate the need for good CRM architecture.

 

The teams that get the most value from these updates won't necessarily be the ones that turn on the most features first.

 

They'll be the ones that know what their data means, have processes their teams actually follow and choose the right places for AI to take work off their plate.

 

Contact us to continue to conversation