Skip to content
unbound-2026-people-joining-event
power2Cloud28/09/2615 min read

HubSpot UNBOUND 2026: Why Context Now Matters More Than AI Tools

"Raise your hand if you've read something this week and instantly known AI wrote it".

Almost every hand in the room went up. Ours did too.

It was a small moment, but it set the tone for UNBOUND (formerly INBOUND) 2026. This wasn't going to be a product launch with some networking on the side. It was a conversation about how all of us actually work now.

Our team spent three packed days in Boston, working through more than 200 sessions, workshops and live demos.

You've probably already seen the big announcements, but the livestream only shows you so much. We spent most of our time in hands-on labs, testing new features and swapping notes with HubSpot's teams and partners from around the world.

Once we pulled together our notes, the testing and all those conversations, one thing stood out. Everyone has AI tools now, so simply having them isn't an advantage any more. The advantage comes from the context you give them: your data, and how well they fit into the rest of your stack.

That's the lens we've used for this recap. Rather than walking through every new feature, we'll focus on what actually changes results: context, data, and how you keep control of AI agents that are starting to work both inside and outside your CRM.

 

Why We Went All the Way to Boston for UNBOUND

We don't want to read about HubSpot's latest features in a release note and then work out how to roll them out for clients. We want to see them first. That's why we make the trip to HubSpot's flagship conference every year, and why our HubSpot Trainers and Community Champions keep signing up for certifications and bootcamps.

Three days at UNBOUND gets you new ideas, people from very different business cultures, use cases you'd never have thought of, and beta features you can try before anyone's writing about them.

The schedule is packed, so we split up. One of us takes a workshop, another covers the keynote, a third gets hands-on with a demo. By the evening, photos, notes and slides are already landing in our shared Drive folders for the team back home.

Some of the best conversations of the week happened over pepperoni pizza. We got chatting to partners from Australia and Colombia, compared notes on the projects we're running, and quickly realised we're all fighting the same battles: messy data migrations, getting sales teams to actually use the CRM, AI workflows, the lot.

That's what we like most about the HubSpot community: the people building on the platform and the people using it every day end up in the same room. In one session, a marketing manager from Boston told us she'd spent years on Salesforce before making the switch. Her verdict? She gets more done now, and spends a lot less time wrestling with her software.

It's a good reminder of why we go. New features are exciting, but what matters is whether they make someone's working day easier. That's the test we'll be applying to everything we brought home from Boston.

 

Inbound Isn’t Dead, It Just Measures Different Things

If one theme ran through UNBOUND 2026, it was outcomes.

Inbound marketing was built on a simple idea: useful content beats constant interruption. That still holds. What's changed is how people buy.

It used to be predictable. Someone searched on Google, landed on your website, read an article, filled in a form and became a lead.

Now buyers do most of their homework somewhere else. They ask ChatGPT or Gemini to explain the problem, then compare solutions and shortlist products long before they visit your site. By the time they reach you, they've often half made up their mind.

So a form fill isn't the start of the journey any more, and counting them tells you less than it used to. What matters is what happens at the end: closed deals, revenue, resolved tickets and qualified meetings.

HubSpot has rebuilt its user experience around that shift. Instead of starting with a tool, you start with the result you want, then set up the features and AI agents that will get you there.

 

AI’s Real Problem Isn’t the Model, It’s What It Knows About You

George Davis, Director of Product at HubSpot, described a scenario most of us will recognise. A company rolls out AI and the team starts using it. Then people start rewriting its outputs, second-guessing its answers and ignoring its suggestions. Eventually someone says it: "This tool doesn't understand how we work."

Fair enough. But what did anyone actually tell it?

That's the gap HubSpot's new Growth Context is designed to close. It's built on three layers of information:

  • Company: brand guidelines, products, positioning and tone of voice
  • Team: roles, processes, methodologies and sales strategy
  • Customers: interaction history and CRM records

Then Davis asked the room how many companies still keep this information in loose documents, spreadsheets, shared folders or on people's laptops.

Almost every hand went up. Again.

That's the real problem. If your knowledge lives in people's heads or in folders nobody can navigate, your AI agents have nothing consistent to work from, so they can't behave consistently either.

Here's a quick test: what do you find yourself explaining over and over to partners and suppliers? Whatever it is, that's the context you need to get out of people's heads and into your systems.

 

Context Home Puts Your Business Knowledge Inside the CRM

This is where the idea becomes something you can actually use.

Context Home brings everything HubSpot knows about your business, team, customers and processes into one place. You can see how complete your context is, spot what's missing and fix anything that's out of date.

That matters because the rest of the platform reads from it:

  • Breeze Assistant uses it to answer questions and complete tasks in a way that reflects how your business actually works
  • AI agents use it to make better decisions inside your workflows

Alongside that, there are two new tools for managing those agents:

  • Agent Hub shows every agent running across your business in one view
  • Agent Builder lets you reuse agents that already work across multiple portals, so you're not starting from scratch each time

The same approach runs through marketing, sales and service. One agent helps build campaigns, another handles lead nurturing, and a third supports deals or picks up customer tickets.

But the real shift isn't how many agents you can switch on. It's that AI is moving from writing text to actually doing the work. And once software starts taking action on your behalf, data quality, permissions and governance matter a lot more than they used to.

 

The Agent Becomes Your CRM’s Second User

Your team isn't the only one working in the CRM any more. AI agents are in there too, querying data behind the scenes, pulling records and coordinating between tools to get tasks done.

That's what makes the idea of a "headless" CRM suddenly practical. The CRM is still your single source of truth for records, relationships, rules and processes. But the screen your team logs into is no longer the only way in. Agents can work with that data directly, without ever opening the interface.

In one workshop, HubSpot shared a framework we liked a lot: design an agent the way you'd onboard a new employee.

  • Role: what it's there to do (instructions)
  • Access: what it's allowed to see and change (permissions)
  • Knowledge: what it needs to know about your business (context)
  • Autonomy: how much it can do without checking with someone first

We saw this in a live demo of a customer health-check agent. It worked from a document that defined what healthy, at-risk and critical accounts look like. To begin with, it needed a person to approve every change before it touched a record.

That's the point. You check its work first, and as it proves itself, you give it more freedom, just as you would with a new hire.

It's a small step, but it's the difference between AI your team trusts and AI they quietly ignore.

 

Fix Your Data Model Before Adding Agents

The same workshop exposed one of the biggest traps. When an agent mapped out a portal's structure, the gaps showed up straight away: missing associations, isolated data, deals with no company attached.

This is where a lot of AI plans come unstuck. No agent can make up for messy data. Unlinked records, inconsistent properties and information that lives outside the CRM don't just cause mistakes, they cause them faster. If your data doesn't make sense to your team, AI will only spread the confusion further.

Many of the partners we met in Boston had restructured their data models before launching a single agent. It's the same order we follow at power2Cloud:

  1. Work out which data actually matters
  2. Clean up properties and associations
  3. Give every piece of data a clear owner
  4. Only then add automation and AI

It's one of the least exciting lessons from UNBOUND 2026, and probably the most important.

 

Start Every Agent on Read-Only Access

The same logic applies to security and permissions.

Just because an agent can access everything in your CRM doesn't mean it should.

Your admins need to decide who can set up AI tools, which records each agent can see and what it's allowed to do. And every action it takes should be traceable.

In our client projects, we always start in the same place: read-only access.

Let the agent work on data and processes your team knows well, so mistakes are easy to spot. Only give it permission to edit or update records once it's proved it gets things right.

Giving AI more autonomy doesn't mean giving up control. It means deciding, deliberately, where control matters most.

 

If Customers Ask AI, Does AI Mention You?

Remember those buyers asking ChatGPT and Gemini to do their research? For marketing, that's the biggest shift of all.

People haven't stopped looking for answers. They've just changed where they ask. And when someone asks an AI assistant to recommend software, the reply names specific vendors and products. The shortlist gets made before anyone clicks through to your site.

That's the problem HubSpot AEO (Answer Engine Optimization) is built to solve. It tracks how AI answer engines talk about your brand, shows how your visibility compares with competitors, and highlights the gaps in your content.

In Boston, we got to see how neatly it fits into the rest of the platform. Once it spots where you're not showing up, you can create content to fill the gap, using the brand guidelines, audience insights and tone of voice you've already set up.

(Full disclosure: we listened to this session through headphones at the HubSpot Academy stand, from the comfort of their massage chairs.)

SEO and AEO aren't rivals. Both reward content that's useful, well structured and written by people who genuinely know their subject. Which brings us back to that show of hands at the start: content that obviously sounds like AI struggles to earn trust from people, and it's hard to see it earning trust from AI either.

 

When Everyone Uses the Same AI, Your Tone of Voice Becomes Context

Here's the obvious risk with AI in marketing: if everyone feeds generic prompts into the same public models, everyone gets the same content.

The way out is to treat your taste and tone of voice as context, just like your data. That means being able to answer questions like these:

  • Who is your buyer, and how do they talk?
  • Which problems actually matter to them?
  • Why should they read what you write?
  • How do you explain things?
  • Which words sound like you, and which are banned?

A brand guide sitting in a shared folder doesn't change anything. It has to feed into your tools before a single word gets drafted.

We followed that rule for this article. Nobody typed "write an article about UNBOUND" into a chatbot. We pulled together our workshop notes, the demos we tested, conversations with HubSpot's product teams and what our own team saw on the ground.

AI can help organise ideas. It can't go to Boston for you.

 

Fast Learning Beats the Perfect Plan

Another lesson from Boston goes well beyond AI. Plan quarterly, and you get four chances a year to check what's working. Work in two-week cycles, and you get 26.

The team that learns fastest wins.

One speaker made the point with a story from 1903. Samuel Langley had serious government funding behind him, yet his aircraft crashed into the Potomac River twice. The Wright brothers, working on a shoestring, built cheap gliders and kept testing and adjusting. Nine days after Langley's second crash, they flew at Kitty Hawk.

AI has made experimenting far cheaper. You can turn one long article into content for several channels in minutes, or test a handful of short ideas, see what people actually engage with and put your budget behind the winners. Moving fast doesn't replace strategy. It means you can change course before a bad idea gets expensive.

 

Unified Doesn't Mean Putting Everything in One Tool

"Unified" was another word we heard everywhere in Boston.

Platforms built for very different jobs now connect far more deeply. They share data, and AI agents can work across them. We saw this ourselves in Agent Builder, which connects natively to the tools teams use every day.

But connecting your tools isn't the same as having a joined-up system. The more tools, integrations and automations you add, the more a few big decisions matter:

  • Which data actually matters?
  • Where should each core record live?
  • What's your single source of truth?
  • Which processes depend on shared data?
  • What should the whole setup deliver for the business?

No piece of software can answer those for you. That's where strategy comes in.

At power2Cloud, we don't pile tech on top of tech. We design systems where your data, platforms and people all work towards the same business goals.

To us, "unified" doesn't mean forcing every workflow into one piece of software. It means every platform has a clear job, and the whole thing stays manageable.

 

Leadership Sets the Direction for Technology

How a team approaches innovation says a lot about who's leading it. And as software starts doing more on its own, clear direction from the top matters more, not less. Several keynote speakers came back to this:

  • Sunita Williams, the astronaut, talked about her extended stay on the International Space Station. For her, resilience means not dwelling on setbacks, and giving teams room to experiment.
  • Cynthia Erivo talked about how professionalism carries teams through the hype, reminding the room that "no red light lasts forever".
  • Amy Webb, the futurist, drew a line between companies that wait for the future to happen to them and those that shape it. Her example was Nintendo, founded in 1889, which is choosy about where AI adds value and where human craft has to win.

Good leadership isn't about adopting every shiny new tool. It's about choosing the technology that serves where you're going.

 

Why We Went All the Way to Boston for UNBOUND

We don't want to read about HubSpot's latest features in a release note and then work out how to roll them out for clients. We want to see them first. That's why we make the trip to HubSpot's flagship conference every year, and why our HubSpot Trainers and Community Champions keep signing up for certifications and bootcamps.

Three days at UNBOUND gets you new ideas, people from very different business cultures, use cases you'd never have thought of, and beta features you can try before anyone's writing about them.

The schedule is packed, so we split up. One of us takes a workshop, another covers the keynote, a third gets hands-on with a demo. By the evening, photos, notes and slides are already landing in our shared Drive folders for the team back home.

Some of the best conversations of the week happened over pepperoni pizza. We got chatting to partners from Australia and Colombia, compared notes on the projects we're running, and quickly realised we're all fighting the same battles: messy data migrations, getting sales teams to actually use the CRM, AI workflows, the lot.

That's what we like most about the HubSpot community: the people building on the platform and the people using it every day end up in the same room. In one session, a marketing manager from Boston told us she'd spent years on Salesforce before making the switch. Her verdict? She gets more done now, and spends a lot less time wrestling with her software.

It's a good reminder of why we go. New features are exciting, but what matters is whether they make someone's working day easier. That's the test we'll be applying to everything we brought home from Boston.

 

Real Value Comes When Technology Meets Human Judgement

AI speeds up the work. But designing processes, making sense of the results and shaping your company's context still need people.

The real value comes from the groundwork: a clean data model, sensible governance and tools that line up with your business goals. Get that right and technology drives growth. Get it wrong and it's just another cost.

That's the balance we help our clients strike at power2Cloud, from redesigning processes to rolling out AI agents that people actually use.

Want to know what the Boston announcements mean for your own HubSpot portal? Join our upcoming webinar on HubSpot UNBOUND 2026, where we'll walk through what's changed and how to put it to work.

Get in touch to reserve your place or receive registration updates.

RELATED ARTICLES