Show & Tell: HubSpot’s New AI Capabilities
- Helena Sampayo
.JPG/v1/fill/w_320,h_320/file.jpg)
- Jun 2
- 3 min read
Updated: Jun 3
Updated 3 June, 2026
One of the things we've been exploring recently is HubSpot's newer AI capabilities.
HubSpot has really boosted its AI capabilities. Which is great, but it also adds another layer of complexity when you're deciding which AI workflow to use where.
While we'll be discussing exactly this (and more) in our new workshop - AI Workflows for GTM Teams, we also wanted to share some of what we've been testing so far.
Below, we'll walk through the three areas we've spent the most time exploring and where we're finding them useful in practice.
Our early observations of HubSpot's new AI capabilities
As a team, we're always more interested in useful workflows than AI for AI's sake.
We're also conscious of not burning credits across five different platforms when some of that work could already happen inside your CRM.
That's what made these recent releases interesting to us.
A lot of the outputs teams are using AI to generate - summaries, classifications, enrichment, and automation - ultimately end up in HubSpot anyway.
So we've been exploring how much of that work can now happen inside the CRM itself.
Most HubSpot portals now include a monthly allocation of HubSpot AI credits. To check yours, head to:
Settings > Account Management > Usage & Limits > HubSpot Credits
From our testing so far, there are three areas worth paying attention to:
Smart Properties
AI Workflow Actions
HubSpot's new Agentic Automation Builder
Smart Properties
What is it?
AI-generated custom properties that use prompts to classify, summarise, or transform CRM data.
What it's good for
Lead categorisation, ICP tagging, summarising notes, identifying intent, classifying inbound leads, and lightweight enrichment directly inside HubSpot.
Limitations
Best suited to working with data that's already in your CRM. It's less useful for deep research, website crawling, enrichment waterfalls, or large-scale prospecting.
What to try first
"Summarise this company"
"Classify lead source"
"Is this company an agency, SaaS business, or ecommerce brand?"
AI Workflow Actions
What is it?
AI-powered actions embedded directly into workflows.
What it's good for
Automating tasks based on lifecycle changes, form submissions, deal updates, or support activity.
We've been particularly interested in use cases such as lead routing, inbound form summaries, churn-risk detection, and internal handover notes.
Limitations
Still designed around HubSpot records and workflows. More advanced research, prompt chains, and high-volume enrichment can become difficult (and expensive) fairly quickly.
What to try first:
Auto-summarise inbound forms for SDRs
Route leads by intent
Detect churn risk from support tickets
Create internal handover summaries automatically
Agentic Automation Builder
What is it?
HubSpot's next-generation AI workflow builder, currently in private beta.
Think workflows, AI agents, and external triggers brought together in a single automation layer.
What it's good for
More dynamic workflows where research, decision-making, or multiple systems need to work together.
Limitations
Still early. Some integrations and workflow functionality are continuing to evolve, so we're treating this one as a space to watch rather than a finished product.
What to try first
AI-powered company research workflows
Triggering workflows from Sheets or webhooks
Automated deal or account research
Cross-sell and upsell recommendation workflows
Our takeaway so far
The biggest thing we've taken away from this testing is that HubSpot can now handle more AI-driven workflows than we expected.
We're not rebuilding everything overnight. But for certain use cases, we're finding ourselves checking what's possible inside HubSpot before automatically reaching for another tool.
We'll keep updating this post as we continue testing and as HubSpot rolls out new capabilities, so keep an eye out!

Comments