HubSpot AI Features Review: Are They Actually Useful?

HubSpot has been aggressively adding AI across its platform since 2023. There are now AI features in the CRM, the marketing hub, the service hub, and the sales tools. The question isn't whether HubSpot has AI — it's whether any of it is actually worth using.

In our testing, the answer is: some of it is, most of it isn't, and the features that matter most are locked behind the expensive tiers.

Here's the full breakdown.

What HubSpot AI Actually Includes#

HubSpot's AI surfaces under several different names and locations. Getting a clear picture requires navigating their documentation carefully.

ChatSpot#

ChatSpot is HubSpot's conversational AI interface — a chat window where you can ask questions about your CRM data, generate content, or trigger HubSpot actions in natural language.

In practice, ChatSpot is useful for simple lookups. "How many contacts did we add this month?" gets you a quick answer without clicking through four dashboards. Creating a contact via chat also works cleanly.

Where it falls apart: complex queries. "Show me all deals over $50k that haven't moved in two weeks with a contact from a company over 500 employees" produces either a vague list or an error. You end up building the filter view manually anyway.

ChatSpot is available to all HubSpot users, but its usefulness is limited to HubSpot's own data model. If you've customized your CRM heavily, expect friction.

AI Content Assistant#

The content assistant is embedded in HubSpot's email editor, blog editor, landing page builder, and sequence composer. You can highlight text and ask the AI to rewrite it, expand it, or change its tone.

This one genuinely saves time for marketing teams. Drafting a follow-up email sequence is faster with the AI suggesting variations. The quality is roughly GPT-3.5 level — good enough for a first draft, not good enough to publish without editing.

The content assistant is included in paid Marketing Hub and Sales Hub plans. Free HubSpot users don't get it.

Predictive Lead Scoring#

HubSpot's AI-powered lead scoring analyzes your historical closed-won deals and scores incoming leads based on patterns. If your best customers are mid-market SaaS companies in North America who attended a webinar, new leads matching that profile get a higher score.

This is one of the more genuinely useful AI features — when it has enough data to learn from. You need a significant amount of historical conversion data for the model to be meaningful. If you're a small team or early-stage, the model has nothing to work with and the scores are noise.

Predictive lead scoring requires Marketing Hub Professional or Enterprise — starting at $800/month.

AI-Powered Forecasting#

Sales Hub Enterprise includes AI-powered deal forecasting that projects likely close amounts and flags at-risk deals. It analyzes rep behavior, deal velocity, engagement signals, and historical win rates.

In our review of published user feedback, the forecasting is most accurate for teams with at least 6 months of pipeline history and consistent CRM data hygiene. Teams that don't log activities religiously get unreliable forecasts.

Breeze Copilot (formerly AI Assistant)#

HubSpot recently consolidated many of its AI features under the "Breeze" brand. Breeze Copilot is the embedded AI that appears in sidebars throughout HubSpot, offering contextual suggestions, email drafts, and summaries.

It's the most ambient of the AI features — you don't have to go looking for it, it shows up where you're working. The quality varies considerably depending on context.

The Pricing Problem#

Here's what HubSpot doesn't make easy to find: most of the useful AI features require Professional or Enterprise tier plans.

If you're on a free or Starter HubSpot plan, you have ChatSpot and limited content assistant access. That's it. The AI features that could genuinely transform your workflow — predictive scoring, forecasting, autonomous agents — require jumping to plans that cost more per month than most teams spend on their entire software stack.

Where HubSpot AI Falls Short#

Not Conversational Enough#

HubSpot's AI features are add-ons to a traditional CRM interface, not a conversational layer over your data. ChatSpot is a chat window alongside the HubSpot UI, not a replacement for it. You still navigate the same menus, click the same filters, and build the same dashboards. The AI assists; it doesn't replace the operational overhead.

A true conversational CRM would let you operate entirely through natural language: "add this person as a lead, set the source to LinkedIn, assign to the enterprise team." HubSpot can do some of this through ChatSpot, but the interface is auxiliary rather than primary.

Data Stays in HubSpot's Cloud#

Every AI feature in HubSpot operates on data that lives in HubSpot's infrastructure. Your contacts, your deal notes, your email sequences — all of it is in their database, accessible to their systems. The content assistant isn't calling your local files; it's pulling from their API.

For most users, this is fine. For teams with data sovereignty requirements or serious privacy concerns, it's a non-starter.

Customization Is Limited#

HubSpot's AI features work on HubSpot's data model. If you've heavily customized your CRM with non-standard objects, the AI often doesn't know what to do with them. ChatSpot has limited awareness of custom objects. The forecasting model doesn't adapt to unusual pipeline shapes.

The Context Window Problem#

Even the best AI features in HubSpot don't have deep context about your entire business. The content assistant doesn't know about the specific deal you're working on unless you tell it. ChatSpot doesn't have access to your email thread history. Each AI interaction is relatively stateless.

What HubSpot AI Does Well#

To be fair: HubSpot has made AI genuinely useful for a few specific workflows.

Email drafting in sequences is noticeably faster. The AI suggestions are relevant and require less editing than a blank template.

Meeting summarization (in the Sales Hub) produces usable call summaries from transcripts that would take 20 minutes to write manually.

Spam filtering and bounce handling work quietly in the background and save time.

The SEO content suggestions in Marketing Hub actually surface useful optimization opportunities.

These are table-stakes features that many tools now offer. They're good, not exceptional.

The Alternative: AI-Native CRM#

The honest conclusion from our HubSpot AI review: these are AI features bolted onto a CRM, not a CRM built for AI.

DenchClaw takes the opposite approach. The AI is the interface. Natural language isn't a sidebar feature — it's how you operate the entire system. You talk to your CRM the same way you'd talk to a team member: "show me deals stalled in the negotiation stage for more than two weeks" and a filtered view appears instantly. The AI writes the DuckDB query, updates the view configuration, and responds conversationally.

Because DenchClaw is local-first, your data never leaves your machine. The AI operates on your local DuckDB files. There's no "AI feature" separate from the product — the entire product is an AI agent that happens to manage your CRM data.

And it's open source. No $800/month plan required to access the intelligence features.

Frequently Asked Questions#

Is HubSpot AI free?#

Some HubSpot AI features are available on the free tier (ChatSpot, basic AI writing suggestions). But the most useful features — predictive lead scoring, AI forecasting, Breeze Intelligence — require Professional or Enterprise plans starting at $800/month.

Does HubSpot AI actually work for small teams?#

For small teams with less than 6 months of pipeline history, most of HubSpot's predictive AI features don't have enough data to produce useful outputs. The content assistant and basic chatbot are functional, but the headline AI features require established data to perform.

Can HubSpot AI access my emails and calendar?#

HubSpot AI can access the emails and calendar events that are synced to HubSpot. It doesn't have access to your local email client or system calendar — only data that's been explicitly imported or synced through HubSpot's integration layer.

How does HubSpot AI compare to Salesforce Einstein?#

Both platforms offer similar AI feature sets at the high end. HubSpot tends to be more accessible for mid-market teams; Salesforce Einstein is better suited to enterprise teams with large data sets and dedicated admin resources. Pricing is comparable at the top tier — both are expensive.

Is there an open-source alternative with AI CRM features?#

Yes. DenchClaw is an open-source, local-first CRM with a conversational AI layer. It runs on your machine, stores data in DuckDB, and lets you interact with your CRM entirely through natural language — at no per-seat cost.

Ready to try DenchClaw? Install in one command: npx denchclaw. Full setup guide →