Your CRM’s biggest blind spot is getting worse
900+ companies had CRM data stolen via connected apps in 2025. Now you don't even need to be a target.
18 August 2026
Intent data has always promised greater relevance. But it’s only really achieved mildly improved targeting, until now.
There’s an important difference between ‘here is a list of companies that fit your profile’ (targeting), and ‘here is a company where something just changed, and now they might need what you sell’ (relevance). The first is a filter, but the second is the reason to pick up the phone.
HubSpot’s standard buyer intent has sat firmly in the first category. It tells you which companies visited your website. Useful only really as a proxy, because a website visit is not a buying signal it’s a possibility.
At the end of June 2026, HubSpot shipped something more intriguing. Custom intent signals, now in public beta, let you define the specific business moments that matter to you, in plain language, and HubSpot monitors for them across the web. When it finds a match, it returns the company, source, date, and the context.
It is the closest HubSpot has ever come to giving every rep the kind of research that good reps already do manually. Whether it works depends, as always, on what sits underneath it.
Custom intent signals are part of a three-feature update to HubSpot’s Buyer Intent, released on 30 June 2026:
The third one is the significant change. Standard buyer intent tracks website visits (first-party, via your tracking code) and research topics (across 200,000+ third-party websites). The pre-built news signals detect known event types like funding rounds, leadership hires, M&A. Custom signals let you go beyond all of that and define the event yourself.
As Alex Bauer, VP of Marketing at DemandWorks, put it after two months of early access: “It lets you define the moments that actually matter to your business.” His framing is useful: “Personalisation says ‘I know who you are.’ Relevance says ‘I know what you’re dealing with today.’”
Here is how the layers of HubSpot intent data now stack up:
| Layer | What it detects | Data source |
|---|---|---|
| Website visitor intent | Companies visiting your site | First-party tracking code and IP matching |
| Research topic intent | Companies researching topics you define | 200,000+ third-party websites |
| Pre-built news signals (20) | Funding, hires, M&A, product launches, growth metrics | Web monitoring |
| Key stakeholder / pain point | Decision-makers and stated challenges | Deal-associated emails, notes, transcripts |
| Custom intent signals (beta) | Business moments you define in plain language | Web monitoring with cited sources |
Custom signals are available to customers with HubSpot Credits and are currently in public beta.
The power of custom signals is that you define the moment. The risk is that vague signals return noise and burn credits on results nobody acts on.
A good signal describes a specific, observable event that would change how you sell to that company. The test is ‘if this signal fired, would a rep do something different?’ If not, it’s not a signal, just trivia.
Example signals worth building:
Signals that waste credits:
The specificity is the feature. If you write a signal as vaguely as a Google search, you will get Google-quality results. Which is to say a lot of everything, and none of it prioritised.
Custom intent signals detect the moment. Smart properties research the company. Standard buyer intent tracks website engagement. So despite the slightly confusing semantics they are complementary rather than competing.
The useful mental model:
In practice, these chain together. A custom signal fires (company hires a VP of RevOps). A smart property enriches the company record with industry, tech stack, and ICP fit score. A workflow routes the record to the right rep with the signal, the enrichment, and a recommended action. The rep gets a timely notification that says something useful.
Versus third-party intent platforms (Bombora, 6sense, G2): These track research behaviour across thousands of publisher sites, giving you surge scores for topic-level intent. HubSpot’s research topic intent partially replicates this (across 200,000+ sites). Custom intent signals do something different in monitoring for specific events you define, not research patterns. They are not a replacement for category-level intent data, they’re a different kind of signal entirely.
If you are already running Bombora or 6sense, custom signals add a layer they do not cover which is the specific, observable business moments that are harder to detect from research behaviour alone.
Custom intent signals use HubSpot Credits from the same shared pool as every other AI feature in the platform. This matters because the cost is not visible in isolation. Rather, it compounds across features.
Buyer intent tracking costs 10 credits per company per month. That covers all signals for that company (website visits, research topics, news signals, and custom signals). This way you don’t pay per signal, you pay to track a particular company.
However that same credit pool also funds:
Your monthly allocation depends on your plan. Starter tier receives 500, Professional 3,000 and Enterprise 5,000. Additional credits come in at $10 per 1,000 and note that credits do not roll over. Consider carefully your spend if you are tracking 200 target accounts (2,000 credits/month) and running smart properties on new MQLs, and your sales team is using the Prospecting Agent, how much headroom you have. Do the maths across all features before you enable a new one. The credit pool is shared and the budget conversations need to be too.
A defined target-account list. You need to know which companies to track. Buyer intent charges credits per company per month. If you do not have a named target-account list, you are paying to monitor companies at random. Start with your ICP, build a segment, and track that segment. Expand when you see value, not before.
Workflows ready to act. A signal is only useful if something happens when it fires. Build the workflow before you enable the signal. What happens when a custom signal fires? Who gets notified? Does it create a task? Does it trigger enrichment? If the answer is “it appears on a timeline nobody checks,” the signal is wasted.
Enrichment in place. The signal tells you something happened. You still need context on the company. Smart properties, a third-party enrichment tool, or even a well-maintained set of standard properties fills that gap. A signal without context is a notification without a next step.
Rep adoption planned. If reps do not know the signals exist, or do not trust them, the feature is shelfware. The advantage of custom signals is that each one comes with source, date, and context. Show reps the evidence behind the signal so they can verify before acting. Trust is built on transparency, and this feature provides it, if someone shows the team where to look.
Credit budget allocated across features. Decide how credits are split across buyer intent, smart properties, and AI agents before someone burns the monthly allocation. This is a team-level conversation, not a feature-level one.
If that list sounds like a lot of groundwork, it is. But it is the same groundwork that makes every other CRM feature deliver value. The pattern is always the same: the feature is ready, but the infrastructure underneath it might not be. That gap is the kind of CRM architecture work Fly does through its Flight Plan retainer: target-account strategy, workflow architecture, data models, and the governance that stops credits being wasted on features nobody is ready to use.
Custom intent signals are the most interesting addition to HubSpot’s buyer intent in a long time. They move beyond proxies (page views, email opens, research topics) toward the kind of specific, observable business moments that good reps have always researched manually.
The feature is new and is still in beta (at the time of publishing). The coverage is thin and the edge cases are unknown but the architecture is sound – define the moment, monitor for it, return it with evidence.
Whether the potential translates to value depends on the CRM underneath. As always, the foundation is still the hard part.
Talk to our team about your platform.