HubSpot's evolving pricing model is changing how businesses should think about AI, automation, and the value of their CRM. And you already have Credits available to start experimenting.

For years, software pricing has largely followed a familiar model: companies pay for subscriptions, users, and access to features. Whether those features are used occasionally or every day, the subscription cost remains the same, predictable.

AI is introducing a different model.

Actually HubSpot has launched a new model (for new customers only - for the moment) based on users (seats) and consumption (credits and workflow steps). This would mean saying goodbye to "the hubs" and moving into an "all hubs included" platform (still with different levels of licenses starter-pro-enterprise).

With HubSpot Credits, businesses can now pay for specific work performed by AI agents and other consumption-based capabilities, alongside their existing subscriptions.

Instead of focusing exclusively on which features employees can access, companies can begin evaluating the work their platform performs on their behalf, and pay only for it.

That distinction is important. A salesperson using HubSpot to manage opportunities is different from an AI agent researching contacts, enriching records, analyzing companies, or responding to customers automatically.

And it's changing the economics of automation.

What are HubSpot Credits, and how do they work?

HubSpot Credits are a consumption-based currency used by certain AI and automation capabilities across the HubSpot platform.

Different actions consume different amounts of Credits, depending on the functionality used.

For example, according to HubSpot's published rate sheet:

• Customer Agent: 50 Credits per text-based conversation handled.
• Data Agent: 10 Credits per prompt, per record.
• Prospecting Agent: 100 Credits per lead for recommended outreach.
• Custom agents: Consumption varies depending on the actions and resources used during execution.

Additional Credits are available at a standard rate of $10 per 1,000 Credits, although the way customers purchase or consume additional capacity depends on their subscription and billing arrangements.

You can review the complete HubSpot Credits Rate Sheet for current rates and eligible features. 

The interesting part isn't simply that some AI functionality now has a variable cost. It's that businesses can start evaluating automation differently.

Rather than asking how much it costs to give another employee access to a feature, they can ask how much it costs to incorporate a new ability to the business, let's say complete a particular operational task, how reliable the result is, and whether that task is worth automating.

Of course, consumption-based pricing also requires more deliberate planning. Not every task needs AI, and not every AI action delivers equal value.

Also think about paying only for the amount of emails you send per month and not for a fixed tool price. This will make business be more conscious about the amount of communications they are doing with prospects and client and what they are getting from them, focusing over quality instead of quantity (always coming back to the basics).

You already have Credits to start experimenting

One aspect of HubSpot Credits that is frequently overlooked is that eligible subscriptions already include a monthly allocation at no additional cost.

Under HubSpot's current pricing, included monthly Credits vary by subscription:

Unused included Credits reset monthly rather than rolling over.

So if your company already have Credits available right now to test capabilities without purchasing additional capacity.

That's particularly interesting for companies that are curious about AI but haven't yet identified where to start.

You don't need to begin with a large AI transformation initiative. You can select a small, well-defined operational problem, test an agent, evaluate its results, understand its consumption, and then decide whether expanding it makes sense.

The goal of that first experiment shouldn't be to use all your available Credits. It should be to learn which business capability is worth developing.

Have unused HubSpot seats? You may be able to swap them for Credits

HubSpot has also introduced an interesting option for companies paying for seats they aren't currently using (and cannot downgrade until your contract renewal)

Eligible accounts can now swap unassigned paid seats for HubSpot Credits or other seat types, redirecting existing subscription spending toward capabilities the business can actually use.

For example, a company with unused Sales seats may be able to convert their value into Credits for AI-powered work rather than leaving that capacity unused.

The important conditions are that the seats must be unassigned, the swap must be available for the account, and the decision should be reviewed carefully: Credits received through a seat swap cannot simply be converted back into the original seats.

You can find the details and instructions in HubSpot's official guide: Swap unassigned seats for HubSpot Credits.

This is another indication of where the platform is heading.

Software budgets are becoming more flexible, allowing organizations to allocate some spending toward employees using software and some toward technology performing work.

But what does that actually look like?

What can AI agents do for your business? Four examples we've built at Andimol

At Andimol, we've been exploring how AI agents can address practical operational problems inside HubSpot.

From enriching CRM records and researching account opportunities to supporting customer service and handling complex operational processes, these agents demonstrate how AI can create useful capabilities across the customer lifecycle.

The most interesting opportunities aren't necessarily the most sophisticated agents. They're the ones that remove friction, provide better context, or support more informed decisions while maintaining the right level of control.

Here are four examples.

1. An AI agent that enriches your CRM automatically after a phone call

Imagine your sales team receives a call from someone who isn't yet properly identified in the CRM.

The call is logged through your calling app (Aircall or any other), but the contact record has little more than a phone number. Someone needs to research who called, identify their company, complete the contact information, and associate the records correctly.

It's a small administrative task, but one that repeats constantly.

We built an AI-powered enrichment agent that is triggered when a call is logged in HubSpot. It analyzes the available context, helps identify the contact, enriches the relevant information, and associates the person with the appropriate company—or creates the company record when necessary.

The agent can also be triggered manually and adapted to other calling platforms that record activity in HubSpot.

Instead of relying entirely on salespeople to complete this administrative work, the CRM can take care of much of the research and record preparation.

The capability isn't simply enrichment. It's making customer information more usable as interactions happen.

2. An AI agent that researches companies and recommends the best product fit

How much does your sales team really know about a prospect before the first conversation?

Understanding a company's business model, industry, challenges, and potential needs often requires research across multiple sources. And even when that information is available, connecting it to the right products or services isn't always straightforward.

We built an AI agent designed to analyze a prospect's or customer's company and recommend which products are most relevant to the business.

The agent gathers and interprets company information, provides additional context about the organization, and evaluates potential product fit based on your company's offering.

For the Sales team, this means approaching prospects with a better understanding of their business and more informed recommendations, rather than starting every conversation from scratch.

But the value doesn't end when a prospect becomes a customer.

Customer Success teams can use the same intelligence to better understand existing accounts and identify potential upsell and cross-sell opportunities based on the customer's business profile and product needs.

Instead of relying exclusively on historical purchases or manually reviewing accounts, teams can use AI-generated recommendations to inform their conversations and account-development strategies.

Of course, identifying a potential product fit doesn't mean a customer is necessarily ready to buy. The agent provides context and recommendations; the commercial decision remains with the people managing the relationship.

The capability isn't simply company research. It's turning business context into actionable commercial intelligence throughout the customer lifecycle.

That's particularly interesting because the same AI capability can support both new-customer acquisition and the development of existing customer relationships.

3. An AI agent that supports customer service around the clock

Customer support is another area where AI agents can perform meaningful work.

We implemented HubSpot's Customer Agent as part of a Help Desk solution covering email and WhatsApp.

The agent helps customers with common questions, using the company's support knowledge, while the broader solution supports classification, routing, escalation, and service management.

During the first month, the implementation achieved some encouraging operational results:

• 95% of tickets were automatically classified.
• 100% of after-hours inquiries were handled by AI.
• 76.55% improvement in SLA performance.
• 65.19% reduction in response times.

These figures reflect the broader Help Desk and AI solution, not the Customer Agent acting alone.

The point is not that every business will achieve the same results. Each implementation depends on the quality of its knowledge, the complexity of customer requests, and the processes supporting the agent.

The opportunity is to provide customers with more consistent assistance while allowing human representatives to concentrate on cases that genuinely need their expertise.

The capability isn't simply answering questions automatically. It's enabling a support operation that can respond more consistently, while keeping people involved where their expertise matters most.

4. An AI agent that helps manage CRM account-deletion compliance

Some of the most useful AI applications aren't customer-facing.

Consider an organization that receives requests to delete customer information from its systems.

Before deleting anything, it may need to determine which records are involved, whether they are associated with other entities, what the request covers, and which actions require review.

We designed an AI-enabled account-deletion compliance solution that combines interpretation with controlled, deterministic execution.

AI helps interpret the request and relevant CRM context, while structured automation handles the operations that should follow predictable rules, with appropriate safeguards. This distinction is essential.

This is an example of how AI and traditional automation can work together, rather than treating an autonomous agent as the solution to every problem.

The real opportunity: start with a business capability, not an AI agent

As HubSpot makes AI capabilities increasingly accessible, companies will naturally begin asking which agents they should activate.

But we think there's a better question:

Which part of your business would benefit from having work performed automatically, intelligently, and reliably?

Perhaps salespeople spend too much time completing CRM records.

Perhaps your commercial teams lack the context they need to identify the most relevant products for each customer.

Perhaps support representatives repeatedly answer questions that could be resolved without human intervention.

Or perhaps sensitive operational processes require time-consuming interpretation before structured actions can take place.

These are opportunities worth investigating.

Then comes the architectural decision.

Some work is best handled by deterministic automation, where the rules are clear and the result needs to be predictable.

Other work benefits from AI because it requires interpreting information, understanding context, or making a recommendation.

And some decisions should remain firmly in human hands.

HubSpot Credits make it possible to assign a consumption cost to many of these AI-powered activities, but cost is only one part of the equation.

Businesses should also consider accuracy, reliability, escalation requirements, operational impact, and ongoing maintenance.

A cheap agent that produces unreliable information isn't necessarily valuable. And a more expensive agent may be worthwhile if it performs an important task consistently and safely.

The objective shouldn't be to consume fewer Credits at any cost.

It should be to use the right technology to develop useful business capabilities.

Your next AI experiment may already be included in your HubSpot subscription

The move toward consumption-based pricing represents a meaningful change in how software creates and captures value.

Companies will continue paying for access to HubSpot, but they can increasingly invest in the work the platform performs on their behalf.

And with monthly Credits already included in eligible subscriptions, plus the possibility of converting unused seats into additional capacity, there may be fewer barriers to experimentation than you expect.

Our recommendation is simple: start small, choose a real operational problem, define what successful execution looks like, and test it before scaling.

At Andimol, we help companies identify AI opportunities, design agents and automation around their existing processes, and establish the architecture and governance needed to operate them reliably.

Already using HubSpot and wondering what you could do with your available Credits?

Book a 15-minute call with Andimol and let's explore a practical use case for your business.