HubSpot has announced a new pricing structure for two of its flagship AI agents.
Effective April 14, HubSpot is moving its Breeze Customer Agent and Breeze Prospecting Agent to an outcome-based pricing model.
The Customer Agent will be charged at $0.50 per resolved conversation; the Prospecting Agent at $1 per qualified lead.
In layman’s terms, this means no resolution, no charge; no qualified lead, no invoice.
This pricing shift will undoubtedly carry real implications for CX and customer service leaders evaluating where to put their budgets.
In HubSpot’s official blog, the company stated:
“We believe AI should be priced on the value it delivers, not the compute it consumes.”
This framing is designed to position Breeze not as a cost center to manage but as a revenue and efficiency lever to invest in, which may land very differently on a CFO’s desk.
The company has some performance numbers to back the confidence.
Across more than 8,000 customers, the Breeze Customer Agent is reporting a 65% resolution rate and a 39% faster resolution time.
The Prospecting Agent, on the other hand, has seen a 57% quarter-over-quarter increase in activation and is contributing to 10% higher close rates across more than 10,000 customers.
Before unpacking what the shift means commercially, it is worth clarifying what each agent actually does:
- The Breeze Customer Agent is a 24/7 AI support tool trained on a company’s knowledge base, website, and blog content to handle inbound customer queries without a human agent in the loop.
- The Breeze Prospecting Agent sits on the sales side: it researches prospects, personalizes outreach, and surfaces qualified leads so that sales reps can focus on closing rather than sourcing.
Part of a Broader Industry Shift
HubSpot joins a growing list of CX vendors rethinking how AI is sold and charged for.
Zendesk made an early move in this direction, launching what it described as an “industry first” outcome-based pricing model in mid-2024, under which customers are only charged for issues resolved autonomously by AI.
Speaking at the time, Nikhil Sane, SVP & GTM of Strategy and Pricing at Zendesk, said:
“Our outcome-based pricing solution is more than just a pricing model — it’s a reflection of our dedication to driving real, measurable success for our customers.”
Salesforce has taken a more iterative approach with Agentforce, making repeated adjustments to its pricing structure in an effort to drive adoption.
Having replaced a $2-per-conversation model with a credit-based system – where individual actions cost around $0.10 – it subsequently introduced pay-as-you-go and pre-commit payment options alongside the existing pre-purchase model.
Bill Patterson, EVP of Corporate Strategy at Salesforce, said the goal was simple:
“We’re removing the friction and lowering the barrier to entry so every company, whether they’re a long-time customer or trying Salesforce for the first time, can get started and see immediate value from digital labor with Agentforce.”
On the contact center infrastructure side, AWS has taken a consumption-based approach with Amazon Connect, charging for usage rather than seats.
When Samsung selected Amazon Connect over traditional CCaaS providers, Reuben Lowenstein, Customer Care Manager at Samsung Electronics, pointed directly to the pricing model as a factor:
“AWS is not a ‘telephony company’, it is an AI company with a modular platform where you purchase what you need and pay for what you use.”
Buying on Seats Is “Lazy Now”
The shift in commercial logic is something Matt Price, CEO of Crescendo, has observed closely.
Speaking to CX Today, he explained that outcome-based pricing has become viable primarily because the measurement infrastructure now exists to support it.
“Before, it was really impossible to measure the true outcome of what was going on. People had to use proxies: hours of labor, or surveyed results.
“For the first time, you get this full picture. What we’re able to do then is say, at the end of the journey, ‘Was the problem solved?’ and ‘Was the customer happy?’ And in that case, we only charge if they were happy at the end of the whole journey.”
On why seat-based models have persisted, Price argued that “it suits the legacy vendors because it’s very hard for them to re-architect their systems to actually do the instrumentation and measuring to actually do the outcome.
“Quite frankly, that’s lazy now, and it’s not doing the best thing for the business.”
He also raised a warning for buyers evaluating any outcome-based model, claiming that resolution rate alone does not tell the full story.
“If you have a 60% AI solve rate, that means 40% of the people have gone through the AI journey and then have to do something else. A lot of AI is actually adding friction to the journey, and those customers end up being unhappier – they switch channels, they reopen tickets.”
He likened it to the fast food industry, stating:
“It’s a bit like a fast food restaurant measuring the number of burgers but not taking into account if people are getting food poisoning at the other end.”
As for whether outcome-based pricing will become the dominant model across the market, Price was forthright in his belief:
“Yes, and it really is the right time to do it… I think hundreds of customers who are adopting it from us are proving that.”
You can check out Matt’s full conversation with CX Today here.
