It's not a price cut. It's a redefinition of what HubSpot is allowed to charge for, and the buyers I've talked to in the past 72 hours haven't caught up yet. The old meter ticked on every conversation. The new meter ticks only when Breeze "resolves" something, but resolution is defined by HubSpot, not by you. At 5,000 resolutions per month, you're paying HubSpot $2,500. The same volume on a flat-rate competitor runs about $99.
This post walks through the math, the competitive timing, and what an SMB should actually do.
TL;DR: HubSpot moved Breeze AI from $1.00 per conversation to $0.50 per resolved conversation on April 14, 2026 (HubSpot, 2026). The headline is a 50% cut. The reality is a redefined billable event that still scales linearly with your support volume. At 5,000 resolutions per month, you'll pay $2,500.
AI support pricing models in 2026
What did HubSpot actually change about Breeze pricing?
HubSpot replaced per-conversation billing with per-resolution billing on April 14, 2026, dropping the unit cost from $1.00 to $0.50 (HubSpot, 2026). Breeze still lives inside Service Hub. What changed is the trigger: the meter only fires when Breeze marks a conversation "resolved" without escalating to a human agent.
The old model
Under the previous billing, every conversation Breeze touched counted. A user types "hi," gets a response, abandons the chat: that's $1.00. An angry customer pings four times, gets escalated to a human within 30 seconds: also $1.00. The meter cared about engagement, not outcome.
This made the cost completely uncorrelated with value delivered. SMBs running high-volume but low-resolution traffic (think e-commerce returns or password resets) were paying for noise.
The new model
Now, only "resolved" conversations bill. HubSpot defines resolved as: Breeze handled the issue end-to-end, the customer didn't request a human, and the conversation closed without escalation within a defined window. If a human picks it up, no charge. If the customer ghosts the bot, also no charge in most cases (MarTech, 2026).
On paper this looks like vendor risk-sharing. In practice, HubSpot controls the resolution definition and the model that decides what "resolved" means. You'll find out at month-end how many resolutions you bought.
Citation capsule: HubSpot moved Breeze AI from per-conversation billing at $1.00 to per-resolution billing at $0.50, effective April 14, 2026 (HubSpot, 2026). Breeze claims a 65% resolution rate and 39% faster resolution time across more than 8,000 users on the platform.
How does HubSpot's framing compare to the buyer's reality?
HubSpot's launch deck says they cut the price in half. The buyer's spreadsheet says something different: at a 65% resolution rate, you're paying $0.50 on 65% of conversations instead of $1.00 on 100% (HubSpot, 2026). That's $0.325 per conversation effective rate, not $0.50. So yes, cheaper. But not half-off cheaper.
The math the press release skips
Take a typical SMB doing 1,000 conversations per month:
- Old model: 1,000 × $1.00 = $1,000/month
- New model at 65% resolution: 650 × $0.50 = $325/month
That's a 67% reduction at the published resolution rate. Sounds great. But here's where the framing breaks down. The 65% rate is HubSpot's average across 8,000+ users. Your actual rate depends on three variables HubSpot doesn't control: your knowledge base quality, your conversation complexity, and how aggressively their model decides to "resolve."
Why the resolution rate matters more than the price
In analyzing pricing models for SMB AI support tools, we've found that vendor-defined resolution rates tend to drift upward in the vendor's favor over the first 12 months of a contract. SaaStr made the same point about HubSpot's specific change, noting that "the vendor decides what resolved means, and that definition will get more generous to the vendor over time" (SaaStr, 2026).
Translation: the $0.50 price will hold. The number of conversations classified as resolved will grow. Your bill will track the latter, not the former.
What does the math look like at typical SMB volumes?
At 500, 2,000, and 5,000 resolutions per month, HubSpot Breeze costs $250, $1,000, and $2,500 respectively at the new $0.50 rate (HubSpot, 2026). The same volume on Intercom Fin runs $495 to $4,950 in resolution fees alone, before the $85/seat charge (Intercom, 2026).
500 resolutions per month (small SMB)
| Vendor | Resolution cost | Seat cost | Total/mo |
|---|---|---|---|
| HubSpot Breeze | 500 × $0.50 = $250 | included | $250 |
| Intercom Fin | 500 × $0.99 = $495 | $85 | $580 |
| Zendesk AI Agents | 500 × $1.50 = $750 | $55 | $805 |
| Corebee | flat | flat | $99 |
At this volume, HubSpot looks competitive. Corebee still wins on raw cost by 60%, but HubSpot beats Intercom and Zendesk handily.
2,000 resolutions per month (mid-stage SMB)
| Vendor | Resolution cost | Seat cost | Total/mo |
|---|---|---|---|
| HubSpot Breeze | 2,000 × $0.50 = $1,000 | included | $1,000 |
| Intercom Fin | 2,000 × $0.99 = $1,980 | $85 | $2,065 |
| Zendesk AI Agents | 2,000 × $1.50 = $3,000 | $55 | $3,055 |
| Corebee | flat | flat | $99 |
The gap widens. HubSpot beats Zendesk by 3x. Corebee beats HubSpot by 10x.
5,000 resolutions per month (scaling SMB)
At 5,000 resolutions per month, HubSpot bills $2,500 versus Corebee's $99 flat fee, a 25x difference (HubSpot, 2026). This is where per-resolution pricing breaks down for growing businesses. Every successful AI interaction becomes a tax on your growth, not a fixed operating cost you can plan around.
When I talk to founders evaluating AI support tools, the question they always miss is: "What does this cost when it works?" Per-resolution pricing punishes the exact outcome you're paying for. Every successful deflection bumps your bill.
Compare Zendesk vs HubSpot for support
Why did HubSpot make this move now?
HubSpot moved because outcome-based pricing has become the default frame for enterprise AI agent deals, and per-conversation billing was already a hard sell against competitors charging only on resolution. Gartner forecasts that 40% of SaaS spend will shift to outcome-based models by 2030 (Gartner, 2026). This is competitive defense, not customer generosity.
The competitive backdrop
Vendors with deep enterprise pockets are racing toward outcome pricing. Several well-funded AI agent startups now bill enterprises only when their agents resolve issues, and they've been winning the same Service Hub accounts HubSpot wants to defend. The narrative had shifted: outcome-based was the new default for AI support pricing. HubSpot couldn't sit on per-conversation while every competitor reframed the conversation.
The Gartner pressure
Gartner published a forecast on January 26, 2026, predicting that GenAI cost-per-resolution for customer service will exceed offshore human agent costs by 2030 (Gartner, 2026). The same analyst track flags 40% of SaaS spend shifting to outcome-based models by 2030 (Gartner, 2026). HubSpot read the writing on the wall.
The buyer behavior shift
Procurement teams started asking for "outcome SLAs" in RFPs in Q4 2025. Per-seat AI tools were being benched. HubSpot's enterprise sales team was losing deals to upstarts that priced on resolution from day one. Moving Breeze to per-resolution was less a product decision and more a sales-enablement decision.
Citation capsule: Gartner predicts GenAI cost-per-resolution for customer service will exceed offshore human agent costs by 2030, with 40% of SaaS spend shifting to outcome-based models in the same window (Gartner, 2026). HubSpot's pricing change reflects that pressure, cutting Breeze AI to $0.50 per resolved conversation.
What's the structural problem with per-resolution pricing?
Per-resolution pricing has three structural flaws: vendors define what counts as resolved, costs scale with success rather than effort, and growth becomes a tax line. Gartner predicts cost-per-resolution will exceed $3 across the industry by 2030, a 6x increase from HubSpot's current $0.50 (Gartner, 2026).
Flaw one: vendor-defined resolution
You don't get to define "resolved." HubSpot does. Their model decides whether a conversation closed cleanly. Their dashboard shows you the count. You pay the bill. There's no third-party arbitrator. Over time, vendors have every incentive to broaden the definition, as SaaStr called out specifically when the change was announced (SaaStr, 2026).
Flaw two: success scales your costs
Flat-fee tools cost the same whether you handle 100 or 100,000 conversations. Per-resolution tools punish you for using them well. The better Breeze gets, the higher your bill. The product getting better is a cost increase, not a benefit.
Flaw three: unpredictable budgeting
CFOs hate variable cost lines that scale with traffic. Marketing spend is variable but controllable. Support cost should be a fixed operating line. Per-resolution AI flips that. Your support OPEX now floats with traffic spikes you didn't forecast. That's a finance problem dressed up as a product feature.
AI support pricing models 2026
What should an SMB actually do about this?
Match the pricing model to your volume. Under 500 resolutions per month, flat-fee tools win on cost and predictability. Over 5,000, per-resolution costs ($2,500+ on HubSpot) get hard to justify against a $99 flat alternative (HubSpot, 2026). The midrange is where the calculus gets interesting.
If you're under 500 resolutions/mo
You're a small SMB or early-stage. Flat-rate wins. The variance on per-resolution will hurt you when traffic spikes (product launch, holiday surge, viral moment). Pick a tool that doesn't make spikes expensive.
If you're 500 to 2,000 resolutions/mo
This is the contested zone. HubSpot's $0.50 looks reasonable, especially if you're already on Service Hub. But run the math on your worst-case month, not your average. If you 3x in volume because of a campaign, can you absorb the bill?
If you're over 5,000 resolutions/mo
Flat-rate is almost always cheaper at this scale. The $2,500 you'd pay HubSpot at 5,000 resolutions buys you 25 months of Corebee at $99/mo. Even if you only hit that volume two months a year, the savings compound fast.
Build the spreadsheet first
Before signing anything, model 12 months of volume across three scenarios: flat, 50% growth, and 200% growth. Compare total cost across HubSpot, Intercom Fin, Zendesk, and a flat-rate option. Whichever wins your worst case is the right answer.
How does Corebee fit into this conversation?
Corebee charges $99 per month for unlimited AI resolutions. No per-conversation fees. No per-resolution fees. No surprise scaling costs. At 5,000 resolutions per month, that's a 25x cost difference compared to HubSpot Breeze (HubSpot, 2026).
If predictable monthly cost matters more to you than enterprise feature parity, the math is straightforward. If you need HubSpot's CRM tie-in and broader Service Hub workflow, Breeze at $0.50/resolution is genuinely competitive against Intercom and Zendesk; just budget for the bill scaling with your success.
Frequently asked questions
What is HubSpot Breeze?
HubSpot Breeze is the AI customer agent built into Service Hub, launched in 2024 and repriced on April 14, 2026 (HubSpot, 2026). It handles tier-one support conversations, claims a 65% resolution rate across 8,000+ deployments, and integrates directly with HubSpot CRM. As of April 2026, it bills $0.50 per resolved conversation.
What's the difference between per-conversation and per-resolution pricing?
Per-conversation pricing charges for every interaction, regardless of outcome. Per-resolution charges only when the AI closes the issue without human escalation. HubSpot moved from $1.00 per conversation to $0.50 per resolution on April 14, 2026 (HubSpot, 2026). The vendor defines what counts as resolved, which is the structural risk.
How does HubSpot Breeze compare to Intercom Fin in 2026?
Intercom Fin charges $0.99 per resolution plus $85 per seat (Intercom, 2026). HubSpot Breeze charges $0.50 per resolution with no separate seat fee for the agent itself. At 1,000 resolutions per month, Breeze runs $500 versus Fin's $1,075. Breeze wins on price; Fin's been in market longer with deeper integrations.
Is $0.50 per resolution a good price?
It's competitive against current AI support tools. Gartner forecasts industry cost-per-resolution will exceed $3 by 2030 (Gartner, 2026), suggesting prices will rise. At $0.50 today, HubSpot looks cheap. At your renewal in 24 months, you may be paying $1.50. Lock in flat-rate pricing if you can.
Will HubSpot raise prices again?
Probably yes, on the resolution definition before the dollar amount. SaaStr's analysis of the April 2026 change notes that vendors typically broaden "resolved" criteria over time, increasing billable events without changing the unit price (SaaStr, 2026). The $0.50 sticker may hold for years while your effective bill climbs.
What counts as a "resolved conversation" in HubSpot Breeze?
A resolved conversation is one Breeze handled end-to-end without human escalation, where the customer didn't explicitly request a human and the conversation closed within a defined window (MarTech, 2026). HubSpot's model decides this. There's no third-party verification, which is the structural concern with vendor-defined outcome pricing.
Should I use HubSpot Breeze or a flat-rate AI tool?
Match pricing to volume. Under 500 resolutions per month, flat-rate wins on predictability. Over 5,000, the cost difference becomes substantial. $2,500 on HubSpot versus $99 flat on Corebee, a 25x gap (HubSpot, 2026). If you're already deep in HubSpot's CRM, Breeze's tie-in may justify the premium. If not, flat-rate almost always wins for SMBs.
The bottom line
HubSpot didn't cut prices. They redefined the billable event in a way that aligns with where outcome-priced competitors were already winning enterprise deals. The $0.50 sticker is real. The cost predictability problem is also real. SMBs scaling from 500 to 5,000 conversations a month will see their HubSpot bill grow 10x, all for using a tool that's working well.
The right answer for most small businesses is still a flat-rate model. Predictable cost, no penalty on growth, no vendor-defined outcomes deciding your monthly invoice. If HubSpot's CRM stack matters to you, Breeze at $0.50 is competitive in 2026. Just build the spreadsheet through 2027 and 2028 before you sign.