Zendesk made a big shift in their AI pricing model. Instead of just charging per agent seat, they now charge per AI resolution: $1.50 if you commit to a volume, $2.00 on pay-as-you-go. This is on top of their existing per-agent pricing of $55 to $115/month.

The pitch sounds reasonable: you only pay when AI actually resolves something. But when you run the numbers at different volumes, the costs add up fast. And since you're still paying per-agent fees for your human agents, the "AI should save us money" math gets complicated.

What a Resolution Costs

Zendesk counts a "resolution" when AI handles a customer query from start to finish without human involvement. If the AI starts a conversation but hands off to a human, that's not a resolution (you don't get charged the per-resolution fee, but you do use agent time).

At $1.50 per resolution (committed volume):

At $2.00 per resolution (pay-as-you-go), multiply those numbers by 1.33.

These costs are on top of your agent seat costs. A 10-agent team on the Growth plan ($89/agent/month) pays $890/month in seat fees. Add 1,000 AI resolutions and your total Zendesk bill is $2,390/month, or $28,680/year.

The Volume Discount Question

Zendesk offers volume discounts starting at 5,000+ resolutions per month. The exact discount depends on your contract negotiation, but published rates suggest the per-resolution price can drop to $1.00 to $1.25 at high volumes.

Even at $1.00 per resolution, 10,000 monthly resolutions cost $10,000/month. That's $120,000/year just for AI, before agent seats.

For context: a human agent handling 25 tickets per day costs about $55,000 to $75,000/year in total loaded cost and handles roughly 6,250 tickets per year. So 10,000 AI resolutions per month (120,000/year) at $1.00 each costs $120,000, the same as about two human agents who would handle only 12,500 tickets.

The AI is faster and works 24/7, which has real value. But the cost advantage over human agents shrinks rapidly at high volumes with Zendesk's pricing.

How This Compares

The per-resolution pricing model is spreading across the industry. Here's how Zendesk stacks up:

Intercom Fin: $0.99 per resolution. Cheaper per-resolution, but you're also paying $29+ per seat for the platform.

Supp: $0.20 per classification, $0.30 per classification plus automated action. No seat fees. No base subscription. At 1,000 resolutions/month, Supp costs $300 compared to Zendesk's $1,500. At 5,000 resolutions/month: $1,500 vs $7,500.

Salesforce AgentForce: $2.00 per conversation. Similar to Zendesk's PAYG rate but on top of Salesforce's much higher per-user fees ($25 to $330/user/month).

The gap between Supp and the enterprise players is significant. Supp charges 5x to 7x less per resolution because it runs a purpose-built classifier instead of an LLM, which is dramatically cheaper to operate.

When Zendesk's Pricing Makes Sense

For teams already on Zendesk with AI resolution volume under 500/month, the incremental cost ($750 to $1,000/month) is reasonable if you value having everything in one platform.

It also makes sense when you need the full Zendesk ecosystem (help center, community forums, workforce management, quality assurance, advanced analytics). The per-resolution fee is just one line item in a larger platform cost. The integration value may justify it.

If you have a negotiated enterprise contract, the per-resolution rates can come down substantially. Large enterprises with 50+ agents and 10,000+ monthly resolutions negotiate rates well below published pricing.

When It Doesn't

Small to mid-size teams (under 20 agents) with AI resolution volume over 1,000/month feel the pain most. You're paying $1,500+/month for AI on top of agent seats.

Teams that don't need the full Zendesk platform are paying for features they don't use. Many teams use Zendesk for ticketing and email, and could get the same functionality from a simpler tool at a fraction of the cost.

The pricing model also works against AI-first strategies. If your goal is to automate as much as possible and minimize human agents, the more you automate, the more you pay in per-resolution fees. At high automation rates, the cost advantage of AI over humans narrows to the point where it's barely worth it.

The Hidden Math

Here's what most Zendesk pricing analyses miss: the per-resolution fee creates a perverse incentive. You want AI to resolve as much as possible (to reduce agent workload), but the more it resolves, the higher your AI bill gets.

At some point, the marginal cost of an AI resolution ($1.50) approaches the marginal cost of a human resolution ($5 to $15). For very simple tickets (password resets, FAQ answers) that take a human agent 2 minutes, the human cost might be $3 to $4. At $1.50 per AI resolution, the savings per ticket is only $1.50 to $2.50. Still positive, but not the 10x savings that AI support is supposed to deliver.

Compare to Supp at $0.30 per resolution: the savings per ticket is $3 to $14. That's the 10x savings. The difference is the underlying technology. A purpose-built classifier costs a fraction of running a large language model.

What to Do

If you're evaluating Zendesk's AI pricing:

Calculate your expected AI resolution volume. Look at your current ticket distribution. What percentage is automatable? Multiply by your monthly ticket count. That's your expected AI resolution volume.

Run the cost at committed and PAYG rates. Include your agent seat costs. Compare total cost to your current support spend.

Compare to alternatives. Get quotes from Intercom, Help Scout (which includes AI in its base pricing), and Supp. The per-resolution differences at your volume might surprise you.

Consider the architecture. Zendesk's AI is good, but it's an add-on to a help desk. If you want AI-first support (AI handles most volume, humans handle the rest), a tool built for that from the ground up might be a better fit than bolting AI onto a legacy ticketing system.