A practical guide to Zendesk automated resolutions in 2025

If you're in customer support, you've probably heard the buzz about AI automation. The big pitch is often "value-based pricing," where you only pay when the AI actually solves a problem. Zendesk’s model, built around what they call "automated resolutions," is their version of this. It's meant to tie your costs directly to the results you get.

But if you’ve started looking into it, you might be feeling a little confused. This new pricing has left a lot of support leaders scratching their heads over unpredictable costs, a fuzzy definition of what a "successful" resolution even is, and the headache of trying to forecast any real return on investment.

Let's cut through the noise. We're going to break down exactly how the Zendesk Automated Resolutions model works, point out the practical snags you might hit, and show you a more straightforward way to scale your AI support without the financial guesswork.

What are Zendesk Automated Resolutions?

In a nutshell, Zendesk Automated Resolutions are how Zendesk measures and bills you for using their AI agents. The idea seems fair on the surface: you only pay when the AI handles a customer's issue from start to finish, without a human agent needing to jump in.

An illustration of an AI agent working within the Zendesk platform to resolve customer issues.

But what does Zendesk count as "resolved"? This is where things get a bit murky. According to their own documentation, the rules change depending on where the conversation is happening. On messaging or email, a resolution might be counted if a customer replies with something like "Yes, that fixed it." Simple enough.

But it’s also counted if a customer just clicks a help article you sent and then doesn't respond for 72 hours. Even if they never actually confirmed their issue was solved, the ticket can be marked as resolved.

To make it even more confusing, Zendesk mentions they use a large language model (LLM) to "verify" that the resolution was relevant. This all happens on their end, creating a bit of a black box where you can't really see or audit what you're being charged for. It leaves a lot of room for interpretation, and for costs that don’t always match up with genuinely happy customers.

If you want to dive into the nitty-gritty, you can read Zendesk’s official article on the topic.

How Zendesk Automated Resolutions pricing works

Okay, let's talk about the actual costs. Zendesk's pricing for this has two main components: a small number of resolutions you get for free with your plan, and the much steeper costs you'll see once you go past that limit.

Included Zendesk Automated Resolutions in your plan

Most Zendesk Suite plans give you a monthly allowance of automated resolutions, but it's not a huge amount. Here's a quick look at their common plans:

It’s important to know that these plans have a ceiling of 10,000 allocated resolutions per year. If you're on an older or more basic Support plan, you might not get any included resolutions at all.

What happens when you go over the Zendesk Automated Resolutions limit

This is where your bill can start to swell. Once you burn through your monthly allowance, the costs add up fast.

• Pay-as-you-go: If you go over your limit, you'll be charged $2.00 for every single automated resolution.

• Committed Usage: You can buy resolutions in bulk beforehand for $1.50 each, but this means you have to accurately predict your future needs, which isn't always easy.

When you see you're getting close to your limit, you're stuck with a tough decision:

1. Keep it on and pay the overage: Your AI agents continue to work, but each extra resolution is another charge on a quickly growing bill. A sudden spike in customer tickets could lead to a massive, unplanned expense.

2. Turn it off and stop the fees: Your AI agent shuts down the moment the limit is reached. This saves you from the overage charges, but it dumps all incoming conversations right back onto your human agents, who can get overwhelmed in a hurry.

You can find these numbers on the Zendesk pricing page and in their article about managing your resolutions.

The hidden costs and limitations of Zendesk Automated Resolutions

Now that we've gone over the basics, let's get into why this model can be a real headache. The problems aren't just about the price tag; they're about the risk and uncertainty this kind of model introduces for your team.

What even counts as a "resolution"?

The biggest problem is the fuzzy definition of a 'resolution.' As one person pointed out on Reddit, the logic is 'really unclear.'

Reddit

You could get charged when a frustrated customer gives up and leaves the chat, or when they click an article link but never find their answer. If they don't complain within three days, Zendesk might just call it a success and send you the bill.

This makes it incredibly difficult to connect your spending to real value. Are you paying to make customers happy, or just to make them go away? With this model, it's hard to know for sure.

Unpredictable costs make budgeting a nightmare

This lack of clarity leads directly to financial instability. When you can't be sure what you're being charged for, you can't possibly forecast your monthly bill. Support volume is never perfectly consistent; a new feature release or a small site outage can send ticket numbers through the roof.

With a pay-per-resolution model, that kind of spike can completely wreck your budget. The alternative, turning off your AI, isn't much better, as it throws your team into fire-fighting mode and leaves customers hanging.

The "double-dipping" problem

What about when the AI tries to help but ultimately fails, and the ticket gets passed to a human agent? In many situations, you end up paying for both. You get charged for the failed AI attempt, and you still have to pay your agent's salary and their Zendesk seat license.

You're literally paying for failure. A good AI model should be set up so you only pay for what actually works. Instead, a bungled AI interaction just becomes an extra tax on a resolution that a human had to handle anyway.

Complicated features and add-ons

Finally, getting access to Zendesk's most powerful AI features isn't as simple as just buying a plan. You often need to buy expensive add-ons like "Advanced AI agents" or "Copilot" to get the functionality you actually need.

This just adds another layer of cost and complexity, making it tough to figure out what your total investment will be. You end up trying to navigate a maze of plans and add-ons, all while keeping an eye on your unpredictable resolution bill.

A simpler, more predictable alternative

If all of that sounds a bit stressful, you're not alone. The good news is, there's a better way to do this. Modern AI platforms like eesel AI were designed to fix these exact issues. Instead of a complicated rip-and-replace project, eesel AI connects directly to your current helpdesk, including Zendesk, to give you powerful automation without the financial anxiety.

Here’s how it offers a much simpler and more predictable alternative.

Predictable pricing with no surprise fees

This is the biggest difference. eesel AI works on a simple, flat monthly subscription based on the number of AI interactions you expect. There are no per-resolution fees and no surprise overage charges. Your bill is the same every single month. That means you can automate more support without worrying that a busy week will turn into a massive, unexpected invoice. It makes budgeting simple and straightforward.

A look at eesel AI's transparent, flat-rate pricing page, which helps teams budget effectively.

Test it out with a risk-free simulation

Not sure what your ROI will be? eesel AI has a simulation mode that lets you test your AI on thousands of your own past Zendesk tickets before you ever show it to a single customer.

This isn't just a generic demo. It’s a real-world test drive that gives you a solid forecast of your potential resolution rate and how much you could save. You get to see exactly how the AI would perform on your tickets and build a business case on your own data, taking all the guesswork out of the equation.

The eesel AI simulation mode, where you can test the AI's performance on past tickets to forecast ROI.

You’re in full control of your workflow

Unlike Zendesk's more rigid system, eesel AI gives you fine-grained control to selectively automate only the types of tickets you're ready for. You can start small by automating simple questions like "where's my order?", show the value to your team, and then slowly expand the AI's responsibilities as you get more comfortable. You’re always in the driver’s seat.

eesel AI’s customization rules, giving you full control to selectively automate specific workflows.

Here’s a quick comparison of the two approaches:

Move beyond Zendesk Automated Resolutions to predictable outcomes

While the idea of paying for results sounds good, Zendesk's automated resolution model brings too much uncertainty with its vague definitions and unpredictable billing. True support automation shouldn't feel like a gamble every month.

Platforms like eesel AI offer a much clearer path. With transparent, flat-rate pricing, powerful simulation tools to prove the value upfront, and total control over your automation, you can finally scale your support with confidence.

Ready to stop guessing about your AI ROI? Sign up for eesel AI and run a free, no-risk simulation on your own Zendesk tickets to see exactly how much time and money you can save.