IT Pricing
What does agentic IT support cost? A guide to per-resolution pricing
Most IT software bills per seat or per subscription, which means you pay for capacity whether or not it is used. Outcome-based pricing works differently: you are charged only for finished resolutions, so cost tracks the value delivered. Dex is $1.99 per resolved ticket, with nothing else to pay and a $100 free credit to start, so the economics are simple to understand and easy to predict.
- Updated
- July 2026
- Read time
- 9 min read
- For
- IT leaders, MSP owners, finance and procurement
- Topic
- IT Pricing
In brief
- Most IT software is billed per seat or subscription, so you pay for licensed capacity whether or not it is used.
- Per-token or per-agent pricing charges for consumption or for AI agents running, which can be unpredictable and is not tied to whether work was actually completed.
- Per-resolution pricing charges only for finished resolutions, so cost aligns with value delivered.
- Outcome-based pricing makes the most sense when the vendor is confident enough in results to only get paid when the work is done.
- Dex, built by SysAid, is $1.99 per resolved ticket, with no per-seat fee and a $100 free credit to start.
Best for
IT leaders, MSP owners, and finance teams comparing how AI IT support tools are priced and which model aligns cost with value.
Grounded in how Dex, an autonomous IT engineer built by SysAid, is priced: $1.99 per resolved ticket with a $100 free credit.
What actually shapes the cost of agentic IT support?
The cost of agentic IT support is shaped less by a headline number and more by the pricing model behind it, because the model decides what you are paying for. Three models dominate: per-seat or subscription pricing, where you pay for licensed users; per-token or per-agent pricing, where you pay for consumption or for AI agents running; and per-resolution pricing, where you pay only for requests that are actually resolved. The same tool can look cheap or expensive depending on which model it uses and how much idle capacity or unfinished work you end up paying for.
The key question is not "what is the price per unit?" but "what is the unit, and is it tied to value?" A per-seat license bills you the same whether a user opens one request a year or fifty, so you pay for capacity that may sit idle. A consumption model bills you for activity even when that activity does not resolve anything. A per-resolution model only bills when a request is finished, so the unit of cost is also the unit of value. Understanding which model you are being quoted is the difference between a predictable bill and a surprising one.
Key takeaways
- The pricing model, not the sticker price, determines what you actually pay for.
- Per-seat pays for capacity, per-token pays for consumption, per-resolution pays for outcomes.
- Aligning the billing unit with a completed outcome is what makes cost predictable.
How does per-seat and subscription pricing work?
Per-seat pricing charges a fixed fee for each licensed user or a flat subscription for a tier of usage, regardless of how much work the tool actually performs. It is the most common model in IT software because it is simple to quote and predictable to budget. The trade-off is that you pay for capacity rather than results: a seat costs the same whether the person behind it raises many requests or none, so a large share of the spend can go toward idle capacity.
Per-seat pricing made sense for tools that a person uses directly, where a seat maps to a user getting value. It fits agentic IT support poorly, because the value is the work resolved, not the number of people licensed to submit requests. In a support context, request volume is uneven across users and across time, so a flat per-seat cost rarely matches the actual load. When volume is low, you overpay; when it spikes, the seat count you licensed may not reflect the surge in work. The cost and the value drift apart.
Key takeaways
- You pay per licensed user or per subscription tier, not per outcome.
- It is predictable to budget but bills for capacity whether or not it is used.
- Request volume is uneven, so per-seat cost rarely matches the real load.
How does per-token or per-agent pricing work?
Per-token pricing charges for the volume of AI consumption, such as the tokens a model processes, while per-agent pricing charges for each AI agent you run. Both tie cost to activity rather than to finished work. That can make spend hard to forecast, because a request that takes many reasoning steps costs more than a simple one, and a request that consumes resources but does not resolve anything still costs money. You are effectively paying for the machine to think, whether or not it finishes the job.
Consumption-based pricing is transparent about compute, but it pushes the risk of inefficiency onto the buyer. If an agent works hard on a problem and still escalates it, the tokens or agent-time are already billed. Volume is also difficult to predict in advance, so budgeting becomes an estimate rather than a commitment. For agentic IT support specifically, this model rewards activity, not results, which is the opposite of what an IT leader wants to buy. The buyer is asked to trust that consumption will translate into resolutions, without that link being guaranteed by the price.
Key takeaways
- Per-token bills for AI consumption; per-agent bills for each agent running.
- Cost tracks activity, not whether the request was actually resolved.
- Spend is harder to forecast, and unresolved work still incurs cost.
How does per-resolution pricing work?
Per-resolution pricing charges a flat fee for each request the engine actually resolves, and nothing for the seats, subscriptions, or activity behind it. The unit of billing is a finished outcome, so the bill is simply the number of resolutions multiplied by a fixed price. Dex uses this model at $1.99 per resolved ticket, with no per-seat fee and no subscription, and a $100 free credit to get started. Because the price attaches to a completed resolution, cost and value move together.
The defining feature of per-resolution pricing is that the vendor only gets paid when the work is done. That places the risk of unfinished or inefficient work on the vendor, not the buyer, which is a strong signal of confidence in the results. It also makes the economics unusually easy to reason about: multiply resolutions by the per-resolution price and you have the cost, with no seat math, no tier upgrades, and no consumption estimates. With Dex, a request that is not resolved is simply not billed as a resolution, so you are never paying for capacity that produced no outcome.
Key takeaways
- A flat fee is charged per resolved request, with no seat or subscription fees.
- The vendor is only paid when the work is finished, shifting risk off the buyer.
- Dex applies it at $1.99 per resolved ticket with a $100 free credit to start.
How software helps
Dex, an autonomous IT engineer built by SysAid, is priced entirely on resolutions: $1.99 per resolved ticket, no per-seat fee, and a $100 free credit to begin. Because Dex resolves L1 through L3 requests end to end inside Microsoft 365 and other systems rather than just routing them, the resolution is a real, completed outcome. That is what makes per-resolution pricing honest here: you pay for finished work, and the tool is built to actually finish it.
Why does per-resolution pricing align cost with value?
Per-resolution pricing aligns cost with value because the thing you pay for, a resolved ticket, is the same thing you are trying to buy. With per-seat or per-token models, cost is tied to capacity or activity, which are proxies for value that can drift far from the actual result. When the billing unit is a completed resolution, there is no gap between spend and outcome: every dollar maps to a request that was actually handled, and nothing is spent on idle seats or unproductive activity.
This alignment also changes the incentives. Under per-seat pricing, a vendor is paid the same whether the tool resolves much or little, so there is no direct financial pressure to improve results. Under consumption pricing, more activity can even mean more revenue, whether or not it helps the customer. Under per-resolution pricing, the vendor only earns when the customer gets an outcome, so the vendor is motivated to resolve more, faster, and more reliably. The buyer and the vendor want the same thing, which is the healthiest possible footing for a support relationship.
Key takeaways
- The billing unit and the value unit are identical: a resolved request.
- No spend goes to idle seats or activity that resolves nothing.
- Incentives align: the vendor only earns when the customer gets an outcome.
Per-seat vs per-token/per-agent vs per-resolution: a direct comparison
| Feature | Per-resolution | Per-seat / per-token |
|---|---|---|
| What you pay for | Each request actually resolved | Licensed users (seat) or AI consumption (token/agent) |
| Pay for idle capacity? | No, unused capacity costs nothing | Yes for seats; consumption bills even without a resolution |
| Predictability | Cost is resolutions times a fixed price | Seat cost is fixed; token or agent cost is hard to forecast |
| Aligned to value? | Yes, the billing unit is the outcome | No, tied to capacity or activity, not results |
| Who carries the risk of unfinished work | The vendor, who is paid only on resolution | The buyer, who pays regardless of the outcome |
| Vendor incentive | Resolve more and more reliably to earn | Fill seats or drive consumption, not necessarily outcomes |
What are the real trade-offs of per-resolution pricing?
Advantages
- Cost tracks value: you pay only for requests that are actually resolved
- No spend on idle seats or unused capacity
- The bill is simple to reason about: resolutions times a fixed price
- The vendor carries the risk of unfinished or inefficient work
- Incentives align, because the vendor earns only when you get an outcome
- A $100 free credit lets you see resolutions before committing spend
Limitations
- Variable cost means the bill moves with volume rather than staying flat
- You need a clear, agreed definition of what counts as a resolved ticket
- High-volume periods cost more, though they also deliver more resolved work
- It is a newer model, so it can require rethinking how IT support is budgeted
What should buyers evaluate when comparing AI IT support pricing?
These criteria help you compare pricing models on the factors that determine whether cost actually tracks the value you receive.
- What the billing unit is
- Confirm whether you are paying for seats, consumption, or resolved outcomes, because that determines what your spend is tied to.
- Whether you pay for idle capacity
- Check if unused seats or unresolved activity still generate cost. Per-resolution pricing charges nothing for either.
- Predictability of the bill
- Understand how the total is calculated. A resolutions-times-price model is simpler to forecast than consumption estimates.
- Definition of a resolved ticket
- For outcome-based pricing, make sure the definition of a resolution is clear and agreed before you commit.
- Hidden or additional fees
- Ask whether there are seat fees, subscription tiers, or setup charges on top of the headline price. Dex has none beyond $1.99 per resolution.
- A way to try before committing
- Look for a free credit or trial so you can see real resolutions before you spend. Dex includes a $100 free credit.
When does outcome-based pricing make sense, and when might another model fit?
Per-resolution / outcome-based
- You want cost to track the actual work completed, not licensed capacity
- Request volume is uneven or hard to predict across users and time
- You want the vendor to carry the risk of unfinished or inefficient work
- You value a bill that is simple to reason about and audit
Per-seat or consumption
- You want a flat, fixed cost that never varies with volume
- Usage is steady and predictable enough that a seat maps cleanly to value
- The tool is used directly by people rather than resolving work for them
- Your budgeting process strongly favors a single fixed line item
A buyer checklist for comparing agentic IT support pricing
Use this checklist to compare pricing offers on a like-for-like basis before you commit.
We know exactly what the billing unit is
Seats, tokens or agents, or resolved outcomes.
We understand whether idle capacity or unresolved activity is billed
Per-resolution pricing charges for neither.
We can forecast the bill from our own volume
For per-resolution, that is resolutions times the per-resolution price.
We have an agreed definition of a resolved ticket
So the outcome being billed is unambiguous.
We have confirmed there are no hidden seat, subscription, or setup fees
Dex is $1.99 per resolved ticket with nothing else to pay.
We have a way to try it before committing spend
Dex includes a $100 free credit to start.
Putting it all together: from problem to platform
Placeholder — a short paragraph framing the challenge and what a modern approach looks like, before outlining where automation, AI, and a purpose-built platform each play a role.
The challenge
IT leaders and finance teams want to know what agentic IT support will actually cost, but most pricing models make that hard to answer honestly. Per-seat billing charges for capacity that may sit idle, and consumption billing charges for activity that may not resolve anything. The challenge is finding a model where the cost is predictable, tied to results, and easy to defend to the people who sign off on the budget.
What good looks like
- Every dollar spent maps to a request that was actually resolved
- There is no charge for idle seats or unproductive activity
- The bill is simple to calculate and easy to audit
- The vendor only earns when the customer gets an outcome
Where automation helps
- Resolving requests end to end so each billed resolution is a real, finished outcome
- Keeping the billing unit tied to completed work rather than seats or activity
- Making spend move with demand, so quiet periods cost less
- Removing per-seat and subscription overhead from the cost structure
Where AI helps
- Investigating and resolving requests across L1 to L3 so more of the queue is genuinely finished
- Escalating only genuine judgment cases, which keeps unresolved work to a minimum
- Adapting to novel requests rather than failing and leaving work unfinished
- Improving over time, so a larger share of requests reaches a billable resolution
Where a platform fits
- Pricing purely on resolutions at $1.99 per resolved ticket, with no per-seat fee
- Offering a $100 free credit so buyers can see resolutions before committing
- Resolving L1 to L3 requests end to end so a resolution is a true completed outcome
- Being built by SysAid, whose ITSM technology is trusted by 3,000+ organizations
Placeholder — short, direct value statement
Placeholder supporting sentence. No jargon. One clear benefit.
See Dex pricingFrequently asked questions
Common questions about this topic, answered directly.
A resolved ticket is a request the engine has actually completed end to end, not merely routed or acknowledged. With Dex, resolution means the request has been carried out inside the connected systems, such as resetting access or provisioning an account in Microsoft 365, so the underlying issue is genuinely handled. Because Dex resolves L1 through L3 work rather than just suggesting answers, a billed resolution reflects finished work.
No. Per-resolution pricing charges only for requests that are actually resolved, so a request that is not resolved is not billed as a resolution. Dex targets resolving over 90 percent of IT requests autonomously, and genuine judgment cases escalate to a human rather than being counted as resolved. This is what shifts the risk of unfinished work onto the vendor rather than the buyer.
No. Dex has no per-seat fee and no subscription tier. The pricing is simply $1.99 per resolved ticket, with nothing else to pay. That is what makes the bill easy to reason about: it is the number of resolutions multiplied by a fixed price, with no licensed-user math on top.
Dex includes a $100 free credit to start, which lets you see real resolutions before committing any spend. Because each resolution is $1.99, the credit covers a meaningful number of resolved tickets so you can evaluate the results on your own requests first. It is a way to try outcome-based pricing with the vendor carrying the initial cost.
Because it ties cost to the outcome you actually want. Per-seat pricing bills for capacity whether or not it is used, and per-token or per-agent pricing bills for activity whether or not it resolves anything. Per-resolution pricing bills only for finished resolutions, so cost and value move together and you never pay for idle seats or unproductive work.
It is variable by design, but it is easy to forecast: your cost is the number of resolutions multiplied by the per-resolution price. Quiet periods cost less and busy periods cost more, but in every case the spend maps to work that was actually completed. For many teams that is more predictable than estimating token consumption or over-licensing seats for peak demand.
The bottom line
What agentic IT support costs depends less on the sticker price than on the pricing model behind it. Per-seat and subscription pricing bill for capacity that may sit idle; per-token and per-agent pricing bill for activity that may resolve nothing; per-resolution pricing bills only for finished resolutions, so cost tracks value and the vendor is paid only when the work is done. Dex, built by SysAid, applies this model at $1.99 per resolved ticket, with no per-seat fee and a $100 free credit to start, which makes the economics simple to reason about and squarely aligned with the outcome you are buying.
See it in action
Placeholder — one sentence describing what the viewer will see in a product walkthrough or demo session.