Skip to content
AI-accelerated delivery · You pay when it works
Plano, TX · MunichAccepting Q4 2026 briefs
Results as a Service

What is Results as a Service?

Results as a Service is a way of buying AI where you pay for each verified piece of finished work, such as a resolved support request or a matched invoice, instead of paying for seats, hours or a flat subscription. It is also called outcome-based or per-result pricing. It is easy to describe and harder to contract for, because both sides have to agree on what counts as a result.

Updated October 2026
Tell us the workflow →See how we ship →
The problem

Why buyers started asking to pay for results.

01

Seats stop matching the work

When an agent does work people used to do, the number of logins no longer tracks the value. Gartner said in July 2026 that up to $234 billion of enterprise application spend is exposed as agents take on seat-based work.

02

Usage bills include failed attempts

Per-message and per-action pricing charges each time the AI runs, so a request that fails and goes to a person is paid for twice.

03

Hours reward slow projects

Time-and-materials contracts pay the vendor more when a build drags on, which is the opposite of what the buyer wants.

04

Pilots with no unit of value

Without an agreed result to count, a pilot can run for months and nobody can say what it returned.

The short answer

Results as a Service in two paragraphs.

With Results as a Service the price is tied to finished work. A support agent is paid per resolved conversation and an invoice agent per invoice matched and approved. Work the AI fails at, or hands to a person, is usually not billed, so the buyer pays more when the AI does more of the work correctly and less when it does less.

The idea has several names. Foundation Capital called the shift "service-as-software" in April 2024, Sequoia wrote that AI companies now sell work in October 2024, and HubSpot co-founder Dharmesh Shah described Results as a Service (RaaS) in March 2025. The phrase itself is older: SAS used it in 2017 for an analytics service. The letters RaaS also stand for ransomware, robotics and recovery as a service, so spell the name out in proposals and contracts.

Who charges per result

Published per-result prices for AI agents, October 2026.

Almost every published per-result price is in customer service. These are list prices from each vendor's own pages, checked on October 18, 2026. Our full comparison, including usage and per-user prices, is on the AI agent pricing page.

VendorWhat one result isPublished price
HubSpot Customer AgentA resolved customer conversation$0.50 per resolved conversation
Fin (formerly Intercom)An outcome, at most one per conversation$0.99 per outcome
Gorgias AI AgentAn interaction resolved without a person$0.90 to $1.00 per interaction
Zendesk AI agentsA verified automated resolution$1.50 committed, $2.00 pay as you go
Salesforce Agentforce Help AgentA question that needs no human escalation$2 per resolution, in packs of 1,000
ChargeflowA chargeback disputed and won25% of the amount recovered
How common it is

Popular with buyers, still rare in signed contracts.

Buyers like the idea more often than they sign it. HFS Research found in July 2026 that 76 percent of customer experience buyers were open to outcome-based or gainshare pricing, while only 5 percent had such a contract. Among software companies, Kyle Poyar's 2025 State of B2B Monetization survey found 5 percent used outcome-based pricing as their main model.

Services firms are moving faster than software vendors. TCS told Reuters in August 2026 that about 80 percent of contracts in its finance, HR and business services segment now use outcome measures, and TP said in July 2026 that about 7 percent of its revenue comes from contracts linked to revenue or savings.

Finance teams remain cautious about variable bills. In Futurum's second-half 2026 survey of enterprise software buyers, per-seat was still the most preferred model for separately billed AI at 42.3 percent, with outcome-based at 21.1 percent. Predictability is the usual reason, and it explains why most per-result contracts add a minimum, a cap or both.

Compared

Results as a Service next to the other pricing models.

ModelYou pay forWho carries the risk of poor resultsFits best
Per seatPeople with accessThe buyerTools people use directly, such as a copilot
Usage or creditsEach message or actionMostly the buyerVaried tasks with no single unit of value
Time and materialsHours workedThe buyerExploratory work with an unclear scope
Per resultEach verified finished itemMostly the vendorSteady volume with a result that can be checked
Risks

Where per-result pricing goes wrong.

Arguments over the definition cause most of the trouble. In an a16z buyer survey cited by BCG, 47 percent of buyers said they struggle to define clear, measurable outcomes, and Constellation Research expected "haggling over definitions" when Salesforce launched per-resolution pricing in June 2026. A resolution that only means the customer went quiet is a weak definition, and vendors selling other pricing models have pointed that out.

A low price today may not stay low. Gartner predicted in January 2026 that generative AI cost per resolution in customer service will exceed $3 by 2030, more than many offshore human agents. The contract should say when the price per result can change and which data the change will be based on.

Items the AI fails are not billed, but someone on your team still handles them, and someone still approves the results that pass. Any saving estimate has to include both, or it will look better on paper than it does on the payroll.

How QueryNow does it

How we apply Results as a Service.

We use the model for the months after launch. The build is $10,000, paid only when it passes the acceptance criteria you signed. After that you pay a fixed price for each result that passes its checks, set from the numbers in your acceptance test, with a $1,200 monthly minimum. Work that fails a check, or that your team finishes, is not billed.

The agent runs in your own cloud tenant, so model usage and hosting stay on your cloud bill and every result is logged where your finance team can read it. The Results as a Service page has the full terms and a calculator.

Sources

Sources for the facts on this page.

Prices and survey figures were checked on October 18, 2026. Vendor pages change often, so confirm current terms before you sign.

  1. Foundation Capital, AI leads a service as software paradigm shift (April 19, 2024)
  2. Sequoia Capital, Generative AI's Act o1 (October 9, 2024)
  3. Dharmesh Shah, Software business models in the age of AI: from SaaS to WaaS to RaaS (March 7, 2025)
  4. SAS, Results as a Service (2017 offering)
  5. HubSpot, Customer Agent and Prospecting Agent pricing change (April 13, 2026)
  6. Fin, pricing (checked October 18, 2026)
  7. Gorgias, pricing (checked October 18, 2026)
  8. Zendesk, pricing (modified October 9, 2026)
  9. CIO.com, Salesforce unveils AI Help Agent with pay-per-resolution pricing (June 25, 2026)
  10. Chargeflow, pricing (checked October 18, 2026)
  11. HFS Research, CX buyers want outcome-based pricing, yet their contracts still pay for effort (July 6, 2026)
  12. Growth Unhinged, 2025 State of B2B Monetization (June 4, 2025)
  13. Reuters via The Star, AI reshapes India's IT services sector contracts (August 21, 2026)
  14. Economic Times, outcome-based AI deals rising, says TP CEO (July 15, 2026)
  15. Futurum Research, The Pricing Split (August 17, 2026)
  16. BCG, Rethinking B2B software pricing in the era of AI (August 13, 2025)
  17. Constellation Research, Salesforce takes a run at outcome-based Help Agent pricing (June 26, 2026)
  18. Gartner, GenAI cost per resolution for customer service will exceed offshore human agent costs by 2030 (January 26, 2026)
  19. Gartner, US$234 billion in enterprise application software spend is at risk from agentic AI (July 1, 2026)
How it works

What every per-result contract has to settle.

A written definition

One sentence that says what a finished result is, plus what is excluded. Most disputes start here.

A check on every result

Zendesk verifies each resolution with a language model before billing it. For documents, the checks are usually rules, such as totals that must reconcile.

Handoffs and reversals

Most vendors do not bill when the AI hands work to a person. Fin deducts a resolution if the customer comes back for more help.

A minimum and a ceiling

Fin's standalone plan includes 50 outcomes a month and Salesforce sells Help Agent resolutions in packs of 1,000. A cap protects the buyer's budget.

Log access

The buyer should be able to open any billed result and see the checks it passed and who approved it.

A repricing rule

When the price per result can change, and which logged data a new price will be based on.

Industries served
Our method

From workflow to working tool in two weeks.

01

Scope and sign

Day one

We define the workflow, the deliverables, and the acceptance criteria, and sign an agreement on them before anything starts.

02

Build

Two weeks

We build the tool in your environment, with Claude Code and automated evaluation against your own data.

03

Pay when it works

On delivery

$10,000, due only after every criterion in the signed agreement is met. Nothing before that.

Compliance & security
Built to SOC 2, HIPAA and GDPR standardsEU AI Act-aligned deliveryMicrosoft Solutions Partner
Pricing
$10,000 to build, then a fixed price per passing result

The build is paid only after it passes your acceptance test. After launch you pay for results that pass their checks, with a $1,200 monthly minimum. Failed work is not billed.

How our per-result pricing works →

Tell us the workflow →
Related solutions
Frequently asked

Questions buyers ask.

What is Results as a Service?

Results as a Service is a way of paying for AI where the buyer pays a fixed price for each verified piece of finished work, such as a resolved support request or a matched invoice, instead of paying for seats, hours or usage. Work the AI fails or hands to a person is usually not billed.

Is Results as a Service the same as outcome-based pricing?

Mostly. Outcome-based pricing is the broader term and can include gainshare, where the vendor takes a share of savings or revenue. Results as a Service usually means a fixed price for each verified unit of finished work.

What does RaaS stand for?

In AI pricing, RaaS stands for Results as a Service. The same letters are used for ransomware as a service, robotics as a service and recovery as a service, so spell the name out in proposals and contracts.

How much does an AI result cost?

Published prices for an AI-resolved support conversation ranged from $0.50 at HubSpot to $2 at Salesforce in October 2026. For document work the price is usually set per contract, because the value of a result depends on how long the same work takes a person.

Who came up with Results as a Service?

Nobody owns the phrase. SAS used it in 2017 for an analytics service, HubSpot co-founder Dharmesh Shah wrote about it for AI in March 2025, and Foundation Capital and Sequoia described the same shift as service-as-software in 2024.

Is AI always cheaper per result than a person?

No. Gartner predicted in January 2026 that generative AI cost per resolution in customer service will exceed $3 by 2030, more than many offshore human agents. Compare any price per result with your own measured cost per item.

When is per-result pricing a bad fit?

When a good result is a matter of taste, when volume is very low, or when you cannot measure what the work costs today. In those cases a fixed price for the build is usually easier to agree.

Ready to ship AI that actually works?

Send us the workflow. We return a fixed scope, price, and acceptance criteria in 48 hours. You pay $10,000 only after it works.

Tell us the workflow →
Free paper

Buying AI by the result: a Results as a Service guide for operators and finance teams

How to define, check and price AI results, with published per-result prices and the contract terms to ask for.

Read it free →
Get the designed PDF