The outcome based license model: charging for the result, not the attempt

This is the fourth article in our series on B2B software license models. It follows earlier pieces on the subscription license model, the usage based license model, and the perpetual license model.

Of the four, outcome based licensing charges for a completed result, not for time, seats, or usage. Therefore, the vendor is only paid when the customer gets what they came for.

Key takeaways

• Outcome based licensing is a license model that charges for a completed result, not for time, seats, or usage, so the vendor is only paid when the customer gets what they came for.
• Intercom’s Fin AI agent popularized outcome based pricing in B2B SaaS support software, charging close to one dollar per resolved ticket rather than a per-seat or per-conversation fee.
• Defining the outcome precisely, with a measurable trigger, a baseline, and clear exclusions, is the single hardest part of running this license model, and most disputes trace back to a fuzzy definition.
• Vendors get the tightest alignment between price and value of any license model in this series, but they also absorb the cost of every failed or repeated attempt that happens before an outcome is reached.
• Customers get the cleanest ROI story of any pricing model, paying only for results, but need contractual clarity on what counts as a result before disputes happen at renewal.

Where the outcome based license model came from

Charging for a result rather than an effort is not a new idea. Legal contingency fees and performance based advertising, cost per acquisition, and cost per lead have worked this way for decades.

What changed recently is that B2B software, specifically SaaS products, became capable of completing discrete units of work end to end. This shift, driven by AI agents that can resolve a support ticket or draft a contract clause without a person in the loop for every step, turned “outcome” from a marketing word into something a billing system could detect and charge for.

Intercom is the company most credited with proving this approach at meaningful scale in B2B SaaS. Its Fin AI agent charges roughly one dollar per resolved support conversation. This price point is built entirely around the outcome, ticket resolved, rather than around seats, messages sent, or hours of agent time.

Salesforce took a related but different path with Agentforce, initially pricing at two dollars per conversation. However, they ran into the same problem this article covers below: a conversation is a fuzzy unit. Since it can branch, loop, or drag on without a resolution, Salesforce introduced Flex Credits to price individual actions, such as updating a record.

Metrics: how an outcome based license model is measured

The core unit of measurement is the outcome itself. It needs to be defined precisely enough that both vendor and customer agree, without dispute, when it has occurred.

In a B2B context outside of support software, one clear example is a fraud prevention platform. Here, a billable outcome is explicitly defined as a transaction approved by the model, processed successfully, and not flagged as fraudulent within a 90-day window. Notice how much definitional work is packed into that single sentence: a transaction only counts if it clears three separate conditions, and the vendor must be able to prove all three before charging for it.

Beyond the per-outcome price, the metrics that matter are outcome success rate, the share of attempts that reach a billable result, and cost to serve per outcome, the vendor’s internal cost, largely compute and any human escalation, to produce one successful outcome. The gap between price per outcome and cost per outcome is the entire margin story for this license model, and it is the subject the next article in this series goes deeper on, because that gap is far less stable than it looks on a pricing page.

Pro tip: Before pricing anything on an outcome basis, write the definition of that outcome down as a contract clause. It should not just be a sentence in a pitch deck. If you cannot state the trigger, the baseline, and the exclusions in writing, you do not have an outcome based license model. Instead, you have a promise.

The vendor side: what an outcome based license model gets right and wrong

For the company selling it, outcome based pricing produces the cleanest sales narrative available: the customer pays only when they get the result they wanted, which removes almost every objection about paying for unused seats or wasted API calls. It also tends to command a premium price per unit compared with usage based pricing for the same underlying work, because the customer is buying certainty of result, not just access to a product that might produce one.

The cost is that the vendor now absorbs all of the risk sitting between an attempt and a successful outcome. If an AI agent needs three tries internally before it resolves a ticket, the vendor pays for all three attempts in compute cost but only gets paid once, on the successful one. This makes revenue recognition and margin forecasting genuinely harder than under any other license model in this series, because the vendor has to reliably instrument, log, and audit what counts as an outcome, and defend that definition when a customer disputes a charge.

The customer side: what an outcome based license model gets right and wrong

For the buyer, this is the easiest license model to justify internally. The finance case is almost self writing: “we pay one dollar for every ticket this system resolves, compared with several dollars in fully loaded agent time for the same ticket handled by a person.” There is no cost for attempts that do not work, and no cost for capacity sitting idle. It is also the pricing approach that best matches how customers already think about value in their own MEDDIC style qualification, tying spend directly to a business outcome rather than a feature list. Our earlier piece on B2B sales pipeline metricsgoes deeper on this connection between quantified business outcomes and deal qualification.

The risk sits in the definition itself. If “resolved” is defined loosely, for instance a support conversation the AI agent simply closed without confirming the customer was satisfied, the customer can end up paying for outcomes that do not feel like real outcomes to them. This is why contracts in this category increasingly include audit rights, sampling of resolved cases, and a rejection or credit mechanism for outcomes the customer disputes after the fact. None of this makes outcome based pricing a bad deal. It makes it a license model that requires more precise contracting than a simple per-seat subscription ever did.

Quick facts

• Intercom’s Fin AI agent charges close to one dollar per resolved support conversation, one of the clearest live examples of outcome based licensing in B2B SaaS.
• Salesforce Agentforce moved from a per-conversation price to Flex Credits priced per action, after finding that a conversation was too fuzzy a unit to bill reliably.
• HubSpot and Zendesk have entered the same B2B SaaS AI support category at roughly fifty cents and a dollar fifty per resolution, showing the license model converging into a defined price band.
• A precise outcome definition needs a measurable trigger, an agreed baseline, a time window, and explicit exclusions, written into the contract rather than left implicit.
• The gap between price per outcome and the vendor’s cost to produce that outcome is the single biggest margin risk in this license model.

Frequently asked questions

What is an outcome based license model? An outcome based license model charges a customer only when a specific, pre-defined result is achieved, such as a resolved support ticket or a completed transaction, rather than charging for time, seats, or general product usage.

Which company pioneered outcome based pricing in B2B SaaS? Intercom is widely credited with proving outcome based pricing at scale through its Fin AI agent, which charges close to one dollar per resolved support conversation rather than a seat or usage fee.

Why did Salesforce Agentforce move away from per-conversation pricing? A single conversation could branch, loop, or continue indefinitely without a clear resolution, making it an unreliable unit to bill against. Salesforce introduced Flex Credits, pricing individual actions such as updating a record, to create a more precisely measurable unit than a full conversation.

What makes outcome based pricing hard for vendors to manage? Vendors absorb the full cost of every attempt that happens before a successful outcome is reached, but only get paid for the successful one. This makes cost to serve unpredictable and puts pressure on margins whenever achieving the outcome takes more attempts, or more compute, than expected.

How should a customer protect themselves when buying outcome based pricing? Customers should insist on a written, specific definition of what counts as a billable outcome, including the trigger, the measurement window, and any exclusions, along with audit rights or a dispute mechanism for outcomes they believe were not genuinely achieved.

Outcome based licensing is the sharpest of the license models, and the least forgiving

Of the four license models in this series—subscription, usage based, perpetual, and outcome based—outcome based licensing gets the closest to the ideal that value based pricing has always promised. The customer pays for what they get. This is why it is spreading quickly through AI powered B2B SaaS products. A defined task can now be completed without a human in the loop for every step.

For a broader look at how these license models compare and where each one fits a given product and stage, see our overview article, B2B software license models: why structure beats value based pricing long term, and our companion piece on SaaS license model design.

It is also the license model with the least room for error in its own definition. A vague outcome clause creates disputes at renewal, and a vendor that has not modeled its true cost to serve per outcome can find its margins eroding quietly, deal by deal, long before it shows up in the quarterly numbers. The next article in this series goes deeper into exactly that problem: why an outcome as simple as “finish this invoice” is harder to price than it looks, especially once generative AI is involved.

If you are evaluating whether outcome based pricing fits your SaaS product, or need help defining outcomes that will hold up in a contract, reach out and let’s have a conversation.

Understanding the intricacies of outcome based pricing is essential for both vendors and customers. This pricing model, while beneficial in many ways, requires a clear understanding and agreement on what constitutes a successful outcome.