How AI agents are reshaping payment logic in the CX market
AI agents are transforming not just customer service, but the entire logic behind how companies pay for technology. For years, the most common pricing model in the customer experience market was based on seats — each human agent represented a paid license, and that made sense when people were at the center of every operation. But now, with AI agents resolving interactions autonomously, that math simply doesn’t add up anymore. If artificial intelligence can solve problems without human intervention, why keep paying for idle capacity?
That’s exactly the question pushing the market in a different direction: outcome-based pricing. The idea is simple — pay for the value delivered, not for the tools you’ve contracted. In practice, though, this shift brings some significant challenges, like defining what a real resolution actually is and making sure the metric works at scale. The topic has gained momentum over the past few months, with companies like Zendesk taking positions and Gartner analyses pointing to this trend as inevitable among customer experience technology vendors.
It’s worth remembering that Zendesk was already signaling this vision back in 2015, when it launched Relate magazine and publicly argued that customer experience wasn’t just an operational function, but a strategic engine for value and long-term relationships. A decade later, the industry is experiencing a similar inflection point — this time, about how to measure and charge for the value AI actually delivers.
The seat-based model has fallen behind
The traditional seat-based pricing model worked well for decades because there was a direct relationship between the number of agents and service capacity. Each person hired had their workstation, their login, their software license. Companies sized their teams according to demand and paid proportionally for platform access. It was predictable, easy to calculate, and simple to audit. However, the arrival of AI agents broke that proportionality in an irreversible way. A single virtual agent can handle hundreds of simultaneous conversations, something no human being could ever do. Sticking with the old model in this scenario means paying for something that no longer reflects operational reality.
The problem isn’t just financial. When the billing model doesn’t keep up with technological evolution, it creates the wrong incentives. Companies that pay per seat tend to measure success by the number of active licenses, not by the quality of interactions or actual customer satisfaction. This creates a misalignment between what the technology vendor delivers and what the buying company actually needs. In a scenario where AI agents resolve a significant portion of requests autonomously, keeping this model is like paying rent for an entire office when you only use one room.
Recent commentary from Gartner reinforces this perception. Analysts at the firm observe that vendors are already exploring pricing models tied to concrete deliverables and business outcomes, as AI agents become more prevalent in enterprise applications. At the same time, customers remain cautious about unpredictable billing models and want clearer alignment between cost and value actually realized. The pressure is coming from all sides, and anyone who insists on the old model will have a hard time justifying the investment to increasingly informed and demanding customers.
Outcome-based pricing — what this actually means in practice
When we talk about outcome-based pricing, we’re talking about a deep shift in the relationship between technology vendors and companies. Instead of charging for access to a tool, the vendor charges for the concrete outcome of each interaction. In the customer experience context, this could mean charging for each ticket effectively resolved, for each problem solved without human escalation, or even for satisfaction metrics achieved after the service interaction.
Tom Eggemeier, CEO of Zendesk, frames this shift in a straightforward way: as AI agents take on more of the customer service workload, customers should expect pricing to reflect the outcomes delivered. According to him, Zendesk is leading this transition because it believes customers should pay for the value they realize, not for the tools they use. The key point, in Eggemeier’s view, is ensuring that outcomes are clearly defined, measurable, and consistent across channels — so that the model remains predictable and scalable, and not just aligned in theory.
This emphasis on consistency is critical. Resolution can’t simply mean that a conversation ended or that a ticket was closed. A customer who abandons the journey after a bad interaction with the AI shouldn’t count as a success. True resolution means the customer’s underlying problem was actually solved — even if that required routing to a human agent for more complex or consultative interactions.
Another fundamental point is the question of transparency. For outcome-based pricing to work, both parties need to agree on how metrics are measured, who audits the numbers, and what the dispute mechanism looks like when there’s disagreement. This requires maturity on both the vendor and customer side. Platforms that invest in detailed dashboards, real-time reporting, and open APIs for integration with internal systems come out ahead in this race. Trust in the model depends directly on the ability to prove, with data, that the promised result was actually delivered. And here’s where an interesting component comes in: the more sophisticated AI agents become, the easier it gets to track and document every step of the interaction, which naturally favors this type of billing.
What happens when scale enters the picture
Outcome-based pricing works pretty straightforwardly when AI agents handle a handful of well-defined use cases. Things get complicated — very complicated — when those agents start covering hundreds of workflows, channels, and different customer journeys. The difficulty isn’t just in tracking activity, but in agreeing on which outcomes can be reliably attributed to AI agents versus broader process or experience changes. Each new automation raises questions about how resolution is defined, how it’s measured, and how pricing stays consistent without turning every deployment into a custom contract negotiation.
Spirit Airlines, a Quiq customer, offers a concrete example of how this works in real life. The airline used AI agents to migrate customers from voice calls to messaging within the same journey, allowing problems to be resolved more efficiently without forcing the customer to restart the interaction from scratch. The use case remains the same, even though the journey spans multiple channels and workflows.
Mike Myers, CEO of Quiq, reinforces this point: outcome-based pricing only works if the definition of resolution is consistent and measurable across all workflows and channels. This requires robust orchestration and observability so that companies can scale AI agents without turning each new use case into a separate negotiation.
As organizations deploy AI agents across dozens, hundreds, or even thousands of scenarios, durable definitions of resolution become absolutely essential. Without them, outcome-based pricing risks becoming inconsistent and difficult to operationalize at enterprise scale.
The strategic value goes beyond cost reduction
It’s tempting to look at this transition only through the lens of savings. After all, paying only when the AI resolves a problem seems obviously cheaper than maintaining dozens of fixed licenses. But the strategic value of this shift goes far beyond the financial. When a company adopts outcome-based pricing, it gains a much clearer view of where technology actually works and where bottlenecks still exist. Every dollar invested can be traced to a specific outcome. This enables smarter decisions about where to invest in automation, where to keep human agents, and how to design service flows that combine the best of both worlds. The customer experience stops being a murky cost center and transforms into an operation with measurable returns.
For technology vendors, the strategic value is also enormous. Companies that adopt this model demonstrate confidence in the quality of their AI agents. It’s almost like telling the market — our technology is so good that we’re willing to make money only when it delivers real results. This positioning works as a powerful competitive differentiator, especially in a market where many platforms are still struggling to prove that their AI solutions are truly effective and not just glorified chatbots with generic responses. The bet on outcomes aligns the interests of those who sell and those who buy, creating a much healthier and more sustainable dynamic in the long run.
There’s also a positive side effect that few people are discussing: the pressure for continuous quality. When the vendor only bills if the AI agent resolves the problem, they have a direct economic incentive to constantly improve the technology. It’s not enough to launch the product and hope the customer figures it out. You need to refine language models, adjust conversation flows, train the AI with real data, and monitor satisfaction indicators. This virtuous cycle benefits the entire chain — the vendor evolves its product, the contracting company receives an increasingly better service, and the end consumer has a customer experience that actually works.
Hybrid models are likely to stay in the game
Despite all the movement toward outcomes, this doesn’t mean outcome-based pricing will replace all other models overnight. Hybrid approaches — combining seats, usage, and outcomes — will likely persist as vendors and customers seek to balance predictability with value alignment. What’s changing is the economic logic behind the decision. As AI agents execute more workflows and resolve more well-defined problems, customers will naturally expect pricing to reflect those outcomes.
In a sense, this transition echoes the evolution signaled by Zendesk’s Relate magazine back then. At that moment, the conversation shifted from operational efficiency to customer experience as a business strategy. Now, the conversation is shifting from paying for capacity to paying for outcomes. It’s a natural shift, but one that requires preparation.
The challenges that still need to be solved
Despite the optimism around outcome-based pricing models, there are real obstacles the market is still trying to overcome. The first one is standardization. Today, each vendor defines resolution in a different way. For Zendesk, it might be closing a ticket without it being reopened within a specific time period. For another platform, it might be the customer’s explicit confirmation that their problem was resolved. Without a market-accepted standard, it becomes difficult to compare vendors and make purchasing decisions based on objective data. Organizations like Gartner have already started working on reference frameworks, but we’re still far from having something universally adopted.
Another important challenge is contractual complexity. Outcome-based pricing requires far more detailed contracts than the traditional seat-based model. You need to define:
- Service levels and volume limits
- Adjustment mechanisms and periodic review
- Penalties for failure and clear audit criteria
- Consistent definitions of resolution by interaction type
- Dispute protocols when there’s disagreement about the numbers
This demands more sophisticated legal and commercial teams on both sides, which can be a barrier especially for smaller companies that don’t have that structure. Additionally, there’s the risk of undesirable behaviors — like AI agents that learn to close tickets too quickly without actually solving the problem, just to inflate resolution metrics. Ensuring that quality isn’t sacrificed in the name of quantity is a delicate balance that requires constant monitoring and frequent adjustments.
Finally, there’s the cultural question. Many companies are used to the predictability of the seat-based model — every month, the cost is the same, and the budget is easy to plan. With outcome-based pricing, the invoice amount can vary significantly from one month to the next, depending on interaction volume and resolution rate. This requires a mindset shift from financial and operations managers, who need to learn to work with variable costs in an area that traditionally had fixed expenses.
What to expect going forward
The customer experience market is at a transitional moment that resembles other major technological turning points. Just as cloud computing changed the way companies consume infrastructure — moving from on-premise servers to on-demand models — AI agents are forcing a complete reassessment of how the value of customer service is measured and charged. Companies that start familiarizing themselves with these new pricing models now will have a significant advantage when the paradigm fully consolidates.
Adaptation will happen, but it requires time, education, and above all, trust in the data that supports the model. The important thing is to closely follow how vendors like Zendesk, Quiq, and others are implementing these changes, evaluate the use cases that make sense for each operation, and understand that pricing, in this new scenario, is no longer just an administrative matter. It has become a first-order strategic decision — just as important as choosing the technology itself 🚀
