ROI of less than 5 percent? Why you should stop paying for “AI-powered” features immediately – Image: Xpert.Digital
Paying for results, not access: How results-based pricing is changing the SaaS market
AI licenses: A billion-dollar boondoggle: The new pricing model that is now causing panic among software giants
Billions are currently being poured into artificial intelligence, yet disillusionment is growing in boardrooms. The reason is not a technological, but a structural failure: outdated pricing models. Those who pay for autonomous AI agents and intelligent workflows using the same per-seat licenses (per user) or purely based on consumption as for traditional software are often only funding the hope of efficiency – without any guarantee of measurable added value. Studies show a dramatic failure rate for AI projects and spiraling, uncontrolled costs in business units. But the SaaS market is facing a tectonic shift: the era of outcome-based pricing is dawning. The following article examines why paying for mere access is obsolete, why many providers are resisting the change, and how smart companies can radically shift the dynamics of negotiation to their advantage in 2026.
Related to this:
Stop paying for AI that proves nothing
Those who cannot measure results only finance the hopes of others
The most uncomfortable silence in any sales conversation about enterprise AI arises precisely when someone asks the following question: How much of your AI budget is tied to measurable business results? Not to features advertised as "AI-powered" on a product sheet. Not to license seats that have been provisioned even though the usage rate barely exceeds ten percent. But to actual results that appear in a quarterly report, a process time measurement, or an auditable improvement log. Anyone who answers this question with "not entirely sure" is in good company. And will pay a price that goes far beyond the obvious.
The pricing model of a bygone era
The per-seat licensing model emerged at a time when the value of software scaled directly with the number of human users. More Salesforce seats meant more salespeople logging activities. More Slack seats meant more teams communicating with each other. The relationship between access and value was never perfect, but its basic direction was understandable: The provider charged for access, and the buyer assumed the value would follow.
Artificial intelligence has fundamentally shaken this assumption. When an AI agent resolves a support ticket, extracts data from a contract, or reviews a compliance document, the value isn't created by a person sitting in front of a screen. It's created by a workflow that may not even have a direct user. Charging per seat for an AI capability is like charging per office for electricity: the unit of measurement has nothing to do with the unit of value.
Yet this is precisely the prevailing practice in the enterprise AI market: a user fee tacked onto an existing platform license, a flat annual subscription for a tool that may produce results the buyer cannot quantify. According to the Zylo 2026 SaaS Management Index, based on the analysis of more than 40 million SaaS licenses and $75 billion in managed spend, 78 percent of IT executives reported unexpected costs from usage-based or AI pricing models. This isn't a budgeting failure on the part of individual companies; it's a structural misalignment between how AI creates value and how vendors generate revenue.
To make matters worse, control over SaaS spending is increasingly migrating away from the IT department: According to the same report, business departments now control 81 percent of SaaS spending, while IT is directly responsible for only 15 percent. At the same time, spending on AI-native applications increased by 108 percent year-over-year, and by as much as 393 percent in large companies with over 10,000 employees. The growth is real. Controllability, however, is often not.
What results-based pricing actually means
Outcome-based pricing is simple in concept but complex in execution. The provider is paid when the buyer receives value, not when the buyer gains access or consumes tokens, but when a defined business outcome is achieved.
The difference between usage-based and outcome-based pricing is more significant than most evaluations acknowledge. Usage-based pricing—per token, per API call, per query—is superior to the seat-based model because it correlates with activity. But activity is not value. Thousands of API calls that produce unrealistic results or irrelevant extractions are worthless to the buyer. Usage-based pricing shifts the cost risk from the provider to the buyer without shifting the performance risk in the slightest.
Outcome-based pricing shifts both of these factors. The provider only earns money if the AI delivers something the buyer has defined as valuable before the engagement begins. This could be a document processed with a defined accuracy threshold, an automated workflow with a measurable reduction in cycle time, or a completed compliance audit with a traceable log. The outcome is specified, the measurement criteria are agreed upon, and the commercial relationship follows from this.
The prime example from real-world practice comes from Intercom: The company charges $0.99 per successfully resolved support ticket by its AI agent, Fin. Bessemer Venture Partners, in its 2026 AI Pricing Playbook, describes this approach as the gold standard for outcome-based pricing. The model works because the value is precisely definable: A ticket is considered resolved or not. The metric is binary, tamper-proof, and directly linked to a cost driver within the buyer's organization.
The underlying structural logic also explains why the model is easier to implement in certain areas than in others. Gartner has already predicted that by 2025, over 30 percent of enterprise SaaS solutions would include outcome-based components, compared to around 15 percent in 2022. Simon-Kucher & Partners found in a recent study that 86 percent of buyers prefer usage- or outcome-based pricing models to traditional seat licenses. The market is signaling a clear direction. The question is not whether, but how quickly.
The AI ROI gap: Billions spent without proof
The necessity of this transformation stems from data that should be uncomfortably familiar to the business units responsible for AI. A comprehensive study by the RAND Corporation documented that more than 80 percent of all AI projects in companies fail without delivering the promised business results—a failure rate twice as high as that of traditional IT initiatives. MIT researchers, in a separate report, found an even higher rate of 95 percent for generative AI projects that fail to deliver a measurable return on investment (ROI).
A Forbes study from 2025, which surveyed several thousand executives worldwide, paints a similarly sobering picture: Less than one percent of the global executives surveyed reported that their organization had achieved a significant return on investment (ROI) – defined as more than a 20 percent increase in profitability or cost savings. Only three percent reported a moderate ROI of between 10 and 20 percent. The vast majority – more than 53 percent – described returns between one and five percent. At the same time, 39 percent of the executives cited measuring ROI as one of their biggest challenges.
This measurement gap is not just an analytical problem. It's a structural incentive problem. If the supplier's revenue isn't tied to the buyer's results, neither side has a structural incentive to diagnose why an implementation isn't working. The supplier has made their money. The buyer has gained access. The fact that nothing measurable has happened is everyone's problem—and no one's priority.
The pattern repeats itself with a certain regularity: First, AI was purchased under pressure from public perception, without a clear definition of success. Then, internal dashboards were created, full of activity metrics with no connection to the profit and loss statement (P&L). And finally, the first contract renewal came – and no one could explain what they were actually paying for. Bessemer Venture Partners aptly puts it in its playbook: Soft ROI positioning, which still worked in 2025 under the motto "AI adoption at any cost," now clashes with the reality of the 2026 renewal cycle – and mere promises don't renew contracts.
Why providers reject the model – and what that reveals
The objections raised by providers against outcome-based pricing are predictable and revealing. The standard repertoire includes three arguments: outcomes are difficult to define, the buyer's internal readiness influences the results, and the provider cannot control all variables. All three objections are factually correct. None of them, however, constitutes a valid argument for continuing to pay for AI that produces no results.
Anyone who honestly analyzes these arguments will recognize the real signal: A vendor who refuses to link pricing to results is revealing their own confidence in their product. If AI works, results-based pricing is more profitable for the vendor, not less. They earn money with every successful implementation, the buyer becomes a reference customer with quantifiable results, and sales costs for the next implementation decrease significantly. Vendors who reject this model are often those whose product delivers impressive demos but only mediocre production results.
A critical counterpoint, however, deserves consideration. Parloa, an AI provider specializing in conversational AI, argues that while outcome-based pricing may appear to promote mutual interests, in practice it often translates the company's efficiency gains into revenue for the provider. If an AI agent performs so well that process costs decrease significantly, the provider participates disproportionately in this value under an outcome-based model—even though they may have contributed only a small portion to the underlying efficiency improvements. This tension is real and explains why many experts view hybrid models as the more pragmatic solution: a base fee that covers the platform and implementation costs, combined with outcome-based fees that scale with the value delivered.
The structural shift in the SaaS market
The resistance of many established providers to new pricing models can also be explained by the financial architecture of the classic SaaS business model. Seat-based pricing produced long, predictable remaining contract terms—the so-called Remaining Performance Obligation (RPO)—because customers signed multi-year contracts for a fixed number of licenses. Usage- and outcome-based models compress this planning certainty in two directions: Contract terms shorten because buyers are hesitant to commit to usage volumes they cannot predict. Furthermore, the ratio of committed to flexible spending shifts in favor of the buyer's flexibility.
The valuation consequences are immediate. In the first months of 2026, a massive revaluation in the software market triggered a decline that wiped out nearly one trillion US dollars in market capitalization for software companies. The SaaS benchmark index fell by 6.5 percent throughout 2025, while the S&P 500 rose by 17.6 percent. The median revenue multiple for software companies plummeted from over seven times to below five times in just over a year. In contrast, companies that implemented hybrid pricing models reported 38 percent higher revenue growth and 38 percent higher net revenue retention than pure subscription providers, according to research by LEK Consulting.
Bloomberg predicts that subscription-based pricing could decline from its current 60 percent to around 30 percent of all software models within a decade, while outcome-based models increasingly fill the vacated space. Gartner estimates that 70 percent of companies will prefer usage-based pricing models to seat-based models by the end of 2026. The direction of this shift is not ambiguous; only the speed remains unclear.
🤖🚀 Managed AI Platform: Faster, safer & smarter to AI solutions with UNFRAME.AI
Managed AI Platform - Image: Xpert.Digital
Here you will learn how your company can implement customized AI solutions quickly, securely and without high entry barriers.
A managed AI platform is your all-inclusive, worry-free solution for artificial intelligence. Instead of dealing with complex technology, expensive infrastructure, and lengthy development processes, you receive a ready-made solution tailored to your needs from a specialized partner – often within just a few days.
The key advantages at a glance:
⚡ Rapid implementation: From idea to ready-to-use application in days, not months. We deliver practical solutions that create immediate added value.
🔒 Maximum data security: Your sensitive data stays with you. We guarantee secure and compliant processing without sharing data with third parties.
💸 No financial risk: You only pay for results. High upfront investments in hardware, software, or personnel are completely eliminated.
🎯 Focus on your core business: Concentrate on what you do best. We take care of the entire technical implementation, operation, and maintenance of your AI solution.
📈 Future-proof & scalable: Your AI grows with you. We ensure continuous optimization and scalability, and flexibly adapt the models to new requirements.
More information here:
Results instead of promises: How buyers successfully negotiate AI contracts
What buyers must demand before the next contract signing
Anyone currently evaluating AI platforms for businesses needs a solid framework to implement results-based pricing in practice. The first and most important step is defining the result before the evaluation begins – not as an abstract promise of efficiency, but as a specific, measurable metric linked to a business process the company already follows. This could include documents processed daily, the average review cycle time, the error rate in data extraction, or the throughput of compliance checks. If such a measurement isn't possible with the existing infrastructure, it must first be built, or a different starting point must be chosen.
The second step is a proof-of-value period on the company's own data. Not a sandbox demo on sample data carefully prepared for presentation purposes. An implementation in their own environment, connected to their own systems and running on the documents and workflows actually used in production. Companies that structure evaluations this way avoid the performance cliff that brings most AI programs to a standstill after early successes—because they have already validated production performance before committing to a budget.
The third step concerns the contract structure itself: pricing that scales with value, not consumption. The ideal structure is a base commitment covering the platform and implementation costs, supplemented by performance-based fees that scale as the AI generates measurable results. This provides the provider with predictable revenue for their implementation efforts, while contract growth is tied to the growth of value for the buyer. The buyer's risk is limited. The provider's potential is unlimited—but tied to performance.
The fourth step, which is often overlooked, is the vendor's responsibility for the implementation timeline. If pricing is results-based, but implementation takes nine months before any results are measured, the model is results-based in theory but a classic waste of time and materials in practice. The platform should be in production within days, not months, so that results measurement begins quickly enough to inform a sound procurement decision within a single budget cycle.
The renewal test: What distinguishes 2026 from 2025
The AI contracts that will last for their first renewal in 2026 and 2027 are the ones where someone can point to a number and say: That's what we got. No dashboard full of activity metrics. No usage report. A result that maps to the business case that justified the purchase.
This scenario is playing out right now. In spring 2026, Salesforce reported $800 million in Agentforce ARR (Annual Recurring Revenue) based on 29,000 results-based deals—a data point that demonstrates the model's commercial viability at scale. On the other side of the table, buyers are increasingly accompanied in renewal discussions by CFOs demanding hard ROI evidence and sustainable unit economics. The AI promise market, which was lavishly funded in 2023 and 2024, is now colliding with the AI results market, which will be settled in 2026.
The advantage of results-based pricing extends beyond mere commercialization. The model acts as a structured imperative for the kind of disciplined implementation that most AI programs skip. When the provider is paid only for results, every discussion about data quality, integration architecture, user acceptance, and process design takes place before deployment—not after the first failed quarterly review. The incentive for thorough preparation is not moral, but financial. This is by far the more reliable mechanism.
Structural implications for the company
Results-based pricing is more than just a commercial model. It transforms the internal organizational logic on both sides of the contract. On the supplier side, this model means that the ability to measure results must become part of the product – and not merely an afterthought for the customer success team. Suppliers who take this seriously build dashboards that show the buyer the delivered value in real time: time saved, improved quality, reduced risk. This visibility itself becomes a differentiator in a market where technological capabilities are becoming increasingly homogeneous.
On the buyer side, the model requires an upfront investment in measurability, which many organizations shy away from. Those who haven't systematically tracked process times can't agree on cycle reduction as a contractual metric. While this may initially sound like an obstacle, it's actually a useful filter. Organizations unable to define metrics for results-based contracts are generally also unable to successfully scale AI implementations—regardless of the pricing model. The measurement requirement forces the level of operational maturity that would be essential for productive AI use anyway.
Bessemer Venture Partners' playbook succinctly summarizes the core logic: AI doesn't monetize access. It monetizes results. Companies like Intercom, EvenUp, and Leena AI are aligning their entire organizational and sales models with the work delivered: resolved tickets, completed documents, and finalized reviews. The winners will charge for what their AI generates—not for what it costs or what it grants access to. The metric for calculation isn't simply a billing decision. It's a commitment to what you value, what the system is worth—and what you're willing to prove with your returns.
The power imbalance and who uses it
Anyone who understands the power dynamics in the current AI procurement market will recognize a temporary asymmetry favoring well-prepared buyers. Competition among AI providers has become extremely intense in several categories, while renewal rates for pilot programs are under pressure. Providers who were selling with mere promises in 2025 are now negotiating extensions with customers who want to see tangible results. This creates a negotiating position that didn't exist in 2024.
Buyers who now enter procurement negotiations with clear deliverable definitions, a proof-of-value framework, and a hybrid contract structure are in a significantly stronger negotiating position than those who arrive with only a functional specification and a rough usage estimate. The data—78 percent unexpected costs, 80 percent project failures, less than one percent significant ROI—provides them with the strongest argument. The methodology provides the tool.
This is especially true for medium-sized and large companies that are making significant expenditures on AI-native applications without having built the corresponding governance infrastructure. The Zylo report shows that spending on AI-native applications in large companies has increased by almost 400 percent—often via employee credit cards and expense reports—before IT teams can even react. The so-called shadow AI effect is not a fringe phenomenon but a structural feature of the current adoption cycle, which will become fully visible during renewal negotiations in 2026 and 2027.
Beyond pricing: The broader maturation period
What's happening in the AI procurement market isn't just an isolated price phenomenon. It's the maturation of a technology, marking its transition from experimental to production mode. The Google Cloud AI ROI 2025 report, based on a global survey of more than 3,400 business leaders, describes a new stage of AI maturity—the so-called "agentic age"—in which AI agents operate autonomously within defined parameters to deliver measurable business results. The 88 percent of agentic AI leaders who reported concrete returns in this study differ from the majority primarily in one key aspect: their ability to precisely measure results and align them with strategic goals.
Results-based pricing is the commercial expression of this maturity. It presupposes what mature AI implementations already require: clear process definitions, high data quality, a clean integration architecture, and measurement tools directly linked to business outcomes. Companies that take this path will pay less for hope and more for impact. This isn't a romantic vision of a fairer technology economy. It's a sober description of which contract structures will survive the next renewal cycles.
The real question for buyers is no longer whether results-based pricing is the right direction. Gartner, Bloomberg, Simon-Kucher, Bessemer Venture Partners, and the purchasing preferences of 86 percent of buyers all point in the same direction. The crucial question is whether their own procurement process can be adapted quickly enough to leverage the negotiating position this maturation phase offers in the short term – before the market consolidates again and suppliers can once more dictate the terms.
🎯🎯🎯 Data-driven B2B industry hub as a quasi-in-house solution
The quasi-in-house solution: How Xpert.Digital closes operational gaps in B2B marketing and sales – Smart Content-Driven Business - Image: Xpert.Digital
Xpert.Digital is a data-driven B2B industry hub led by Konrad Wolfenstein . The company acts as an external, quasi-in-house solution for industrial partners, closing operational gaps in marketing, content, and sales – without requiring additional resources on the client side.
More information here:
Your global marketing and business development partner
☑️ Our business language is English or German
☑️ NEW: Correspondence in your native language!
Konrad Wolfenstein
I and my team are happy to be available to you as your personal advisor.
You can contact me by filling out the contact form here wolfenstein@xpert.digital:or simply call me at +49 7348 4088 965. My email address is
I'm looking forward to our joint project.
