Enterprises must break legacy contracts to unlock GenAI return on value

Nothing in Generative AI is ‘business as usual.’ GenAI is underwriting loans, reshaping pharma compliance, and refactoring codebases at an industrial scale. Yet many enterprises still use outdated commercial models of hourly rates, fixed scopes, and endless change orders, even as the link between billed hours and value collapses.

In HFS’ Horizon study, Generative Enterprise Services Horizon 2025, we gathered information from 40 key service providers in the GenAI domain. On average, 30% of 2024 GenAI engagements already include outcome, consumption, or risk-sharing components. Providers expect this to exceed 50% by 2026.

An EVP of Data and AI at a leading service provider explained why commercial models are shifting: “We’re no longer contracting for people. We’re contracting for digital labor. We’re paying for the performance of the model, not the hours of a person. It’s the single biggest shift we’re seeing in how enterprises buy AI-driven services.”

However, enterprise buyers are not satisfied with the commercial models included in service providers’ GenAI deals (see Exhibit 1).

Exhibit 1: Enterprise leaders indicated lower-than-average satisfaction with creative commercial models from their service providers

Sample: N = 75 GenAI partners, 71 client references provided as a part of Generative Enterprise Horizon, 2025
Source: HFS Research, 2025

Service providers may push gainsharing or transaction-heavy pricing for AI/agentic solutions. However,  buyers are unsatisfied and strongly resist for these reasons: decades of experience with software and services pricing that prioritize predictability, a deep psychological preference for certainty and fairness, and risk aversion tied to career consequences, not just financial waste. Providers must abandon purely rational pricing assumptions and strategically leverage behavioral insights, anchoring on outcomes rather than effort, transparently addressing fairness, decoupling payment from incremental usage, and carefully mitigating loss aversion.

Boards and CFOs must also understand that contracting GenAI like traditional IT leads to overspending on experimentation, rigid change requests, and losing top talent who expect shared value. GenAI’s dynamic and automated nature demands non-traditional commercial models that enable real-time adaptation, attract strategic partners, and deliver meaningful ROI in fast-evolving environments.

Key factors driving the shift to non-traditional models in pricing GenAI deals

GenAI can create economic value faster than human staff hours can track—contracts need to change or be left behind. Also, commercial models are no longer procurement housekeeping. They are an operating model choice that can double or halve an organization’s GenAI ROI. The following six complex factors drive this disruptive shift.

• C-Suite ROI focus. Some boards prefer to release funds when outcomes are hard coded into the deals, such as cost reduction, revenue uplift, and efficiency gains. This leads to outcome-based or gain-sharing models.
• Sharing risk. Some enterprises expect service providers to share risks for scaling GenAI solutions instead of leveraging purely time and materials commercial models. The typical response to this hybrid risk-reward model (base fee and risk-reward sharing model). Additionally, the shared risk helps build trust.
• Token-based economics. The provider references unit pricing from leading GenAI platform vendors such as OpenAI, Anthropic, and Google. This is triggering transaction-based / consumption-pass-through-based / metered platform charges.
• Digital Labour and automation. Tasks can get highly automated as we move toward agentic AI solutions. For e.g., one GenAI engineer can now orchestrate digital labor that replaces five to ten full-time equivalent (FTE) coders. This leads to digital labor contracts that price the bot, not the human FTE, converting operating expenses (OpEx) to utility.
• Speed of change. Proof of concept (POC) cycle time is falling sharply, triggering catalogue-based pricing, fixed-scope sprint pricing, or subscription for GenAI-as-a-service accelerators.
• Reputational risk > technical risk. Governance failures in GenAI can cause substantial reputational damage to brands. Thus, buyers want performance to be measured, monitored, and visible through transparent KPI dashboards.

GenAI solutions might still progress under legacy contracts, but the value could be limited. To realize its full potential, enterprises must adopt commercial models designed for experimentation, shared risk, and measurable outcomes.

Overcoming enterprise roadblocks in the adoption of non-traditional commercial models

Replacing legacy contracts can be complex, particularly when they are well-established and widely accepted by stakeholders. Enterprise buyers should demand total cost of ownership (TCO) transparency, push back on gainsharing unless clearly fair and measurable, anchor pricing on outcomes and trust, and favor hybrid pricing models with clear governance levers. This section explores some of the key challenges and considered approaches to adopting new contracting archetypes.

KPI ambiguity remains a major barrier. There are usually endless debates on what really improved. Consider organizing a joint value-charter workshop, third-party audit, and shared dashboard to align all stakeholders on measurable outcomes.

Accounting and procurement inertia also hold back progress. Effort-based systems are hard-wired into the enterprise fabric. One way forward is to create a new value-contracting lane with pre-approved clauses to fast-track non-traditional deals.

Provider balance sheet fear adds complexity. Considerable at-risk exposure stresses quarterly margins. Enterprises should consider establishing hybrid models with a fixed base fee plus upside and downside risk, along with caps and corridors and governance controls to manage volatility.

Legal and compliance departments often act as brakes on innovation, especially in regulated industries, where firms worry about sharing proprietary data on outcomes. Solutions include clean-room data environments, privacy guardrails baked into the statement of work (SOW), and stage-based access.

Finally, cultural change is critical. Delivery teams are rewarded for utilization, not business impact. Consider revising the KPI hierarchy, embedding a product-owner from the client in the vendor pod, and including reskilling funds in contract fees.

A delivery lead at a product-engineering firm pointed to enterprise hesitation as a key factor slowing the shift toward more agile, value-driven contracting. He explained, “Eighty percent of our projects are still time-and-materials today, but the only brake on faster outcome pricing is client readiness.”

Contract models must evolve to match GenAI speed

Most forward-looking enterprises are pioneering new contract archetypes that tie fees directly to business impact instead of hours spent. Exhibit 2 reveals the rising hybrid and consumption-linked contracts across top GenAI service providers, highlighting the waning relevance of traditional T&M models.

The traditional time and materials (T&M) and fixed-price models are used for budget certainty, simple invoicing, and low contract overhead. Still, these might lead to misaligned incentives, scope creep, change-order warfare, and risk premiums in hidden costs. Non-traditional commercial models can mitigate these drawbacks, but adoption is not without friction. Exhibit 2 illustrates the potential challenges of various types of contracts, such as audit costs, hidden integration effort, bot ownership complexity, and tricky forecasting. Thus, before finalizing any new model, organizations and its partners must ensure that sound KPI and baseline definitions, advanced monitoring capabilities, and organizational maturity are in place.

Exhibit 2: The top contract models use cases, benefits, and challenges

Sample: N = 40 service providers, Generative Enterprise Services Horizon, 2025
Source: HFS Research, 2025

Real-world impact through contracts that deliver GenAI ROI

Below are some scaled projects of enterprise clients that successfully use innovative contracts

• A European bank’s multi-agent customer service chatbot slashed contact center costs by 30% in nine months through a hybrid cost-ops deal in which a base fee covers platform run and the provider earns a percentage of verified contact-center cost savings.
• A global logistics giant’s order-query automation cut query handling time by 45% with a risk corridor structure that included fees varied by +/-20% against turnaround-time service level agreement (SLA), plus share of hard-dollar compliance savings.
• A US health insurer’s claims-coding optimization helped it realize a $38M annual benefit by sharing 17% of net adjudication savings above baseline with its GenAI vendor.
• A global pharma freed up $50M in marketing capacity via subscription-plus-success fee models.
• A telecom provider’s GenAI sales agents helped save $2M annually and improve revenue by 11% through a digital labor deal paying per lead and service provider funding agent stack.

The Bottom Line: To unlock GenAI ROI, enterprise leaders must shift from effort-based to outcome-based contracting. This requires rethinking how they define, govern, and reward value or risk being left behind.

In GenAI, how companies charge shapes how they build and operate. Enterprises’ GenAI success will not be decided by who writes the best prompt; instead, it will be decided by how you buy. To lead in GenAI, stop asking, ‘How much per hour?’ and start asking, ‘ What value will this create?’ That’s how you unlock ROI, retain talent, and stay ahead of disruption.

Further reading:

Flat-rate pricing isn’t rational but psychological—that’s why it wins

Services-as-Software and psychology biases break services pricing

How to avoid agentic sticker shock