Rethinking managed services in the age of AI
A retailer adjusts pricing and promotions daily in response to customer demands. A bank experiences sudden spikes in transaction volumes tied to market volatility. A utility provider responds to seasonal fluctuations. A healthcare system adapts quickly to changing patient volumes, claims activity and regulatory requirements. Government agencies respond to changing funding mandates.
Across industries, organizations are becoming more dynamic, responsive and driven by rapidly changing conditions. Yet many of the service contracts supporting those organizations still rely on assumptions of predictability. Fixed scope, stable demand and pricing models were designed for a more stable operating environment.
Artificial intelligence (AI) is accelerating that pressure, with dramatic implications for how services are delivered, priced, governed and measured over time. For managed services providers and their clients, the assumptions that once made pricing and delivery predictable no longer consistently hold. But as uncertainty increases, so do the opportunities to rethink how value is created.
At CGI, we are seeing these trends emerge across industries as clients rethink how services should be priced, governed and measured in increasingly dynamic environments.
Why are managed services becoming less predictable?
AI is fundamentally changing where value is created and how that value can be priced. Tasks that once required hours can now take seconds, while the appetite for faster delivery, continuous innovation and real-time responsiveness grows just as quickly. AI is also accelerating how rapidly organizations can generate insights, deliver work and realize value. At the same time, faster execution is raising expectations for speed, flexibility and continuous responsiveness. As a result, many of the traditional relationships between productivity, cost and value are becoming harder to forecast.
Historically, managed services and consulting models were built on predictability, relying heavily on time-and-materials pricing where effort closely correlated with both cost and perceived value.
But AI is weakening that relationship. Activities such as coding, testing, support and operational management can now be completed with significantly less human effort, making traditional labor-based pricing models harder to align with how value is measured. This is driving growing interest in pricing approaches tied more closely to outcomes, business impact and speed of execution.
These developments are challenging the foundation on which managed services pricing was built. Traditional models rely on predictable demand, defined scope and measurable effort. Clients and providers are now being asked to define pricing models, savings targets and business outcomes before the full impact of AI is fully understood. This introduces uncertainty on both sides of the partnership, particularly as productivity gains, operational impacts and future demand remain difficult to forecast.
At the same time, AI is introducing new forms of operational and commercial volatility that are becoming increasingly difficult to predict and govern.
How can managed services adapt to greater uncertainty?
Amid this uncertainty, clients are increasingly looking for more flexible ways to structure, price and govern services. At CGI, we are responding by working with clients to develop more adaptive pricing and partnership approaches, including outcome-based, consumption-based, agent-based and hybrid models. However, no single model has emerged as the standard. Each introduces trade-offs, and none fully resolves the underlying challenge: uncertainty.
At the same time, business leaders are increasingly willing to pay not only for lower-cost execution, but for the ability to accelerate modernization, compress business cycles and realize business outcomes faster. Simply put, speed itself is becoming economically valuable.
The issue is not selecting the right pricing model. It is designing commercial partnerships that can operate effectively in uncertain conditions.
What new managed services approaches are emerging?
In practice, we’re advising clients to consider:
• Building for adaptability by introducing mechanisms to adjust scope, volumes and pricing over time
• Applying value-based models selectively to align pricing to outcomes where they can be clearly defined and measured
• Designing for value reallocation to reinvest savings into modernization, innovation and customer outcomes
• Increasing transparency and auditability through clearer baselines, metrics and visibility into performance and costs
• Enabling greater flexibility in capacity and consumption by aligning cost structures more dynamically to changing demand
Together, these points reflect a broader evolution in managed services, from static delivery models to more adaptive partnership structures. In this environment, competitive advantage will increasingly depend not only on cost efficiency but also on how quickly AI-driven capabilities can be translated into business outcomes with greater speed and at scale.
Why are traditional managed services models struggling with uncertainty?
AI is creating structural friction within traditional labor-based pricing models. As the amount of human effort required for activities such as coding, testing and operational support declines, pricing structures built around hours worked become harder to align with how value is delivered and measured. In many cases, organizations are achieving outcomes faster, even as the amount of human effort declines significantly. Increasingly, the value of AI is measured not only by efficiency gains but also by speed, responsiveness and accelerated outcomes.
Can AI enable a zero-incident operating model?
We are already helping clients rethink operational support models through AI-enabled approaches such as our DigiOps and AIOps Nova. By combining predictive operations, automation and autonomous incident management, these environments are increasingly evolving from reactive support models toward more self-healing operational structures. In some client environments, this has resulted in:
- Up to 70% fewer auto-generated tickets
- Autonomous resolution of approximately 20% of operational events
- 30–35% faster root-cause identification
- 28–45% fewer major outages through predictive modeling
- 18–25% operational cost savings through reduced manual intervention and technical debt
These capabilities allow internal teams to be redirected from operational firefighting and toward higher-value innovation and transformation initiatives.
Is there a standard pricing model for AI-enabled managed services?
Rather than converging around a single approach, we are seeing a growing willingness among clients to explore a broad mix of commercial structures. Some models prioritize measurable outcomes. Others focus on consumption, automation or shared value creation. Each offers advantages in specific situations while introducing new operational, financial and governance trade-offs.
When does outcome-based pricing work?
One of the clearest trends we are seeing with clients is growing demand for outcome-based pricing. Rather than paying for labor effort alone, business leaders increasingly want pricing structures tied more directly to measurable business results, operational improvements and defined service outcomes.
In some environments, pricing may be tied to resolved support tickets, transaction volumes, automated workflows or operational performance targets. In others, clients are showing growing interest in gain-sharing models tied to business metrics such as reduced claims-processing times, faster onboarding cycles or improved customer responsiveness.
These approaches can create stronger alignment between pricing and value delivered. But they also introduce new complexity. Outcomes must be clearly defined, measurable and attributable. As workflows become more dynamic and AI-driven, productivity gains become harder to isolate, and establishing credible baselines and governance mechanisms becomes increasingly important.
How can consumption-based pricing balance flexibility and volatility?
Consumption-based pricing models are becoming more common, particularly in AI-enabled and cloud-based environments where costs are tied to usage, computing resources or AI activity.
These models offer flexibility and scalability, allowing closer alignment between costs and actual demand. In some cases, they can also reduce barriers to experimentation by enabling organizations to adopt AI capabilities incrementally rather than through large upfront investments.
But greater flexibility can also introduce greater volatility. For finance leaders, highly variable pricing structures can make forecasting and budgeting more challenging, particularly when AI usage patterns and productivity gains remain difficult to predict. As a result, many clients are looking for approaches that balance consumption flexibility with greater cost transparency, governance and operational control. Some organizations are responding by exploring pricing structures tied to measurable units of work. For example, data services may be priced per terabyte of AI-ready data processed, while AI-augmented development work may be priced per story point or function point delivered. For CIOs, this enables treating IT services more like a variable operating expense that scales with business activity, much like cloud consumption models today.
How should organizations price AI-enabled operational capacity?
While still nascent, clients are beginning to explore models that incorporate AI agents and automation in service delivery. In these models, pricing may be tied to automated interactions, AI-driven activities or operational outcomes traditionally associated with human-supported delivery.
The appeal is understandable. AI-assisted delivery models can dramatically accelerate activities such as coding, testing and operational support.
However, the long-term implications of these models are still evolving. As AI capabilities continue to progress, the relationship between delivery effort, cost and value becomes less predictable. Faster outcomes and higher levels of automation may be achieved without a clear or consistent relationship to the underlying pricing structure.
As AI agents increasingly augment traditional delivery approaches, clients may look beyond staffing-based pricing structures toward models that reflect a combination of AI-enabled delivery capacity and human expertise. This could create a simpler transition toward automation than highly granular consumption-based approaches tied directly to API calls or token usage.
What is the AI "token paradox"?
One emerging challenge is what some describe as the “token paradox”—the reality that while the unit cost of AI consumption continues to decline, overall AI spending can still rise significantly as organizations adopt more complex, agentic AI workflows.
As enterprises move from simple prompt-response interactions toward continuously running AI agents and automated workflows, consumption becomes more difficult to predict, govern and budget. This creates a new pricing challenge for both providers and clients. Traditional labor-based pricing models become less relevant as AI accelerates delivery. At the same time, pure consumption-based approaches can expose organizations to rapidly escalating costs unless clear guardrails are in place. As a result, organizations are increasingly looking for pricing structures that balance flexibility, predictability and accountability.
How can pricing models balance predictability and adaptability?
In practice, hybrid approaches that combine elements of multiple pricing models are becoming more common. These structures often blend predictable base pricing with variable components tied to consumption, outcomes, automation levels or business performance metrics.
Hybrid models can create greater flexibility while still preserving a degree of financial predictability. They also allow enterprises to adopt AI capabilities incrementally, creating room for experimentation without fully exposing either party to operational or financial volatility.
But even hybrid approaches involve trade-offs. Simplicity, transparency and governance remain critical. Models that attempt to perfectly align pricing with every dimension of value can quickly become overly complex to manage, benchmark and scale.
What is emerging is not a single dominant pricing structure, but a more adaptive commercial landscape. In our experience, clients making the most progress are not necessarily those using the most innovative pricing mechanisms, but those able to align pricing, governance and operational flexibility as business conditions continue to evolve.
While the market continues to explore a range of pricing approaches, we believe hybrid structures have the greatest potential to become the preferred commercial model for AI-enabled managed services. A typical model may combine a fixed base fee for core services and governance, a variable component tied to AI consumption and outcome-based incentives linked to measurable performance improvements. For CIOs, this approach provides greater budget predictability while creating opportunities to leverage AI-driven efficiency and improve business outcomes. For providers, it helps manage the volatility associated with rapidly changing AI consumption patterns while creating incentives for continuous innovation and performance improvement.
Why are static service contracts failing dynamic businesses?
Across industries, the businesses these service models support are becoming far more dynamic. Pricing changes in response to demand. Volumes fluctuate more rapidly. Priorities change in real time.
In banking, transaction volumes rise and fall with market conditions, yet many service contracts remain fixed around steady-state assumptions. Utilities must respond to seasonal demand swings and infrastructure volatility, while their operating models are often designed for predictability and consistency. In healthcare, patient volumes and claims activity can fluctuate dramatically, but delivery structures may struggle to flex accordingly.
More broadly, many enterprises are adopting increasingly dynamic business models of their own. Retailers may adjust pricing and promotions daily based on demand signals and customer behavior. Public sector agencies may face sudden surges in demand for citizen services during economic disruption, weather events or policy changes.
In our experience, while the front end of the business is becoming more adaptive, the service and commercial models beneath it often remain relatively static. The result is growing friction between how businesses operate and how services are structured, delivered and priced.
How can organizations reinvest the value created by AI?
Managed services have historically been built on stability, not just in delivery, but in economics. Contracts assume a degree of revenue continuity, while efficiency gains are often framed primarily as cost reduction. But in our work with clients, we are hearing business leaders ask a different question: not simply how much cost can be removed, but how the value created by AI can be applied to higher-impact business outcomes?
This represents a broader change in how decision-makers define value. Rather than focusing on efficiency and cost optimization, CIOs are recognizing that AI can create value not only by reducing effort, but by accelerating modernization, compressing timelines and enabling faster operational and business outcomes. Cost savings are no longer viewed as the end game, but increasingly as a source of reinvestment to fund continued transformation and innovation.
Speed has entered the value equation. When AI-assisted delivery compresses a modernization initiative from several months into a matter of weeks, the value created often extends well beyond delivery efficiency alone. Benefits may include faster time-to-market, quicker deployment of digital capabilities and the ability to respond more nimbly to changing demands. In some cases, this is contributing to the emergence of self-funding transformation models, where operational efficiencies help finance future digital initiatives and AI-enabled capabilities.
Which AI-enabled managed services pricing models should leaders watch?
Hybrid value models
Pricing structures increasingly combine fixed base pricing, variable AI consumption and outcome-based incentives. This model provides a practical middle ground for CIOs seeking budget predictability while still creating incentives for efficiency, innovation and measurable business outcomes.
AI-enabled operational capacity
Organizations may move beyond staffing-based pricing structures toward models that combine AI-enabled delivery capacity with human expertise. This model simplifies the transition toward automation while avoiding the complexity of highly granular token- or API-based billing.
Consumption-based unit pricing
Some services may increasingly be priced per unit of work delivered, such as AI-ready data processed or AI-augmented development output. This model allows IT service costs to scale more dynamically with business demand and operational activity.
Gain-sharing and performance-based models
Providers may assume greater delivery risk in exchange for a share of verified business outcomes or operational improvements. This model enables self-funding of modernization and AI-enabled transformation initiatives.
What does an adaptive managed services partnership look like?
As the market transitions beyond static pricing structures, partnership models must become more adaptive, transparent and responsive to changing business conditions and AI capabilities.
This represents a broader change in how decision-makers define value. Rather than focusing on efficiency and cost optimization, CIOs are recognizing that AI can create value not only by reducing effort, but by accelerating modernization, compressing timelines and enabling faster operational and business outcomes. Cost savings are no longer viewed as the end game, but increasingly as a source of reinvestment to fund continued transformation and innovation.
1. Build for adaptability, not certainty
Traditional managed services contracts were often designed around fixed assumptions for scope, volumes and delivery patterns. But in more dynamic operating environments, organizations increasingly need partnership models that allow pricing, capacity and service expectations to evolve over time.
This may include flexible scope bands, periodic recalibration points or pricing structures that adjust as automation levels, demand and business objectives change.
2. Establish baselines before defining gainshare
More adaptive pricing models also require stronger baseline discipline. AI-driven gains, productivity improvements or operational outcomes cannot be measured credibly unless the starting point is clearly defined.
As a result, we advise clients to invest more heavily in discovery, benchmarking and operational assessment. Establishing measurable baselines for effort, cost, service levels and operational performance reinforces trust and alignment around how value is measured over time.
3. Use gainshare selectively
Gain-sharing approaches can create stronger alignment between providers and clients, particularly when AI enables measurable operational improvements or accelerated business outcomes. However, in our experience, these models are most effective in targeted use cases where outcomes can be clearly defined, measured and attributed.
Attempting to apply gain-sharing broadly across highly complex, evolving environments can introduce additional governance and measurement challenges.
We anticipate that gain-sharing models will become increasingly important in high-stakes transformation environments where CIOs are looking to self-fund AI-enabled modernization initiatives. These approaches can create stronger alignment between investment, performance and measurable business outcomes over time.
4. Increase transparency and auditability
As pricing models become more dynamic, transparency is a strategic imperative. Business leaders want greater visibility into how value is measured, how automation impacts cost structures and how AI-driven outcomes are being governed in the long run.
This is driving greater focus on shared metrics, operational visibility, auditability and governance mechanisms that reinforce trust between clients and providers. As AI becomes more embedded in decision-making, clients are looking for greater transparency, stronger human-in-the-loop guardrails and clearer oversight of how AI-driven outcomes are monitored and governed.
5. Enable flexibility in capacity and consumption
Clients are also looking for more flexible approaches to capacity and service consumption. Rather than structuring services around fixed delivery assumptions, some are adopting flexible capacity pools, variable consumption models or dynamically allocated operational support structures.
These approaches can help organizations respond more effectively to fluctuating demand while preserving greater operational and financial flexibility.
6. Align the model to how the business operates
Above all, we advise clients to ensure that service and pricing models reflect the realities of the business environment they support. As enterprises become more dynamic, commercial structures designed primarily for stability can create operational friction rather than flexibility.
The most effective partnership models align closely to strategic priorities, operational responsiveness and business agility rather than fixed delivery assumptions.
Ultimately, the goal is no longer to optimize cost within a static operating model. It is to create partnership structures that can evolve alongside the business itself.
How can managed services partnerships be designed for uncertainty?
AI is forcing the managed services industry to confront a reality it has long avoided: not everything can be predicted upfront.
For decades, managed services economics were built around relatively stable assumptions about demand, labor effort and delivery expectations. But AI is changing how work is performed, how quickly value is realized and how business needs evolve. As a result, the traditional relationship between effort, cost and value is becoming harder to define.
In response, business leaders are becoming more open to adaptive pricing approaches, including outcome-based, consumption-based, agent-based and hybrid models. While no single structure has emerged as the clear winner, each offers advantages in some situations while introducing trade-offs in others.
The future of managed services will not be defined by rigid pricing structures or fixed delivery assumptions. It will be shaped by partnerships built on greater flexibility, shared accountability and a mutual commitment to delivering business outcomes in a less predictable world.
The goal is no longer to perfectly price the future. It is to design for it.
