Practical Commercial Tender and Contracting Approaches to Support AI-Enabled Outsourcing
Artificial intelligence (AI) is breaking the two commercial and governance assumptions on which outsourcing contracts have relied for decades: 1) people perform specified work and customers pay for the effort, and 2) services evolve slowly enough for contracts to keep pace.
AI is transforming the nature of managed service delivery by increasingly replacing human effort with autonomous agents and introducing inherent uncertainty through large language models (LLMs). At the same time, AI capabilities are advancing faster than contracts can be negotiated or updated. The result is a widening gap between traditional outsourcing contracts and AI-enabled service delivery – a gap that requires a fundamental rethink of how managed IT services are tendered and contracted in the age of AI.
The New Uncertainty
AI is rapidly moving from experimentation to expectation in managed IT services. This shift aligns with trends reported in the quarterly ISG Index™ call and other ISG research, which show enterprises increasing rely on managed service providers as strategic partners. They are accelerating AI adoption, addressing skills shortages and industrializing GenAI to capture the structural benefits from agentic operating models. At first sight, this may look like a familiar IT technology shift and efficiency story. It is not.
First, AI redistributes decision rights. Customers and providers must decide where AI can recommend, execute or learn autonomously, and where humans remain accountable. AI works best with clear processes, structured data and stable ownership; in blurry customer environments, it may accelerate dysfunction rather than improve outcomes.
Second, traditional IT managed services are deterministic, producing the same output for a given input; the nature of AI, and especially generative AI and deep learning, is probabilistic. LLMs can be tuned, constrained, grounded and monitored but not made deterministic by contractual provisions. Hallucinations arise because standard training and evaluation reward guessing over acknowledging uncertainty. In outsourced environments, hallucinations may cause performance, reliance and accountability issues: who relied on the output, what controls applied, was the error foreseeable and did the customer approve the use case?
These AI-driven changes to the enterprise risk profile expose the limits of traditional outsourcing contracts, requiring RFPs and contracts to evolve.
Please fill out this form to continue.
When Rules Run Out
Traditional outsourcing contracts assume that future service delivery can be specified upfront by customers in sufficient detail. AI challenges this assumption. New models, new agents, new regulations and new AI use cases emerge continuously making the future shape of services inherently uncertain. As a result, attempting to anticipate every scenario through detailed contractual rules and requirements is no longer practical.
Customer-protective RFP requirements and contract clauses remain necessary because outsourcing creates dependencies that are hard and costly to exit. Security, privacy, intellectual property, audit, regulatory support and liability still require contractual precision. But conventional rule-based clauses, while important, cannot carry the full weight of dynamic AI-enabled services. Taken too far, they encourage customers to hide behind contract positions rather than collaborate and providers to deliver only minimum compliance. Neither behavior creates value in managed IT services relationships.
Principle-based provisions are better suited to uncertainty because they set the standard for the parties’ conduct and response when the facts change. Principle-based provisions complement conventional rule-based provisions to more effectively deal with the elevated levels of uncertainty that AI introduces and to address specific topics such as responsible AI use.
Since AI requires continuous experimentation, learning and rapid adaptation, service outcomes and operating conditions will evolve throughout the contract term. Continuous service redesign will drive change at a pace that traditional contract change procedures will not be able to handle. Starting with an honest and transparent conversation about governance principles for managing changes in technology, volume, regulation and risk is therefore likely more valuable than prescribing meetings and reporting obligations. The contracting challenge then shifts from answering "What happens in scenario X?" to answering "How will we make decisions when scenario X inevitably changes?"
Trust Requires Transparency
Trust in AI-enabled outsourcing is earned through transparency and control. Customers and regulators require visibility into AI use that is proportionate to risk and sufficient to assess its operational, regulatory, risk and pricing implications, without requiring disclosure of proprietary models or trade secrets.
Contractual transparency can be established in multiple ways. Customers should require an AI use-case register, an AI value roadmap (for planned improvements), risk classifications, data-use transparency, testing evidence and clear change control triggers for material increases in autonomy. RFPs and contracts should also define (ideally by service tower, process and use case) where AI may assist, recommend or execute decisions, who retains accountability, what evidence must be retained, when customers must be notified, and where human oversight is required. The EU AI Act provides a useful reference point for this. Human-in-the-loop must be an operating requirement in RFPs and contracts wherever trust and accountability require it.
The Price of Autonomy
Pricing is where AI commitments become real. If AI reduces human effort, changes skills or shifts work from people to AI agents, the commercial model should reflect it: pricing should progressively shift toward business outcomes and the actual level of execution autonomy achieved. Otherwise, customers may overpay for increasingly automated services, while providers may lack incentives to invest in AI.
ISG's Autonomy Level Pricing™ (ALP) framework addresses this challenge by linking price to execution maturity, human oversight and risk. It replaces static units such as "per ticket" with execution-aware dynamic pricing that reflects increasing levels of autonomy, providing a transparent mechanism for commercializing and tracking AI-driven transformation.
Fighting AI Lock-In
Providers are promoting proprietary AI platforms and agent ecosystems, but customers continue to operate in multi-vendor environments. While service integration is familiar, AI raises the bar: customers must orchestrate an ecosystem of AI-enabled providers, platforms and enterprise capabilities that operate consistently under common governance and interoperability standards, rather than simply manage individual providers. As AI solutions remain fragmented and rapidly evolving, interoperability is becoming more important – not less.
Customer-first providers should prioritize customer AI requirements over pushing proprietary AI technology. Contract for AI interoperability, not lock-in, backed by robust exit rights, data portability, open integration and effective multi-vendor cooperation.
Contract for the Long Run
AI is becoming a standard feature of managed IT services. Contracts for AI-enabled managed IT services should therefore be designed for years of service (r)evolution, not just for contract signature. This requires greater reliance on principle-based requirements that are better suited to uncertainty than rule-based provisions. Customers should require transparency around AI use, define where human accountability must prevail, align pricing with increasing autonomy and protect long-term flexibility through interoperability and robust exit rights. Organizations that embed these principles into outsourcing RFPs, governance models and contracts will be best positioned to capture AI-driven value while maintaining control of risk and outcomes.
Ultimately, AI-enabled outsourcing should be designed not merely to manage emerging risks but to continuously improve business outcomes. Contracts should therefore align commercial incentives with business outcomes while enabling continuous innovation, measurable value realization and adaptive governance throughout the sourcing lifecycle. After all, the greatest risk for customers tendering and contracting for outsourcing in the age of AI is not losing control to AI – it is locking tomorrow's operating model into yesterday's contract.
ISG helps enterprises navigate a rapidly changing AI market and negotiate sourcing contracts that set both parties up for success. Contact us to find out how we can help you.
About the author
