In the current evolving technology landscape, generative AI (gen AI) has been leveraged as a transformative force, promising opportunities for innovation and efficiency, in most of the contemporary IT and BPO services deals.
Proposing gen AI in the solution is no longer considered as a differentiator for service providers but has become a table stake in the past two to three quarters. In line with the evolving nature of gen AI, contractual language must also adapt swiftly to address the unique challenges and opportunities presented by these solutions in information technology (IT) and business process outsourcing (BPO).
Here’s our take on which critical contractual considerations must be prioritized during the contracting stage, to improve the effectiveness of such solutions in services engagements.
Reach out to us to discuss this topic further with our expert analysts.
Critical contractual aspects for gen AI sourcing
Ensuring robust contractual clauses in service agreements is paramount for the success of any partnership involving gen AI solutions. While numerous factors come into play, the following critical elements should be considered in the framing of any contract to establish a solid foundation for collaboration:
- Input data treatment: The legal landscape surrounding gen AI is as complex as the technology itself. Enterprises must navigate a maze of industry-specific regulations and regional laws that directly impact data procurement and model training. Enterprises need to actively monitor the origins of input data to ensure it has been legally procured. Contracts must explicitly define the sources of data—whether they are server logs, IT service management (ITSM) ticket dumps, or other proprietary datasets. The nature of this data must be clearly understood whether it is proprietary, copyrighted, or sensitive. The risks associated with potential IP infringement cannot be overstated, and enterprises must protect themselves by incorporating robust clauses. Aligning and outlining data handling protocols with the vendor is essential to safeguarding data privacy and security. Contracts should define a clear RACI matrix for how data will be managed
- Large language models (LLMs) and 3rd party original equipment manufacturers (OEMs) leveraged: The parties should have clarity on the LLM which will be used for the solution. Will it be an open-source model, a pre-trained model, or a custom-built one? Additionally, is there clarity on who will procure the LLM and any associated tools—the vendor or the client?
- Acceptance criteria: Defining the Definition of Done (DoD) and acceptance criteria to agree upon the tasks that must be completed, to consider the unit of project delivery to be successful is of utmost importance, especially when the vendor is managing the project delivery. Tracking and monitoring progress towards these goals can be effectively managed by having robust service level agreements (SLAs) and key performance indicators (KPIs) in the contract
- Commercial construct: a critical consideration for parties engaging in gen AI services is the commercial construct. Options include time and materials (T&M), fixed fee, or per-use case. Exploring outcome-based models, even in a hybrid approach, can also be a valuable avenue to align incentives and measure success based on the desired results
Selecting the desirable considerations will make a significant difference in procurement
When procuring gen AI solutions, the contracts governing these deals are not merely administrative formalities—they are strategic instruments that dictate the success or failure of the entire initiative. To truly capture this potential, enterprises and service providers must iteratively refine contractual clauses, ensuring they align with the unique demands and risks associated with implementation.
In conclusion, sourcing gen AI solutions is a high-stakes endeavor, with contracts serving as the linchpin for success. Enterprises and service providers must adopt a strategic approach to these agreements, emphasizing and negotiating critical clauses to ensure optimal outcomes. By meticulously crafting contracts that address critical aspects, both players can navigate the complexities of gen AI sourcing and position themselves to maximize the benefits.
Everest Group partners with both enterprises and leading IT service providers on mandates related to pricing strategies, productivity gains, and pricing impact related to gen AI.
If you found this blog interesting, you can read our Meeting The ROI Bar: Client’s Expectations For Generative AI In IT Outsourcing Deals | Blog – Everest Group (everestgrp.com) blog, which delves deeper into the topic of gen AI.
If you found this blog interesting and you want to discuss this topic in more detail or for a detailed analysis, contact us at [email protected], or please contact Prateek Gupta and Swapnil Pamecha.
