Agentic AI raises the stakes 9 min read
The premise of outsourcing is simple: organisations reduce their costs by leveraging a third party to perform labour-intensive work in a more efficient manner and with a lower cost base.
However, the emergence of AI tools, especially AI agents, is revolutionising outsourcing arrangements. By developing tailored AI solutions, vendors are turbo-charging the traditional outsourcing model—finding new ways to drive down operational costs and maximise value creation.
Whether your organisation is planning to undertake an outsourcing or has already outsourced services, there are a number of legal and commercial issues to be mindful of when deploying AI agents as part of your outsourcing.
Key takeaways
- The role of the outsourced service provider has changed: both service providers and customers increasingly approach outsourcings as strategic partnerships.
- AI agents expand the possibilities of outsourcing: instead of simply completing labour-intensive tasks at a lower cost, AI agents can act as drivers of revenue and provide highly elastic, responsive resources to meet organisational needs.
- Advancements present new legal risks: pricing models, service levels, data and information security, AI safety and continuous improvement are all implicated by the rollout of AI agents and need to be carefully considered in light of the particular technology proposal.
- Outsourcing agreements need to be fit-for-purpose: legacy outsourcing arrangements are unlikely to appropriately deal with AI-powered outsourcing, and the risks and opportunities it introduces, and should be revisited and uplifted. Equally, when embarking on a new outsourcing, a clear understanding of the vendor's AI tools, expertise and risk mitigants is necessary, as well as alignment on how commercial upside will be shared, so that the outsourcing agreement can be tailored accordingly.
Outsourced service providers as strategic partners
Traditionally, outsourced service providers have served as trusted vendors, addressing organisations’ labour requirements, often in lower-cost jurisdictions.
While this model persists, the role of these providers is evolving as they increasingly integrate AI tools—particularly AI agents—into their service offerings. This shift positions vendors not just as service providers but as strategic partners, enabling organisations to leverage the significant investments providers have made in developing or licensing cutting-edge AI technologies. Put simply, the vendor is facilitating the customer's entry into the world of AI. They are not just offering cost savings, but also innovation.
In practice, these strategic partnerships go far beyond simply deploying general-use tools like chatbots. Instead, the most competitive providers are reimagining entire workflows, strategically embedding agents at every stage of the business process to reduce friction, optimise resources, generate valuable insights and maximise overall impact.
The hybrid approach
'Hybrid models' are emerging as the optimal approach for the rollout of AI in an outsourcing context. By combining AI tools with human expertise, service providers can offer the best of both worlds: the efficiencies of AI agents with the monitoring, governance, expertise and empathy afforded by human agents.
From a commercial perspective, the need to scale up or down at speed is a strong selling point of the hybrid model. For instance, a retail business that faces a surge in demand during the holiday rush might traditionally have needed to hire more customer support personnel during this period, thereby increasing full-time employee (FTE) costs. However, with the adoption of scalable AI tools, these fluctuations in FTE costs can be significantly reduced, as AI agents can handle tasks like customer support and stock forecasting more efficiently.
Similarly, from a customer experience perspective, the speed at which AI tools can resolve customer queries, requests or claims can improve customer satisfaction, particularly when matched with escalation to human teams for tasks requiring relationship building and greater human empathy.
Evolving pricing models
As the role of AI agents increases in outsourcings, it is having dramatic implications on pricing models.
Most outsourcings are priced on some variant of a time and materials / number of FTEs basis (ie taking people costs and applying a margin). This creates a dilemma when AI agents are leveraged in an outsourcing context, and in response we are seeing a growing desire to use gain share models.
Pricing schedules should be tailored to ensure gain share models are carefully drafted, and that there are no holes or abilities for one party to 'game' the model.
Workforce transformation implications
With any outsourcing, consideration should be given to workforce requirements, including whether there will be any redundancies and if any employees will transfer to the service provider. However, with an AI rollout, the impact on workforce may continue during life of the agreement—both on the customer side and the vendor side—as further roles are able to be transitioned to AI agents.
This reality is resulting in more involved negotiations around redundancy costs and the commercial model for an outsourcing. For instance, some vendors are seeking to recoup their own redundancy costs from customers when vendor personnel dedicated to the customer are replaced by AI agents during the term of the agreement.
Early engagement on employment matters and legal implications can help mitigate risks associated with workforce changes.
10 legal issues to consider
- Responsibility for oversight, inaccuracies and biases stemming from AI services should be clear. AI safety controls need to be included, such as human monitoring, audit rights and incident management.
- Organisations should ensure they have the means to monitor performance and obtain visibility over how responsibilities are allocated between AI agents and human agents.
- Ensure there is a clear and transparent audit trail regarding what decisions are made by AI agents and the criteria for decision-making. This will become even more critical as agentic AI technology develops and facilitates greater autonomous activity by AI agents across multiple systems.
- Continuous improvement clauses should be tailored to ensure vendors' technology offerings remain competitive, given the pace at which AI technology is evolving. Consider how pricing mechanics can enforce continuous improvement obligations (eg by building in cost-savings commitments or business savings incentives).
- Service level regimes should define responsibility for errors or failures caused by AI agents. Service level regimes should reflect expectations of improved performance by AI agents over time as systems are fine-tuned.
- Document what happens to AI agents at end of the contract term, particularly where AI agents are heavily embedded in business processes or have been trained on business data. Ownership and portability of agents will help ensure your organisation can extract its business from the outsourcing relationship if desired.
- Review any third-party terms governing the AI technology. Outsourced service providers will often try to pass-through third-party terms, which may undercut the negotiated positions under your contract.
- Data flows should be mapped and understood, with a clear process for ensuring applicable privacy laws are complied with when personal information will be handled by AI agents.
- Compliance with AI laws should be addressed separately from general compliance with laws (similar to the approach now taken to compliance with privacy laws), given the pace at which AI regulatory change is occurring.
- Cybersecurity risks should be addressed by implementing contractual, technical and operational safeguards to protect against threats to the AI agents' confidentiality, integrity and availability.
Practical considerations
- Revisit existing outsourcing arrangements to consider:
- whether service providers are already leveraging AI without an explicit contractual framework to govern its use and, if so, whether the contract adequately deals with associated risks; or
- whether existing arrangements are not leveraging AI agents but should be, to improve productivity and cost savings and ensure your organisation remains competitive.
- Design commercial outsourcing models with AI-use in mind. Ensure sufficient detail is provided on how AI will be used, whether these tools are proprietary to the vendor, and what the implications of AI use will be on pricing, service levels and productivity.
- Conduct due diligence around prospective outsourcing vendors' technical expertise, and AI safety and security standards around deploying AI agents.
- Ensure business resilience is prioritised, eg by designing fall-back options where AI agents fail or are unavailable and considering post-termination extraction of data and portability of agents.
Actions you can take now
As the adoption of AI agents accelerates, their expanding use cases present significant opportunities to enhance outsourcing arrangements. However, it is crucial to ensure that your contracts—whether for new or existing outsourcing arrangements—are fit for purpose to safeguard your organisation’s commercial and legal interests.
