Many existing IT and business process outsourcing deals did not anticipate the introduction of new technologies like artificial intelligence (AI) and robotic process automation (RPA). Nor do these legacy contracts prevent providers from unilaterally introducing such capabilities. And while RPA and AI can bring enormous efficiency and cost savings benefits to the outsourcing engagement, a number of issues can arise if the original contracts don’t take such possibilities into account.
[ Related: Why automation doubles IT outsourcing cost savings ]
For example, the pricing for the deal would likely have been based on an assessment of the number of full-time employees necessary to deploy or provide the service. The implementation of AI or RPA would eventually reduce the number of employees required to perform a task. But without a contractual mechanism to adjust for the introduction of this new automation or technology, the customer is unlikely to see any of that cost savings.
A move to an AI- or RPA- centric solution could make it more difficult for the customer to switch providers down the line. The service levels put in place at the time of contract signing may no longer be appropriate in a more highly automated environment. On the flip side, customers whose deals do not address these capabilities may fall behind if competitors begin to deploy them and their IT service providers do not adopt these newer approaches. And those are just a few of the potential impacts.
That’s why it’s important for IT outsourcing customer to proactively address the use of RPA and AI in their existing engagements today, say Peter Dickinson and Paul Roy, partners in the technology transactions practices at law firm Mayer Brown.
“The speed of technological change is accelerating. This combined with the fact that AI and RPA solutions are becoming commercialized and more widely deployed, means that the consequences of not addressing new technologies now will become increasingly significant,” Dickinson says.
[ Related: Outsourcing trends to watch in 2017 ]
IT leaders should take action on a number of fronts in order to reap the most benefits from and mitigate the risks associated with RPA and AI capabilities in their existing and future outsourcing engagements:
1. Include requests for RPA and AI capabilities in RFPs—and include their use as criteria in the selection of providers.
“The two most important reasons for including requests for RPA and AI capabilities is cost reduction and visibility to long-term impact on the customer’s operations,” Roy explains. “Without a focus on these capabilities, providers who rely on labor-based pricing may have little incentive to replace labor with machines and other providers may want to retain the cost savings for their bottom line.”
2. Consider the changes that will be required to incorporate RPA and AI capabilities and require a transformation plan — with corresponding commitments and incentives written into the contract.
Implementing RPA and AI often requires reconfiguration of workflows and associated interdependencies which require additional time and effort to map, configure, and test the new systems and processes. “In an outsourcing context, the provider will take on most of these responsibilities, and the efficiency of completing the implementations will have an effect on costs and disruptions to the customer’s business as well as the time period to realizing cost savings,” says Roy. “Applying the appropriate incentives to the provider will be important to ensuring provider makes the required investments in a timely manner.”
[ Related: Should CIOs be chief robot wranglers? ]
3. Consider new contractual commitments appropriate for RPA and AI.
These may include testing rights or the right to review configurations or coding. “Commitments such as testing are essential to ensuring the desired outcome and avoiding disruptions to the customer’s operations,” Roy says. “Visibility to the configurations and coding is important for the customer’s control of their environment in the course of the outsourcing arrangement—and to manage its exit form that arrangement.”
4. Determining what service levels will be used for RPA- and AI-enabled work or functions.
“Service levels are designed for measuring the quality of a service solutions,” explains Roy. “New service levels may be required when the service solution changes.” Measures appropriate for measuring the effectiveness of manual functions will no longer be suitable. Others may not account for changes in process flows. “This could drive increasingly to transaction- or outcome-based pricing given the lower cost and higher level of predictability and control that automation can bring,” Roy predicts.
5. Revisit pricing to factor in productivity commitments for RPA or AI.
The introduction of RPA or AI capabilitie will take time and cost savings will not be immediate. “The best approach for enabling the customer to realize the promises of cost savings through automation is to build a minimum level of price reductions in the contract based on the provider’s plan for implementing automation,” Roy says “This way, the customer will receive the savings even it the provider is delayed.” That approach also motivates the provider to meet its implementation commitments. “Any new introduction of automation during the course of the agreement should trigger a discussion of additional pricing reductions,” Roy adds.
6. Demand visibility into the use of RPA or AI solutions by providers.
The customer should be involved in—or at the very least be aware of—the introduction of new capabilities. “That will enable the customer to share in the cost savings and to avoid unexpected cost or disruption to its operations on exit from its then current outsourcing arrangement,” explains Roy.
7. Analyze whether the use of RPA and AI software affect compliance with software licenses.
“Third party software, such ERP software, may have restrictions or pricing based a specific number of resources, such as the number of users, that are replaced by RPA or AI software,” Roy says. Licenses may also restrict software’s use with other systems not approved by the licensor.
8. Think through ownership and use rights for RPA and IA solutions and whether such solutions could be managed in-house once deal is done.
Many providers have their own RPA or AI software or configurations that are not generally commercially available. “The question then becomes how the customer manages the risk of unexpected cost or disruption to its operations on exit,” says Roy “In some cases, it may be sufficient for the customer to have RPA configurations specific to the customer’s operations or the databases, algorithms and insights generated by AI solutions [in order] to allow the customer or its replacement provider to recreate similar solutions.”
9. Considering who owns what, particularly what AI systems learn as they get smarter.
“Since AI software in its most sophisticated form does not generate any discernible code, the customer’s ability to replicate it is questionable,” says Roy. But customers must be clear in understanding the extent of their ongoing reliance on AI-enabled software. Another important issue is who owns the intellectual property generated by AI-enabled systems. “This should be explicitly addressed by the customer in its agreement with its provider since there may not be any clear answer under applicable law if the parties are silent on the issue,” Roy says.
10. Include provider commitments to adapt RPA and AI systems to the customer environment.
“RPA solutions, in particular, are designed to interact with existing systems, so the configurations of RPA solutions may have to be adapted to changes in the systems with which they interact,” Roy explains. “To the extent these systems include customer’s systems, the provider should be required to make corresponding adaptations in the RPA configurations when needed.”
11. Evaluate hybrid customer/service provider RPA or AI solutions.
On one hand, the customer may be best able to control their own destiny by licensing AI or RPA software and requiring that their service providers use those systems. On the other hand, providers may be most successful by leveraging their own RPA and AI solutions. “Determining where the balance is between those alternatives will be important for customers who want to achieve cost savings while mitigating risks,” advises Roy.
When addressing the impact of RPA and AI on existing outsourcing deals, customers should take care to address the issues that will arise with these newer technologies, but not be so specific that these new provisions are too rigid to adapt as technology evolves.
“Customers need to recognize that it is not possible to predict, with accuracy, how the technology landscape will develop over the next few years. The only certainty is that such a landscape will look materially different from how it looks now,” Dickinson says. “Customers need to adopt ‘future-proof’ provisions, which recognize the introduction of new technologies, but generically, as opposed to by reference to specific technologies and solutions.”