Find out how RPA and AI will potentially impact the workforce and how to make the most of these tools
By Brad Peterson
By offering the prospect of better service quality at a faster rate and lower cost, robotic process automation (RPA) and artificial intelligence (AI) promise to radically transform human resources and services. So what are RPA and AI?
- AI involves sophisticated pattern-matching software designed for specialized tasks. The pattern-matching software “learns” from sample data and can be “trained” by providing it with the right data. AI has a long history but is becoming practical for HR because of improvements in processing power and algorithms.
- RPA does not actually involve robots, but is a software that mimics a human user. So to a software application or cloud-based system, an RPA “robot” appears to be a human user. RPA can partially or fully automate processes that are manual, repetitive, and rule-based. Lucky for HR practitioners, a substantial part of recruiting and hiring processes fit that description.
For example, a company might outsource a new employee- intake process that involves updating the company’s payroll, employee directory, facility management, and IT and security systems. These processes may require hundreds of clicks, copies, and pastes in order to bring new employees on board. With a human user, it also might take hours and only be 98 to 99 percent accurate. But with an RPA robot, the same process could take minutes and be 100 percent accurate. The programming of an RPA software can also be done in weeks and without altering other code.
At the same time, HR can leverage AI tools to capture information from these processes and convert it into useful insights. For example, voice recognition— a common AI application—can allow employees to perform self-service tasks by voice instead of typing on a keyboard. AI search functions can help employees easily find information in an unstructured database.
Going forward, AI will increasingly be used to screen for indicators of performance ability or gaps, for compensation out of line with standards, or for problems. After identifying problems, AI can then recommend or effect change. AI tools do this by looking at the full data set, then applying what it has learned to each employee.
RPA and AI: Differences
RPA and AI are fundamentally different technologies.
- RPA is programmed and requires a structured process.
- AI can train itself or be trained.
- RPA can copy and paste from field to field in other softwares.
- AI can process natural language and unstructured data.
The return on investment also differs greatly. One specialist from IBM estimated that AI can take eight to 10 times as long to implement, but can have 10 to 25 times the value compared to RPA investments. Also, while RPA will increase efficiency, AI will increase effectiveness. This distinction will likely blur as more products arrive with both RPA and AI capabilities.
RPA and AI: Impact In the coming years, RPA and AI will likely replace low- skill, repetitive HR tasks with solutions that can scale easily. HR service providers are likely to add higher-skill positions for people who configure, train, and work with the RPA and AI software to provide more accurate and insightful service at lower costs. Organizations are likely to restructure traditional job roles into tasks that can be handled by both workers and robotic tools. Many HR service providers are already using RPA and AI to dramatically lower their costs without passing their savings onto customers. This is because many service contracts— drafted years ago—have no barriers to the provider’s use of RPA and AI. So organizations need to be proactive in demanding to share in the benefits of these RPA and AI innovations that are already taking place.
A good first step is to include RPA and AI capabilities as criteria in evaluations and selection processes for HR service providers. Organizations should also modify their sourcing contracts so that they t services provided substantially in an automated manner. It is no longer sufficient—and no longer sensible—to contract for good and workmanlike conduct by adequate numbers of suitably trained and qualified employees. Instead, consider terms such as the following:
- Visibility and perhaps approval rights on the use of RPA and AI solutions;
- Obligations for RPA or AI to meet contractual specifications;
- Access, license, and support clauses similar to a SaaS contract;
- Change control protections;
- Obligations to modify RPA and AI solutions as your business changes;
- Service-level measures designed for automated service delivery; and
- Data privacy and security. Watch out for data leakage. Data has little protection under current intellectual property laws, so strong contract rights are critical for preserving competitive advantage. Contracts with outsourcers may leak data rights in subtle ways, such as through provisions stating that the provider “may use organizational data in order to improve services and for other business purposes” or that the “supplier may aggregate and utilize customer data in combination with other similar data so long as the customer data is anonymized.” The increasing value of data coupled with the growing power of RPA and AI solutions makes these clauses increasingly common and risky.
HR should also rethink ownership and use rights for RPA and AI solutions. Consider what ownership or use rights are needed in the tool to avoid lock-in or unanticipated costs. For AI, consider whether the organization can separately obtain a copy of what the AI software learned while providing the services and, if so, whether the company needs that information during steady state or after disengagement. A key intellectual property point is to specifically state that work produced by RPA or AI tools will be treated as if it were produced by a supplier’s personnel to address the risk that the provider will claim that IP laws do not apply to developments created by software.
Another consideration is restructuring pricing for automated services. For example, in upcoming contracts, organizations can build in cost-reduction commitments from the provider to take advantage of cost reductions available with RPA and AI capabilities. It is a good idea to fund transformation projects to implement RPA and AI as well. If the HRO contract price is based on full-time employees or other effort-based measures, consider repricing with task-based or outcome-based measures that encourage use of RPA and AI tools. Organizations should also reconsider cost-of-living adjustments because the cost of running robots is dropping quickly.
Analyze whether RPA and AI affect compliance with other software licenses. If the company has a license for software that is priced based on the number of users, how will the substitution of a software robot in place of humans be counted? As zero users? As a single human user? As middleware, turning the former stakeholders into users? Alternately, do an organization’s other software licenses impose limits on interfacing RPA software with other licensed software?
Consider hybrid solutions. Because RPA and AI software are geographically agnostic, customers may retain responsibility for RPA and AI software and outsource the human labor portion. This solution would require balancing the costs of licensing RPA and AI software and acquiring staff to configure and train software instead of utilizing a service provider’s leveraged capabilities. However, it can also be helpful in avoiding regulatory restrictions (such as privacy concerns) and creating a smooth path to a successor provider.
RPA and AI are likely to have a major transformative effect on how companies operate and what services they buy. But the industry is still in a relatively early phase of this trend. The traditional HR service delivery model is bound to evolve, and many service providers appear to be responding by building or incorporating RPA and AI capabilities. Service providers are seeing dramatic reductions in their costs, but existing customer contracts may not enable organizations to share in those reductions. There is substantial value in securing commitments, options, and incentives for HR service providers to leverage RPA and AI with appropriate customer protections in the coming years.
Brad Peterson leads the Technology practice at Mayer Brown.
SIDEBAR: Gearing Up
HR service providers are responding to the potential growth and impact of RPA and AI solutions by building up their capabilities. For example:
Cognizant acquired Trizetto;
Genpact acquired Rage Frameworks;
Wipro has created an AI platform called Holmes;
Accenture created myWizard;
Infosys created Nia; and
TCS created Ignio.
In addition, HR service providers are partnering with specialist RPA and AI vendors such as AutomationAnywhere, Blue Prism, UIPath, Ipsoft, Automic, and Celaton.