Benefits could include lower labor costs, increased process throughout, improved process quality, greater delivery model flexibility, better scalability, and better ROI due to the relatively low cost to implement.
With more companies deploying “bots” through Robotic Process Automation (RPA) in organizations every day, automation is clearly the future, and in many cases, it is already the present. However, bots are not magic fix-it machines. They are part of the solution to strategic advantage through automation. As was shown at the June Financial Executives International Committee on Finance & IT (CFIT) meeting, co-hosted by Nike, Inc. and Intel Corporation in Portland, Oregon, RPA can add significantly to finance’s value add, but automating an already broken business process is not a recipe for success.
In a collaborative presentation, Nike’s Director of Global Accounting Operations Eric Rant, Ranjit Rao, Director with Deloitte Consulting, LLP and Patrick Edwards, HR Automation Strategist with Intel discussed the potential opportunities RPA offers organizations. They offered several insights on what they had learned but also cautioned CFIT members on the challenges and potential stumbling blocks bots can present.
For some the idea of bots conjures images of I, Robot or Rosie from the Jetsons processing your accounting and finance work. While these images can be funny or scary, depending on your point of view, that’s not RPA. Simply put, RPA is advanced macro-like capabilities that can be deployed at an enterprise or business unit level.
What RPA is
- Computer coded software
- Cheaper and faster way to automate processes
- Cross-functional, cross-application macros
What RPA is not
- Physical hardware
- A multi-year technology deployment
- Artificial intelligence
According to a July 2016 Deloitte survey of 143 global shared services leaders, 76 percent plan to pursue RPA within the next year; those who have piloted or implemented at scale expect significant benefit. Those benefits could include lower labor costs, increased process throughout, improved process quality, greater delivery model flexibility, better scalability, and better payback / ROI due to the relatively low cost to implement.
As part of their pilot, Nike simulated a few aspects of their process being automated in a safe environment. The pilots are small-scale through a short-term automation platform to test and prove changes. Current organizational processes that were previously performed by humans are not being performed by bots. Following the pilot, business cases can be created and next steps recommended. Additional areas can then be identified for RPA expansion opportunities within Controlling.
RPA can present a short payback period which provides the opportunity for recognizing significant benefits quickly. However, one-time costs such as license cost (platform, bots, developer); training cost, and design, development and implementation costs should be weighed against full-time equivalent (FTE) benefits such as FTE capacity. Additionally, on-going costs such as bot management cost, maintenance and operations cost, and managed services cost should be consider against bot efficiency – processes per bot and improvement in efficiency.
Key success factors for successful RPA programs include:
- Right process or activity
- Stable not broken processes
- Quality and exceptions management
- Adequate education and user adoption
- Checklist for requirements
- Stakeholder management
- Approach to measure and track benefits
The RPA business challenge for Intel was that their enterprise capabilities did not match every requirement. Nearly half of their Human Resources Employment Solutions (HRES) business requirements are satisfied by thousands of disconnected, manual or semi-automated solutions. Through the use of RPA, the organization hoped to deploy a unified, business-friendly platform to automate and consolidate disconnected processes and point solutions – a more end-to-end experience with higher quality and lower costs. Ideally, in the future HR would offer no point solutions outside enterprise capabilities or unified platforms. Human Resources would then be enabled to continually drive quality, efficiency and velocity with exceptional, end-to-end customer experiences.
With their first RPA pilot HRES’ expected:
- Automation results are fast and low-cost.
- Easy to accomplish.
- IT engagement optional.
- Fast-pass due diligence would adequately identify material risk.
- Find and deploy use cases with minimal impact to operations.
- Technology will meet our requirements and process paradigms.
However, what they found was the following:
- Initial results are not fast; expect a steep learning curve and more time to market at first.
- Played down the customer reference calls where they stated that the first use cases would be difficult.
- Underestimated the technical dependency.
- Our pilot platform was unable to process disparate tasks in a sequence.
The valuable learnings from the pilot were:
- Assume risk when you undertake risk taking.
- Provide equal weight to customer references and vendor/consultant claims.
- Recommend time and materials contracting governed by a robust SOW and Change Management.
- Allow sufficient time to validate vendor claims with proofs of concepts.
- Use every ounce of ‘business development’ offered by vendors.
- Budget for vendor due diligence process.
- Be selective with first use case – ensure adequate business engagement.
- Don’t assume any technology solution is completely business friendly.
- Don’t establish artificial deadlines.
- Cast a wide net when evaluating technology platforms.
- Engage with IT & IT procurement when exploring any technology.
- Know your internal process paradigms.
As Patrick Edwards astutely pointed out, “If you’re not making mistakes; you’re not innovating”. And in today’s business world, if you’re not innovating; what are you doing?
It should be noted that Intel HR forged ahead with RPA after reflecting on and learning from our pilot experience. They now have RPA use cases in production producing business value and a backlog of promises use cases.•