Thinking

Optimising financial processes

Posted on:

10 Insights from Shared Services Leaders on RPA and AI Experiences


It was a privilege to be asked to speak recently at the European Shared Services Leaders’ Summit held in London on 15th and 16th March.

This conference, delicately nurtured, tended and moderated by SharedServicesLink was a case study in how to facilitate information and experience exchange between Executives. It was in stark contrast to the glitzy ‘convention’ style events that are often marked by style over substance, confusing arrays of tracks and streams, sales pitches and huge exhibitor areas. funded by the “well-nourished” marketing budgets of large software and services firms. The Shared Services Leaders’ Summit was an intimate collection of 90 senior managers from Shared Services organisations across Europe. Many represented the European presence of the largest global businesses. The focus was on attendee collaboration, down to the structure of small tables of 8 attendees that were gently led into engaging in table discussions on priorities, experiences and challenges as well as debating the nuances of what we had heard.

The event was themed on “The Impact of RPA on Shared Services Target Operating Models” with interesting discussions on the realities of applying RPA in business, with reality often confounding the marketing mythology.

There were great insights from SharedServicesLink themselves as well as GE, Johnson & Johnson, GSK, NBC Universal, Solvay, Philips Lighting, Tarmac, Royal Mail, Caterpillar, Brambles, SAP, Hackett Group, Nyrstar, and Leiden University.

There was some healthy discussion around experiences of RPA implementation, scaling RPA results, the investment required to achieve the benefits, how RPA affects the operating model we have been working with for 20 years, the impact on outsourcing relationships, the opportunity for business partner roles, the impact on talent, culture, process, workforce and change management.

I was intrigued by some statistics shared by SharedServicesLink from surveys that had been conducted with the participants just prior to the event, not least that “35% of RPA implementations have not achieved their planned benefit.” Further figures gave some sense of the state of the union with regard to RPA;

  • Key Benefits of RPA (multiple responses permitted)
    • 81% responded that a key benefit was productivity savings
    • 63% responded that reduction of human error was a key benefit
    • 58% responded in favour of improving the customer experience
  • Shared Services maturity
    • 44% of the represented SSOs had been in existence for over 10 years
    • 62% have operated for over 5 years
  • RPA Journey
    • 48% of respondents were piloting or deploying RPA
    • 27% were active with RPA across multiple processes
    • 44% were active in the P2P cycle
    • 38% in O2C
    • 35% in R2R
  • Productivity Improvements in practice
    • 22% of respondents had experienced improvements in SSO productivity
  • Organizational attitude towards RPA
    • 87% of respondents indicated that executives were positive about RPA
    • Only 17% of respondents felt that transactional workers were positive
  • Relationship to BPO
    • 50% of respondents had captive Shared Services
    • Only 2% were fully outsourced
  • RPA impacting BPO contract renewals
    • 28% respondents agreed that RPA will impact their BPO contract renewal negotiations
  • Talent Management
    • 17% of respondents say that RPA is already impacting talent management / HR strategy
  • AI & Machine Learning
    • 20% of respondents say they are already using AI in business processes!

There was an interesting trend displayed by presenters in naming their new RPA software components or “bots”, with one named “Marvin”, which will bring a wry smile to the face of any readers of “A Hitchhikers Guide to the Galaxy”. The irony is not lost that in Douglas Adams book, it is the android that is paranoid, whereas in the real world it seems the humans are!

Many presenters shared that whilst “it is easy to sell an FTE headcount reduction business case up the line”, in many cases the resultant productivity increases did not directly produce headcount reduction. Indeed, it was argued that effective change management required that the participants in the process should not be in fear of their jobs.

There was some thought provoking discussion about “what BAD looks like” and how to avoid these characteristics. However, it is clear that RPA can cost effectively augment ERP for business needs with a 2-4 year horizon that cannot be justified for ERP extension/customization. However, many experienced leaders indicated that ideally, these new capabilities should be incorporated into the core ERP. Which reminds me of the great observation by BD (formerly Becton Dickinson) that “RPA is about fixing the potholes in the road. Do not confuse it with building the new highway”.

There was consensus that we need to understand our business processes much better to get the real value from RPA. I gave a talk entitled “The Role of AI in Shared Services” sharing some ongoing research which was stimulated by my co-founder’s role as a key designer of Apple’s Siri, the first Virtual Personal Assistant, now on every iPhone. The machine learning experiences and lessons from this have created some great results in advanced business process analytics, which caused quite some discussion. AI can be a powerful technique in driving analytics and smart recommendations, but critical task execution is still best left to human experts.

So, I have meandered around some of the themes of the summit, but what are these ten insights?

  1. We know there is a lot of confusion and hype around RPA, so we need to be clear where it fits into the workplace, in business processes and in our digital strategy. It is a tool, not THE tool. RPA is about productivity and freeing time and fixing issues around small, irritating, repetitive, error-prone tasks.  
  2. RPA impact – Whilst the individual productivity gains can be immense in percentage terms, be careful of the “magic of small numbers”. Work out what these gains mean at scale. You may have to balance investments between a 300% improvement in a small task, or a 3% improvement in an end-to-end process. Which is more significant? 
  3. Human centred design and “Humanistic AI” is the future. It is about augmenting human creativity, decision making and effort, not replacing the workforce with robots.
  4. The BPO firms are very active in RPA, as you might expect. Should they run your RPA program? What is the risk versus opportunity? Some feel their BPO knows more about their process than they do, but RPA eats away at the labour arbitrage element of the case for action for BPO.
  5. BPO and scale – the general view is that at large scale there remains value in BPO but at smaller scale, it may be that effective RPA eliminates the business case for BPO.
  6. Cost savings in RPA implementation need to be balanced against the cost of setting up and sustaining a Centre of Excellence and the overhead of governance processes. Team structure and placement for successful RPA is crucial, requiring a balanced set of business process understanding and skills with logic, data and systems skills. Engage IT, CISO and Audit early, as well as the business process stakeholders.
  7. The very term “Robotic Process Automation” is misleading. In reality it is “Robotic Task Automation”. Whilst the vision of RPA label may help a business case, someone will be held to account for delivering the benefits. Be clear that the claimed benefits are achievable.
  8. “RPA makes bad worse faster”, so be sure to understand the end to end process, the context, role and issues with the task to be “RPA’d”, ensure you have detailed desk/task level Standard Operating Procedures and test all exception conditions early. Bots have no judgement and will repeat errors at scale if rules are not well designed and monitored. Ultimately, remember that “the best automation is elimination”.
  9. Analytics – we have been discussing this topic for years, and it has been a core part of the value proposition of centralised shared services. Dynamic insight into process operation, performance, exceptions, outliers and improvement opportunities has been a much prized goal, but few have achieved it due to cost, effort and technical complexity. Artificial Intelligence and Machine Learning specifically offers a genuine opportunity to make this vision a reality.
  10. If RPA and Intelligent Automation with AI will help us do less of some activities, what could we do more of to drive value for the business? It is not a zero sum game, as the workforce must be part of the solution, not just part of the problem!

I came to the Summit feeling I knew quite a lot, and left having learned even more.

The more you understand, the more you realise what you don’t know! That must be why they say “ignorance is bliss”.

Thanks for reading, as always I welcome your comments, criticisms and thoughts…