In the UK and US there have been many recent publications and conferences about Robotic Process Automation (RPA) and the potential influence on the job markets. Unfortunately, so far this has not been the case in continental Europe.
Originally, outsourcing was meant to reduce costs by using low-wage countries (off- and nearshoring), but the aim was to also get access to skills which were not internally available, to innovate processes and practices and to free up internal resources.
This set-up has come under pressure not only because of local loss of jobs but also because of dissatisfaction of customers caused by cultural differences, perceived low quality and high attrition rates. Expectations have not been met.
Many organisations outsourced work and sometimes whole departments dedicated to performing standardized, low-level routine tasks including manually copying and pasting data from one system to another. By giving this work away, they transferred any connected problems as well – one of the common mistakes made in the early days: never outsource a problem – it will come back bigger!
Still, businesses continually seek new lower-cost countries with low cost, low-skill workforces (so called FTE – Full Time Equivalent – or transaction-based outsourcing).Whereas the middle and back end of the crowd is still running into this direction, many businesses are now starting to hear the calls from the front: high volume (including a high volume of people) does not produce value, or sometimes any.
Conversely we can observe that previously outsourced business activities, particularly those that interface with customers directly through contact, shared services and delivery centres, are significantly moving towards being on-shored and in sourced over the last five years. This remedies many early outsourcing mistakes for companies that have experienced only the above mentioned problems.
But reshoring or bringing the functions in-house again means another major, disruptive, and expensive transition: reorganisation will have to be made, new or lost processes need to be built again and executed, specific roles and responsibilities have to be defined and then either new staff hired or existing personnel retrained. Loss of skills and knowledge is also a risk which is often overlooked in the first step.
This is why many customers are confused: a) the old set-up does not produce the anticipated advantages, b) bringing the function back in-house might be too expensive so what might be the alternative?
Here Robotic Process Automation emerges as a new, reasonably mature alternative available on the market. But, what exactly is it? It is definitely not R2D2 wheeling through the office although the name suggests that robots take over administrative tasks. We are talking about software which is able to mimic human interaction with different IT Systems by following (“learning”) routine tasks and then executing them independently. Such tools are bridging the gap between disparate, non-integrated IT systems. The problems of cultural differences, low quality and attrition should disappear, because the goal is to relieve humans from standardized, low-level routine tasks, such as the aforementioned manually copying and pasting data.
Robotic Process Automation corresponds to an emerging trend for technology to replace at least some functions performed by humans, particularly in the service sector – and this is why it is now so much discussed in the context of Outsourcing / Offshoring of Business Processes (BPO). At what point this work would or could be completely taken over by machines or software is a question of availability and balancing costs.
Why tolerate issues with consistency, quality and attrition rate if software does this work for you? Looking at contemporary call centre technology, the intelligence sits already in the (CRM) system – soa highly skilled workforce is no longer necessary. On the other hand, front office work with direct customer interaction is a classic instance in which you can’t just replace headcount with software. The problem can be the process itself.
Some of the fundamental problems with business processes and outsourcing are not caused by technology or by the people, who are disgruntled by repetitive and in the end compartmentalized to meaningless work, but with the compartmentalization of work itself.
RPA without intelligence will only support the reorganisation of the parts of work that consist of routine tasks. Actually, outsourcing already did this to a certain extent: dividing the work into routine parts which were offshored and the intelligent part, which stayed in-house at the customer’s site. Practically, business processes, regardless of whether it’s human resource administration, sales or even activating SIM cards (Telco industry specific sub-process), are defined through instructions for every possible person which often means an excessively fragmented, detailed, and controlled process. Most of those processes are mundane but – according to the outsourcing contract – need to be executed step by step without any shortcut or deviation, to measure the amount of transactions, time spent and in the end proof that SLAs are met.
Years ago we saw this phenomenon described in a satirical way (“Lean Brain Management – more success and efficiency by saving intelligence”, Gunter Dueck, 2010 Springer Verlag) – and now with the maturity of software, we find outsourcing at the intersection of “Lean Brain Management” where the intelligence is in the system, which is itself intelligent and “operates with robot-like, cheap standard humans who are best when not allowed to be intelligent. (…) In Lean Brain Systems, only the system shines, and nothing else.[i]” You see where RPA can lead to if not done in an intelligent way?
This development is actually a part of global standardisation and commoditizing of people’s work and adjusting them, and their needs, to IT – not adjusting IT to meet people’s needs.
There are some undesirable outcomes we experience as customers when we are routed to call centres:
Brainless obedience of “good practice” without considering whether it is appropriate for all the individual circumstances has limitations. The scripts which follow on dealing with a complaint, with a sales pitch, or finish with “is there anything else I can help you with?” when no help has been offered.
Unintelligent adherence to process – help desks going through standard scripts irrespective of the information the customer provides; or directly pushing back the work to the customer in the way of “self service,” which simply translates into “let the customers do the work for you.”
The current status is exactly why more and more customers prefer to be serviced by machines or using social media instead of jumping through such second rate “process loops.” In my opinion, this push to social media (Twitter) is a result of simple necessity. Current research suggests that customers use more channels to communicate. In intensive consultation areas like banking and insurance, customers prefer to have face to face meetings within their local branch [ii]instead of being served via any kind of indirect media channel where ease and elaboration are combined.
Would you like to discuss remortgaging your home with a faceless computer or an anonymous remote call centre? If it’s the former, feel free to do so, but just be prepared that the response is often that the “computer says no,” an actual response that the former Chairman of the Federal Reserve Bank in US (Ben Bernanke) recently had as a firsthand experience. He uses this now as an anecdote in his speeches. Well, if you don’t earn $200,000 for a speech, you might be less enthusiastic about automated credit checks (a sub-process of the finance industry), executed without intelligence and lacking context, but delivered cheaply following an automated business process performed from somewhere remote.
So in which ways will RPA be useful?
It should have been done before, but if it hasn’t or has been put in badly, now is the time to implement it.
Investigate the business, including the processes. Identify the goals which that need to be achieved. According to case studies from Genfour many companies do not have a particularly good insight into their own processes or in the different workarounds within the IT landscape which need to be made. Many companies are also not aware of existing analytics software tools which bring transparency into those processes. They could have already been used when working with an outsourcing provider to get more real time information about the status of SLAs, but this is a different matter. Then, target the areas for automation and build a benchmarking system. You would rather not get an increase in service quality replacing superfluous, brainless, rules based process steps especially when those have been created without taking the human qualities into consideration. It’s crucial to identify the right processes and cut down to the really necessary execution before implementing RPA, which also needs to be aligned with the corporate strategy. You might not need to change a customer’s address three times in three different systems, neither through human intervention nor through a robot. Do it once, and do it right the first time. RPA and outsourcing in general needs to be aligned with the overall business strategy.
Widely discussed in the media and in specialized interest groups, but hardly even recognized by customers is the question: will RPA make outsourcing of centralising functions such as key ITO, BPO, shared services, call and delivery centres in low-wage countries obsolete?
According to a recent survey published October 2014 by SSON, 33.1% of CEOs regard cost reduction as top priority for shared services centres but 29% have not even heard about Robotic Process Automation. We need to ask the same questions and face the same results as we did with standard outsourcing. So what will RPA do here? If those centres are mainly concerned with rules-based, excessively defined, standard, high-volume and repetitive tasks the population will not be engaged or actively disengaged. Outcomes will vary from high attrition rates to high profile cases of dishonest employees selling customer data, or participating in other unethical or even unlawful practises such as fraud and bribery or blasting customer relationships on purpose. This has raised questions about the risk levels in many offshore locations, but not concerning the set-up itself.
For many years we see employee engagement surveys finding a measurable correlation between disengagement and failure of services. You can imagine or have probably experienced results of frustration and disengagement: rumours that ruin reputations, poison relationships and halt careers. As a result individual performance plummets when people are more engaged in passively creating worst case scenarios for themselves and the company than doing the actual work. Or they get actively involved in political fights about who is allowed to use this exact coffee cup and why does the other person use the printer so much.
Is it possible that humans are divided into two categories: employees and customers? Whereas customers are the most important thing in business, employees are just a replaceable asset it seems: current call centre work is often intelligence-wasting, oppressive and puts qualified people through conditions where they are forced to work under “Big-Brother” style control. This is often discussed in the context of location in India – logically as UK and US use the English capabilities there, but the conditions remain comparable also in other lower-wage and near shore locations.
Do you think that people from anywhere in Eastern Europe (CEE) or South America are more thrilled and motivated than Indians when the time spent on single transactions or even in the bathroom is measured? Check anonymous employee feedback tools such as glass door. Employees describe those exact conditions in centres, regardless of the location and regardless of the companies they are working for.
And if strategy is not aligned, it might be that the legal function buys then legal process services, such as multi-lingual document screening for a due diligence project from an external provider, where there is a multi-lingual shared services centre readily available. One hand does not know what the other is doing. This service should and could have been performed in the established centre, particularly as those are single, short term projects, which could be taken onto balance workload in low-peak times.
RPA itself will take away the standard, repetitive tasks, but not change the working conditions. I doubt that attrition rates will decrease significantly and consequentially there will be limited potential to allocate sophisticated work into such surroundings – besides the fact that trust will never grow properly (if they can’t perform the low-level work, we do not give them the high-level tasks). If the workforce is completely new every two years, the cost case is hardly convincing any more (hiring, training time) plus you cannot rely on building capabilities for human judgement in the process. How could you if there is limited experience?
If tasks can be automated, what will be left over, what skills are required and how will work in general look?
Actually I think we have arrived at another important crossroads. Should we go ahead with “Lean Brain Management” and wait until Artificial Intelligence (AI) software is ready for production and available at a reasonable price? Which means the remaining tasks with the human judgement factor would then be taken over by systems which are able to process natural language (IBM’s Watson or IPSoft’s Amelia)?
In order to be successful, business processes and business process outsourcing must be analysed and worked back from the results. Ultimately business will change when holistically focusing on relationships, rather than only implementing automation tools.