In March 2017 there was a headline in the Daily Mail which said ‘Will Sam the robot kill off the brickie?’ and then 2 weeks later it seemed that Sam the robot was out of a job as a house was ‘printed’ in concrete.
Earlier this year it was reported that a McDonalds restaurant produced a burger to order with no human involvement and a well known high street coffee shop would no longer need Baristas as the latte could be produced by a robot. You wouldn’t even need to order it as it would recognise you as you came into the shop and have your ‘usual’ ready by the time you got to the counter.
All of these activities have one thing in common which allows these roles to be automated – a predictable outcome – the robot / process knows what it has to do and therefore does it. If you then apply this rule to everyday activities in the workplace, the potential for automation is staggering and the future impact on the workforce is significant. The impact isn’t really known as this sort of automation in the workplace, in the grand scheme of things, is in its relative infancy, none the less it is here and Oxford University researchers have estimated that 47% of US jobs could be automated in the next 20 years, another estimated that 30% of UK jobs could be under threat.
We are according to many in the ‘2nd Machine Age’. The ages of man used to last for thousands or hundreds of years like the Ice Age or Stone Age, but with man grasping new technology these ages are getting shorter. In the last 100 years we have had the ‘1st Machine Age’ (early 20th Century) the ‘Space Age’ (which started in 1957), the ‘Information Age’ or ‘Internet Age’ which is a term used for the 21st century and the speedy movement of data globally. The Information Age is technically only 17 years old and yet academics and economists are now saying we are moving into the 2nd Machine Age.
According to the author of the book ‘Second Machine Age’ Andrew McAfee, he argues that this age involves the software based automation of tasks that replace labour unlike the 1st Machine Age that complemented labour. The first labour casualties will be the more unskilled workers where the work is more predictable and repetitive and then moving up the scale as technology develops. Examples of this would be farm labourers but we see now on-line diagnostic tools for illnesses and more proactive and objective tools for analysis of business performance, stock markets and even grading exam papers. So what does this mean for Credit Management?
Many of the tasks we perform in credit management, let’s be honest, are predictable. Credit Assessment taking data from different sources to understand business performance has been available for some time and these tools, ‘what-ifs’, scorecards and predicting failure based on trends are getting better. Simple Collections activity has been automated in a complimentary way with systems generating letters and requesting calls through workflow again these have been around for a while as has automated Litigation and automatic Cash Allocation.
The differentiator comes when these activities and the software that we use to drive them are linked together and they begin to learn what happened previously and then apply these learnings automatically. Cost reduction in the workplace involves all sorts of activities, process improvements, waste reduction, but the single most expensive thing is still people. Getting the systems to drive the people has been a consulting mantra for some time, and moving people away from contacting everyone on their ledger to managing the exceptions has allowed large organisations like utilities to manage millions of customers without aircraft hangers full of people. The question asked by people who don’t understand operational customer management ‘What is best practice in terms of number of customers to a credit controller?’ is becoming an even more ridiculous question.
Behavioural Scoring and Predictive Analytics have been around for a while, banks and financial institutions have used these for years to understand spending and payment patterns, this is why you get an unsolicited gold credit card application through your door… or not…but these tools are becoming more commonly available and cheaper so can be used in everyday smaller businesses and with the developments in cloud technologies and application based development what was once impossible, not least due to cost, is now available on ‘pay-per-click’.
As mentioned the real benefit comes from linking these various customer lifecycle activities and processes together. So what if you understand the payment method, the point in the cycle, what and when a customer pays? Organisations follow a pattern so the outcome is predictable they have a standard process, a payment method and a timeframe or internal payment term that they follow, in many cases what you do makes little or no difference, unless there is a problem. Add this knowledge to your risk assessment, propensity for late or non-payment and you start to see the potential.
Software companies just in the collections space suggest that 70% of the collections effort is wasted, all we have to do is identify where to put that 30% of effort to get the same result. Most organisations have this data somewhere, the trick is getting hold of it in the large ERP systems which many organisations have. We have all heard the experienced credit controller say ‘He always pays on the second of the month’ capturing and managing this knowledge, understanding why and then build a system that captures this the allusive 30% becomes more obtainable.
Specialist systems managing specific activities are built now with this level of information available for extraction. A great example of which is in cash allocation. Studies show that 85% of organisations use ERP systems for allocation with the required human intervention to perform the simplest of allocation tasks. These ‘Jack of all trade’ systems only achieve around 50% to 55% auto allocation, but smaller specialist ‘best of breed systems’ dramatically increase this auto allocation to over 90% also completing the task faster.
As mentioned many Credit Management processes are predictable and there are a number of outcomes that we expect and with smart systems that learn these things tasks can be automated. The founding CTO of Uber Oscar Salazar said at a technology conference in Los Angeles a few weeks ago ‘We are all responsible for adding technology to a society without thinking about the consequences.’ and that ‘…it may not replace jobs but it will change industries…’ and we all know that Darwin quote that it isn’t the strongest species that survives but those most responsive to change.
As Credit Management professionals we should try and understand the impact of this sort of Robotic Automation technology. For example; Will it be necessary to offshore / outsource to reduce costs now if we can do this better and cheaper ourselves? We can automate the ‘call centre’ type low value high volume collections, how far up the value stack will this go? What will be the staffing and capability requirements of my future credit management team? Where does this leave SSCs, training, process management? What is the next generation of measures required to ensure effectiveness and performance?