Legal technology sponsored briefing: RPA in legal – be cynical of the hype and do not get left behind

 

Scott Littlehales of Littlefish on the importance for law firms of investing in automation

Picture the theatre of high-profile legal cases that feature in the mainstream media and on television dramas, and you can see why law remains an attractive profession.

But the screen-worthy glamour of boldly presenting arguments in front of packed courtrooms is a far cry from the reality for many legal practitioners. Below the surface of every law firm, even those practising case law, a high proportion of the workload comprises mundane background tasks, form filling, and many painstaking and, at times, unforgiving and copious amounts of research.

The same is true for robotic process automation (RPA) and artificial intelligence (AI).

Even the name is misleading, with robotics conjuring images of Hollywood interpretations of robots, marching in time towards the obliteration of the human workforce. Media hype tends to focus on a dystopian vision of the future where automated machines have ridden rough shod over humanity, giving rise to hotly debated topics such as the need for universal income as automation decimates the employment market, effecting an impact predicted to cut far deeper than the industrial revolution ever did and at a greatly accelerated pace.

Writers are already considering topics such as whether children should be taught not to be polite to voice assistants, for fear of breeding into future generations a collective human psyche of subservience to the all-powerful machine. The late Professor Stephen Hawking warned back in 2014 that the development of ‘thinking machines’ threatened the existence of humanity, our pitifully slow biological evolution allowing us to simply be superseded.

If we believe the premise of numerous articles, then we are standing two steps back from the edge of a very dangerous precipice.

‘Some firms are investing now and gaining a huge advantage over the laggards as they achieve efficiencies that positively impact the bottom line.’

The legal industry’s interpretation of the threat automation poses to job prospects appears to show little difference in some quarters. Grabbing the headlines back in 2016 was Deloitte’s prediction that the legal industry would suffer a 39% reduction in employment by 2036 with over 100,000 jobs becoming automated. The Law Society recently made a similar prediction: due to automation and process improvements, ‘by 2038, employment in the sector could be 20% less than it would otherwise have been’, according to its August 2018 Legal Sector Forecast.

But just like the gulf between high-profile, media-friendly cases and the less dramatic reality of most legal work, much of the opinion surrounding RPA is hype.

Perhaps the hype, and apparent misunderstanding of RPA, which is revered or feared in equal part as some mythical future being, is clouding judgement and preventing some firms from making sufficient investment. Failure to understand the basics and grasp the benefits of automation (and no, benefits is not simply a pseudonym or spin to disguise job losses), and a lack of knowledge of where, how and why to begin, engenders a tendency in some to ‘put it off’. But others are investing now and gaining a huge advantage over the laggards as they achieve efficiencies that positively impact the bottom line.

We prefer to call RPA ‘intelligent automation’ and refer to bots as the ‘intelligent workforce’. At their essence, they are simply some clever computing code packaged for the sake of simplicity and quantification into an intelligible concrete form, which once built and instructed, can automatically and digitally perform some tasks that ordinarily required human input. RPA is the grassroots of automation, but it is just the beginning of the journey.

Move up the automation spectrum from RPA and you reach machine learning and natural language processing (NLP), which will enable the processing of vast quantities of data in the form of legal documents and other unstructured data, presenting relevant cases and precedents to researchers, dramatically improving research quality and efficiency when preparing cases. Many firms are employing this technology right now and the immeasurable future significance of life-like natural language processing is demonstrated by the hiring strategy of tech giant Apple, which is currently targeting AI developers focused on NLP as a priority hire. Adding greater technical complexity, webchats and webbots are making automated user interaction and self service possible, while the ultimate destination of the automation journey is the much-misunderstood AI, which currently delivers the ability to let a machine make a self-determined decision, understand sentiment and give insight.

The potential implications and possibilities for the legal sector are vast. While it may sound like a distant destination, when considering mainstream services provided by the likes of Google built on AI and the rapidly shifting technology environment, what may seem unachievable today could be the ‘norm’ within a handful of years. But for many, focusing on pushing the boundaries of the future potential of automation and AI in the legal sector is attempting to fly before learning to crawl. Start simple, start small, prove return on investment (ROI) and value generation – then get creative.

‘The generation that takes slick digital operators like Amazon for granted is growing, and they demand and expect convenience.’

To think small, so we can begin to think more creatively, we need to get back to reality and go back to basics.

Away from the bright lights of case law, the buyers instruction form for a conveyancing solicitor is a mundane task but one that is essential to initiate the property-buying process. Consider a law firm that receives and processes just 25 buyer instruction forms each week. This is a staggering 750 minutes of human effort per week, and that does not even consider those forms that might contain invalid fields or errors that require further communication to and from the client for rectification.

The same task assigned to a bot from the intelligent workforce could comprise of one check every hour into the law firms email inbox, 24 hours a day, seven days a week, with each inbox check taking two minutes – equating to 48 minutes per day or 336 minutes per week. Next comes the form validation and processing, at five minutes per form. Assuming the firm continues to receive 25 forms each week, that equals 125 minutes per week. In total, processing the forms takes 461 minutes of machine effort each week. That is a 38.5% increase in efficiency. And that is just the start.

By checking the inbox automatically, forms emailed in the evenings and at weekends can be processed, moving the process along with no delay. Automated email acknowledgements can be sent, giving the client immediate gratification and reassurance, and the data can be input into case-tracking software so the client can see instant progress – delivering clients an improved user experience that satisfies increasingly high customer expectations. The generation that takes slick digital operators like Amazon for granted is growing as a proportion of the working population, and they demand and expect convenience in line with their everyday experiences of retail and tech in the workplace.

Invalid fields can also be addressed quickly, with an automated response highlighting issues to the client and enabling them to rectify errors. For example, a client sending a form containing errors on a Friday evening after the practice has closed, could be notified, rectify any errors and have correct data transferred to the next steps in the process before the weekend begins. The alternative could have been finding out on Monday, three days later (assuming the human managing the inbox was not ill or on holiday). And with accurate and error-free data inputting, with a clear audit trail to follow, there is no remedial action required to amend any incorrect data, or worse, incomplete or erroneous data causing costly problems at a later stage in the process, potentially when a contract has already been drawn up based on the flawed data.

Accuracy, efficiency and improved customer experience are not the only benefits.

More importantly, a human has been saved from 750 minutes, or 12.5 hours, of menial work each week – 35% of a typical employee’s working time. Over a year, even a conservative assumption on standard 9-5 working hours delivers a saving of 93 working days of effort. This saving can be framed as a cost efficiency, eventually leading to a human job loss – after all, it does not take many processes to reach a full working day, at which point the human can simply be replaced.

But think of it another way: imagine what you could achieve if you saved 35% of individuals’ time each week, starting from the bottom with your ‘low end’ activities and leaving the good stuff untouched? Imagine if that time spent on menial tasks could be replaced with greater time spent on value-adding, rewarding activity? Could fee-earners take on more work, or spend more time meeting clients face to face, further improving the customer experience and the revenue per partner? RPA has the potential to improve what work means to individuals, increasing job fulfilment and engagement, and ultimately delivering greater value to law firms.

‘RPA has the potential to improve what work means to individuals, increasing job fulfilment and engagement, and ultimately delivering greater value to law firms.’

If you are not already investing in a virtual workforce and taking advantage of the huge potential RPA offers in terms of efficiency, accuracy and improved customer experience, there is one thing you can be sure of: your competitors are.

Whether the scaremongering or headline-grabbing stories are justified or otherwise (the latest Law Society figures indicate a reduction of 9,074 FTEs, or 3.07%, in the legal sector between 2016 and 2018, and suggests automation is one of the drivers), legal is undoubtedly one of, if not the most turbulent industry today, with the most potential for transformation. As it grapples with the entrant of new business models, new methods of service delivery, changing customer expectations, increased M&A activity and the Big Four accountancy firms sustaining ascendancy, the industry is also now contending with transformation-enabling technologies that could have a wide-reaching impact for decades to come. And it is a wave of transformation you cannot afford not to be a part of.

Which begs the question, if you are not automating already, where and when do you start?

The first step is engaging with stakeholders from all departments and establishing which processes make good candidates for automation. The questions you need to ask of any process are:

• Structured process – could a set of task instructions easily be given to a new employee? If processes can be easily defined and communicated, they are typically good automation candidates.
• Defined input and output – is the data in a consistent and known format? Automation needs to work with data in a known format.
• Clear decision logic – could the decision points be defined and documented? Automations follow logic and defined workflows – they are not artificially intelligent.
• Definable workflow – is there an existing workflow guide or run book? A pre-existing one is not essential but can speed up the process.
• Uses multiple applications or tools – are multiple systems and applications required to execute the process? Processes that use humans as the glue between systems make good automation candidates.
• No subjectivity – is there room for ambiguity or emotion in the process? Processes requiring human judgement are not good candidates for ‘hands off’ RPA but might be suitable for ‘assisted’ RPA (ie goes to a human for approval).
• No spoken/handwritten inputs – are all inputs and outputs digital? RPA processing requires inputs and outputs to be digital only.
• Prone to error – are errors common and/or costly? RPA is human-error free so provides significant benefit.
• Unfulfilling – is the process mundane, repetitive and unsatisfying? Removing low-satisfaction effort improves satisfaction and makes better use of humans.
• High volume/low < medium complexity – is there a high volume of activity? Supports ROI.
• Low volume/high complexity – is activity less frequent but highly complex? Supports ROI.
• 24-hour operation – does or could the process run 24/7? RPA runs 24/7 so suits operations that already run 24/7 or could benefit from doing so.
• Suitable for self service – could the process be presented to users/customers and executed on demand? RPA processes can be executed on demand via a web portal, reducing human effort.
• Legacy or external applications – does the process use legacy applications and/or external systems (ie owned by a provider or customer)? RPA can automate processes that could not otherwise be automated.
• Variable/unpredictable volume – are there seasonal or time-based volume variations that could benefit from scalable resourcing? RPA can cope with demand variations better than the human workforce.
• Requires auditability – does the process require an audit trail for regulatory or compliance purposes? RPA provides consistent execution and a complete audit trail.
• Limited by human workforce – would running the process more frequently be beneficial but unviable for human resources? System checks and data validation are often carried out on smaller samples due to resource limitations, but these can be increased with RPA.
• Partial automation – is there a volume of repetitive or defined work that requires some human judgement to initiate, approve or define? RPA processes do not need to be 100% automated and can be initiated, approved or defined at any point in the process.
• Worth automating? – assuming it takes five days to build, test and release an automation – does that feel justified for the process? Given the broad potential of RPA, it is worth sense checking the potential process – if it only saves minimal time, are there other benefits that justify the investment?

Once you have established which of your candidate processes are suitable for investment in RPA, then engage with a specialist provider able to provide honest, open consultancy, to support you by calculating ROI and building a business case. Choosing a supplier that offers consumption-based pricing models reduces risk and ensures efficient expenditure.

Successfully proving ROI in one or a small set of trial processes should give stakeholders the confidence to support further investment in expanding the number of automated processes, which can build a platform for a more advanced automation, including user interaction and self service, and even the heady heights of AI.

Begin your automation journey with investment in simple cases with irrefutable ROI and who knows where those small steps might lead.

Go to the Law Tech Special main menu