Why Is My Automation Program Failing?

He had his head in his hands, metaphorically speaking. As the executive sponsor for a major automation program at a large financial institution (let’s call him John), he was supposed to be the person to deliver greater efficiency, reduced cycle times and vast cost savings. His CEO had read the articles, attended the seminars and bought the hype – every firm on Wall Street was undergoing a massive transformation to a digital workforce, and he needed to do the same.

John built a business case based on assumptions of job cuts, assembled a team based on technology-heavy resumes, and selected a robotic process automation (RPA) tool vendor based on reputation. The initiative was given the green light.

As John spoke about how the program sprang to life, it was reminiscent of many conversations I have had with financial institutions around the world – not surprising when one considers the overwhelming case for automation in the industry. Financial services is brimming with high-volume, repetitive workflows performed by large numbers of manual resources. Sure enough, financial services companies are among the early adopters of RPA technology, which built the businesses (and staggering market valuations) of tool vendors such as Automation Anywhere, UiPath and Blue Prism.

The more obvious processes in the back office (HR, Finance and Accounting and aspects of Procurement) were quickly automated as were certain middle-office elements of risk management and treasury. A small number of progressive financial institutions even managed to automate isolated processes in the front office, for example in structured notes, foreign exchange and bond repurchase.

The case for RPA was clear. Increasing throughput, reducing errors to improve accuracy and driving efficiency while improving risk posture – these were the prizes that attracted John to push ahead with the automation program. As he set out on this journey, he wasn’t concerned that the initiative was highly visible to the executive board. On the contrary, he fully expected this to further his own career aspirations. The phone call he just received from the CEO, however, was short and unpleasant.

John is not alone. Our research suggests that almost 80 percent of financial institutions have implemented RPA to some extent, with much of the “low hanging fruit” now automated. So why is John not being feted as a hero? The truth is that the large majority of automation projects have not delivered the expected return on investment, whether that return is defined in financial or non-financial terms. Our work suggests there are three primary reasons.

1. Expectations are too high. Before understanding the nature of RPA, many executives went on record announcing their intention to replace thousands of humans with digital labor. This was a fundamental misunderstanding of the nature of RPA, which tends to automate tasks – not roles.As a result, many business cases that were predicated on significantly reducing the workforce were never realized.
2. Many organizations hit the “RPA wall.” After an initial burst of activity, automation initiatives often peter out due to multiple factors, including a lack of immediate success, inadequate in-house capabilities and a sub-optimal (or non-existent) automation center of excellence (CoE). Our research shows that almost 60 percent of organizations have automated fewer than 20 processes to date. Only 12 percent had automated more than 50 processes. An inability to scale represents the single biggest reason why enterprise-level benefits are not being realized across the financial services industry. Another indicator that RPA progress has stalled is that average bot utilization is hovering around 50 percent, suggesting that many firms have been unable to move beyond the “one bot, one automation” mindset, and do not have a strategic plan (and the associated governance) to effectively scale across the enterprise. This is certainly the case for John, who has encountered both these impediments.
3. The cultural change is more significant than many realize. Moving to a hybrid digital-human workforce is about much more than implementing RPA software. Most financial institutions fail to recognize the need for a well thought-out, robust organizational change management (OCM) plan, and its absence means that every individual employee will struggle to make the link between automation, their job and the company’s goals. Of those institutions that are considered market leaders in automation, a strong OCM plan is always a telling characteristic.

As I describe these impediments to success for automation programs to John, he adds two further reasons his own program stalled. The pace of change in the industry means that the term automation now includes far more than RPA; his CEO and lines of business are exploring myriad other “intelligent automation” technologies, from natural language processing (NLP) and optical character recognition (OCR) to machine learning (ML) and genuine artificial intelligence (AI). This means John’s team is expected to be capable of assessing, selecting, procuring and combining all these technologies in a way that is seamless to users and customers, but his CoE and associated program were not built for this spectrum of tools.

Finally, John underestimated the degree of conflict that would occur between his program team and the IT function. Even as he was rolling out RPA pilots in operations, the IT team was telling him they could provide the same functionality with existing tools and, in some cases, they were openly obstructive to the program team. This is a common story among financial institutions. Of course, automation initiatives must rely on underlying infrastructure and integration with other systems, but first and foremost it should be thought of as a force for good within operations and the lines of business.

Here’s what I told John – and what I’ll tell you – as the three most important things to do now:

1. Reimagine the CoE. The CoE needs to define the overarching automation strategy from RPA to AI, including the implications for an organization’s operating model from back-office shared services to front-office interaction with customers. It’s the CoE that should lay the foundation for building and supporting automation, establishing KPI metrics and reporting standards that highlight value to the business and, in turn, drive re-investment in automation. This will help the organization break through the “RPA wall” by increasing bot utilization and working across lines of business to establish a pipeline of prioritized automations based on likely value.
2. Leverage intelligent automation to uncover new sources of business value. For example, don’t think about automation purely in terms of cost savings, but consider how it can help improve the experience of your customers – reducing the time it takes a customer to originate a mortgage or receive a contact center response. AI can be deployed to cross-sell / upsell based on data analytics and can predict (and therefore avoid) deficiencies in new product design to hasten the time to market. In the middle office, machine learning can reduce false positives by 95 percent in the anti-money laundering process, and AI can transform a firm’s know-your-customer effectiveness.
3. Design and deploy a strong OCM plan to run alongside the automation program. No major transformation program is optimized without an OCM plan to align expectations, identify required skills and training, develop the optimal communications strategy and establish a formal governance plan to scale effectively. This will require expertise that may need to come from outside the organization.

Even hearing some clarity on how to remediate the current impasse made John feel better. He realizes what he needs to do to recharge the program and re-engage stakeholders across the organization. Above all, he knows that his career prospects are inextricably linked to his ability to overcome the impediments and drive measurable value. If he can achieve this, the next conversation John has with his CEO will be a lot more pleasant.

ISG helps financial services enterprises leverage intelligent automation, build an effective CoE and manage the organizational change to make an automation initiative successful. Contact us to find out how we can help you.

About the author

As the Lead Partner for Banking and Financial Services at ISG, Owen Wheatley has more than twenty years of industry experience and has responsibility for senior client relationships, business growth and delivery excellence through the management of high-performance advisory teams in multiple regions. He has served major banking and financial services clients around the world, leading major transformation programs, advising on complex technology strategies and continues to be a widely published thought leader in the industry.