Four Easy Steps to Intelligent Automation Success

Our work with hundreds of clients over the past three years to introduce robotic process automation (RPA) and intelligent automation (IA) programs into their businesses has given us a rather unique line of sight into some of the chief worries and perceived obstacles that many businesses have about their IA journey. So I thought it could be useful to share some key learnings from those experiences with those of you who may be contemplating further automation in your enterprise.

Let me start by saying that whatever issue keeps you up at night about IA, you are not alone. Reflecting on past experience, 90% of concerns are common to the early stages of most implementations and fall into four categories addressed here in the form of questions I often get asked:

If you see your own fears in one or more of these questions, it may be comforting to know you definitely are not alone. Dealing with these issues likely is less problematical than you may have anticipate and actually can present an opportunity to manage risks better.

{image 2}

Who should own and drive the process?

Based on our experience, ultimate ownership should reside with the business teams whose processes are being automated — finance, tax, supply chain, etc. — the business is making the investment and stands to benefit most. A business-led approach, which is IT-enabled with the IT teams supporting the business, rather than owning the project, helps accomplish the outcomes more effectively. Whomever it impacts directly should be engaged and involved right from the start, setting goals, driving each step, being accountable for the end result and achieving the benefits outlined in the business case.

As long as the pipeline for automation and associated priorities are conceptually designed and driven by the business unit that owns and understands the process, the fears along the way regarding whether the project will work and add value are minimized.

What should we automate — how far should we go?

There is no question that a lot of “stuff” in your business area can be automated, so the true question becomes “But why should you?” Just because the capability exists does not mean IA is the best answer. Would a heart surgeon replace valves if inserting a small stent would solve the problem? Be strategic about the opportunities IA provides, thinking about what it can do best --optimization, not just utility. Sometimes we find 80% of the value lies in just 20% of the effort and provides a better view of what should be done next.

One of the main benefits to be derived from an IA project is development of an intuitive design that makes for a positive user experience, which in turn makes for rapid adoption among your teams, scaling as needed, and sustained use beyond the initial implementation. The importance of a well-handled change management process at the design phase cannot be overstated. The temptation is to over-code the bot, when sometimes use of simpler existing tools — like an Excel spreadsheet —would be a better answer.

I advise clients to think like a businessperson, not a technologist. Just because the bot can handle 500 steps doesn’t necessarily mean it should be given 500 steps. Let the bot do what it does best and humans do what they do best. This simple rule will enhance the ability to update and tweak as needed, further contributing to steady use and sustainability.

Is automation — the bots — going to get me into trouble and create risks?

There is a great deal of buzz about cybersecurity and privacy issues these days, and rightfully so. But bots alone are not the culprits. It’s about how they are used, what controls are embedded into processes and how the infrastructure is set up. I advise clients to build into the configuration as many controls as needed to enhance the desired level of control.

Humans intuitively know how to protect themselves; that intuition should be embedded in the bot, since it will not “think” of it on its own. The reverse can also be comforting; a bot that can’t think to add a control also is equally incapable of forgetting to follow through on a control it was configured to carry out. Furthermore, with advance planning, the bot can actually help you meet the growing demand for assurances associated with security. Configure it, for example, to automatically send out that required validation or report, time-stamped and filed for future reference. In fact, the project can be an impetus to find new opportunities to embed desired controls and real-time documentation that the controls were actually executed, thus avoiding a rear-view mirror activity.

The beauty of tools like RPA is that one can get excruciating detail about what the bot does and when. Bot logs are like a telephone bill in that they share information about what was done and when. They also tell you which bot did it! This data and tracking can be incredibly useful in an RPA implementation, but rarely is available with a manual process.

How can we sustain and leverage our investment?

I saved this for last, but obviously it is of paramount importance. No one wants to create a “wild west” of bots with everyone using IA tools to do their own thing, possibly duplicating efforts and lacking consistent standards. Clear and strong governance is absolutely crucial to sustained success. Start with setting out a strong governance framework and a clear operating model with an engine that can execute desired policies and standards. Consistent coding, solution architecting and control standards make it more likely that best practices will be leveraged and value will be optimized.

IA is not without its challenges, but with thoughtful planning, it will prove well worth it, bringing your organization — regardless of sector or size — to a whole new level of sustainable efficiency, control and quality.

Sharda Cherwoo is EY Americas Tax Intelligent Automation/RPA Leader.