Walking Through the Gateway of Digital Transformation [ARTICLE]
Lee Coulter wonders if we’ve got the talent to do this while laying out how we can do this
CEO Ascension Shared Services
The state of the market is that some distinct patterns are emerging. The words Digital Transformation have been both extraordinarily ambiguous and remarkably ubiquitous in consultants’ speak. Now, they’re beginning to take some level of shape. What we’re really beginning to understand is that RPA - which is an incredibly useful harvester of tasks is an almost essential gateway to the more sophisticated digital transformation capabilities.
The state of the market is that there’s an emerging realization that most companies can’t “just go do” AI. Even to do IA, you probably don’t have what you need. People are becoming aware that they don’t have the right resources; they don’t have the right knowledge, experience, and infrastructure. This is a maturation point. For Intelligent Automation, we have crossed the Valley of Despair - as all new technologies eventually do - and we’re actually coming to the question now of “how do I scale this thing?” Increasingly, the conversations are moving to scale.
That means that for those that are awake and have a heart beat, they are aware of this thing Called IA or Intelligent Automation (which includes RPA and RDA) - they have an aspiration to move to AI, and they’ve probably stalled somewhere. From a recent study of automation programs, we know about half of all the adventurers that set out, are stalled. They stall for a whole variety of reasons. The conversation is shifting now to ‘how do you scale? There are some very important changes that need to take place in how you’re running your program if you want to scale.
What we also know is that to enable intelligence in your automation program means you are needing to bootstrap a business-rules engine into your automation platform(s) and using that to drive some level of intelligence into the task orchestration. This is also a maturation point for the IA industry. As a set of products and competing manufacturers, they are all evolving. Customers too are evolving.
The early adopters have been doing this for three, four years, or more and are now practitioners. We have tinkerers — they’ve been doing it for a year or two, and they want to become practitioners. We have folks that are just at the pilot stage and finally those who just learned to spell RPA.
The answer to the question “what are you doing with your IA platform?” continues to be a crucial important question. What use cases are you solving? I firmly believe that intelligent automation platform software is literally a box of possibility. People look at me cross-eyed when I say that, “What does that mean?” Well, what it means is the R - in ROI - the return, is undefined. It can be zero or infinite. It totally depends on the business. Not IT, not anybody else, not your partner or consultant. It relies on how many use-cases the business has queued up to move into automation.
The costs of an IA program are largely people based, not the technology. Technology is by far the smaller cost when you begin to get into a scaled program. Large programs may have as many as 65 or 70 people doing discovery, configuration, and production management. Technology will be less than 10% of the total program cost.
But another key cost is maintenance. It’s directly related to the issues around scaling. Robots don’t get sick, and they don’t go on vacation, but they do go on strike, and they don’t tell you why. They’re very sensitive to environmental change. Whether that’s process change or technology change, they’ve very sensitive to change. This is what I call a “Stall Point” – an inevitable challenge you will face when scaling. It is critically important you have a partner who can teach you how to detect and to manage environmental change at a level you’ve never managed before. Because once you’re far enough along, the
sensitivity to environmental change begins to go up - I won’t say exponentially - but it’s not a linear curve either. The ability to learn how to detect that and manage it is, in a lot of cases, one of the primary Stall Points that programs will inevitably confront.
There are several Stall Points every program will face. Another is what I call TAPY or Total Automation Program Yield – that is, “all in”, what are we getting in return. Eventually the business will ask what the hard-dollar benefits are, so a well managed program will always be concerned with some level of task harvesting for FTE savings and importantly measuring it with an approved methodology. This generates the core savings that pay for the parts of the program that are moving to strategic automation and eventually adding intelligence.
But once you move into IA - you’re now moving into a world in which you cannot proceed without IT at your side. BUT, with the right roles and expectations. You can do basic RPA with very limited support from IT. They need to provide you with infrastructure for the platform, security provisioning, and access management for the automation. The business can drive a very long way, with just that limited support. The conversations in the back-half of 2018 are shifting to, “Okay, so now I understand what it is, what it does, how it works, what is the right role of IT? What is the right role of audit?” I now understand these things. So, what does it take for my program to get over that exponential hill in front of me?
The role of IT very much changes, as you now want to bring pieces of real IT systems into your automation platform or program. Most enterprises begin the IA journey with a single platform - often an RPA platform. That is absolutely the right way to start. Harvest those tasks to pay for the more advanced stuff. Now you realize that you need to use advanced analytics to give your automation the smarts it needs to make a decision on whether or not to process a certain transaction, or fork left or fork right. Now you are asking to query live production systems for data to fuel intelligence and you will need IT with you for this.
Well, now you’re starting to change - you’re starting to move into this world of IA. It’s very rudimentary, but nonetheless, you’ve brought in some intelligence. It’s, “Hey automation, pick your head up, look. Is there a rock in your way? Or can you keep going straight? Basic automated decision making is the logical next step.
The manufacturers are starting to narrow the gaps in functionality. Some easing the connection to third party analytics and some offering “in-platform” machine learning capabilities. The market is realizing we need an easy start up, along with a pathway to advanced Intelligent Automation.
We’re also starting to see that the combination of multiple technologies is really where IA is achieving next-level success. There are some products out there that claim to deliver IA out of the box. I’m not aware of any production use cases I’ve seen that would convince me that that is a common occurrence. Not saying they’re not out there, just that I haven’t seen them first hand. There are advisors available that have truly been there and done that- and have five years more experience than you do - or six or seven years more experience than you do. I was recently talking to a very large services organization. They were proudly touting their top person as having two years’ experience. Trust me, they haven’t even gotten to the really hard Stall Points. We’re beginning to see some separation in terms of the kind of partners that you need to have to do this work successfully and well. You need somebody who’s actually run those traps before. Power point won’t do it.
I continue to believe and there is growing agreement, that RPA is the gateway to digital transformation.
It really is. It is a way to get in, get started in a way that’s very consumable, low-threat, and low
investment from a cost perspective. But it brings the organization to an important level of capability and awareness in a pretty short time - in as little as two years - to the point where they can now begin to invest in more sophisticated tooling, and then finally bringing intelligence into the automation platform. This is a non-trivial exercise. I can’t emphasize it enough. I’ve not seen anyone able to jump the basics. It’s like going to college when you didn’t go to highschool. VERY difficult to get a passing grade.
Particularly when you’re looking to disrupt work. It’s one thing to do a data science experiment, build a data model, get some predictive analytics, and then realize, “Well, that’s just a fascinating and cool intellectual exercise we have here.” You’re looking through a microscope and saying, “Hmm, that’s umm, really cool, but what do we DO with it?” As opposed to changing the way that 4,000 sales people work, or 4,000 wealth managers work, or 5,000 service delivery personnel work. Transforming work in a large enterprise is completely different than creating a machine learning model in a petri dish. For that, you will need the experience of a partner.