RPA is Dead. Long Live Intelligent Automation!


I have been researching the notion of Intelligent Automation and, in particular, the rapid uptake of Robotic Process Automation (RPA) for more than 4 years. It was this work that eventually brought me to HfS. Over the years, I have made many good friends and acquantances in the automation community – and many of those good folks have graciously suggested I have become a spokesperson for this community. Yet, it is time to take a stance and declare: RPA is dead!

This is not meant to try to grandstand my esteemed colleague Phil, who’s eloquently stated that RPA 1.0 is a done discussion. Phil suggested: “We know what it is, we know what it can do, we know how it can augment operations and help digitize broken processes.” I rarely disagree with Phil, but I would argue that that broader market, outside of specialized services, technology and process areas, has no clue as to what RPA really is. We have no common reference points, we have no definitions, we have no clarity as to how successful deployments really are. What we have, is a set of reference technologies and respective case studies that demonstrate potentially significant efficiency gains, if implemented effectively. What we have is a set of innovative technology providers who use the term “RPA” to get a seat at the table to transform service delivery. However, to quote Lee Coulter, who chairs our Sourcing Executive Council, “In the context of automation we have a Tower of Bable; we have many languages, but we don’t really understand each other”.

RPA has dominated the conversation with industry novices and this needs to advance to the broader automation outlook

Yet, it is about much more than just semantic nuances. RPA currently is largely about the automation of tasks that can be implemted from the bottom up, then land and expand within enterprises. RPA is about rules-based processes. RPA is being delivered for specific accounts, often at sub-process levels, but does not sit at the heart of a delivery backbone. Nevertheless, RPA, for the wrong reasons, or just the lack of reference points, is the focal point for all the innovation that we at HfS tend to subsume under the notion of Intelligent Automation. However, if we move beyond the confines of the BPO and operations world, we see automation deployments at much larger scale – and beyond clearly defined processes. The current market development and opportunity in IT-centric automation scenarios in a lot more advanced, has more breadth and scale than business process centric scenarios, yet, in the discourse on Intelligent Automation these tend to get marginalized. It is in those IT-centric scenarios where the impact of Cognitive Computing and Artificial Intelligence is most pronounced. Yet, as an industry, we appear to be stuck in the “RPA” mindset, largely due to the heavy influx of marketing investments from the emerging RPA software suppliers which is influencing sourcing advisors and analysts new to the automation discussion.

Examples for a broader, more holistic approach to Intelligent Automation is Accenture’s artificial intelligence engine that provides an architecture abstraction layer for interacting with various autonomics services such as natural language processing and machine learning. Thus, underlying components can be swapped out according to client preferences or as new solutions become available, leveraging a broad autonomics ecosystem. Suffice it to say, not many providers have merged IT and operations like Accenture has done. But we also see a convergence of IT and business process centric scenarios starting to happen on a tool level. IPsoft and Arago trying to get traction in operations while the prominent RPA tool providers invest in analytical and cognitive capabilities. Thus, we are seeing virtual agents to gain traction of virtual agents in in operations and RPA tools in IT helpdesk use cases. These activities have to be reflected in the discussions on Intelligent Automation – not in isolated use cased but as part of a holistic strategy to scale out automation initives.

Therefore, as an industry, we really have to change the perspective. The focus should be on top down, the point of view of process owners and on taking a much more holistic point of view. That is what we had in mind when we launched the HfS Intelligent Automation Continuum back in 2015. All the approaches on the Continuum are both interdependent as well as over-lapping. Put in simple terms, life is complex. There is no such thing as a silver bullet or a turn-key solution. Intelligent Automation is about transformation. We have to wean ourselves off the drug that is task automation. Off course, it is prudent to automate low-hanging fruit. Though how do we optimize the processes after the effect of that drug starts to subside. Thus, we shouldn’t lose sight of the direction of travel. We at HfS call it the journey toward the As-a-Service Economy. Crucially, Intelligent Automation is just one building block among others along this journey.

Away from the headlines and away from the confines of business processes, we see some indications for market maturity as providers like Atos, TechMahindra and Hexaware start to standardize delivery on ServiceNow, linking it up with service orchestration engines that allow the plethora of Intelligent Automation tools to be plugged in. Put in simple terms, we see a much more holistic approach to Intelligent Automation. The crucial question is how do we cross fertilize all these approaches across the boundaries of trational and established business units? Flipping this to the client side, how are service provider supporting their clients with Intelligent Automation on their journey into the As-a-Service Economy? This is not about RPA alone anymore, but the broad notion of Intelligent Automation in all its complexity. To assess how far the industry has matured on this journey, HfS has just launched the RFI for the Intelligent Automation Blueprint to take stock where the industry really is at.

The Bottom-line: All the stakeholders in technology and operations need a much more effective education on the impact and potential of Intelligent Automation. Confining the discussion to “RPA only” is doing everyone a disservice

What I hope to hear in our discussions on that Blueprint is around broader notions of service orchestration, about integrating broad data sets, about moving beyond clearly structured processes. My gut tells me much of the future of Intelligent Automation will be about deep and unsupervised learning, about vertically-infused insights, about technologies for which we don’t even yet have monikers. To advance all those discussions, we urgently need a much more effective and actionable education of all the stakeholders. So once again I declare:”RPA is dead – long live Intelligent Automation.” But I want you to challenge me in that, be it to convince me that RPA is the best thing since sliced bread, deconstruct my arguments, unmask my assumptions. But most importantly I love hear about new approaches and new ideas.

Posted in : Robotic Process Automation



Leave a Reply

Your email address will not be published. Required fields are marked *

  1. Thanks for this Tom. Couldn’t agree more and many of our clients feel the same. So let’s elevate the discussion but, to Lee’s point, a common nomenclature would certainly help in the movement.

  2. Well, we just contracted with Accenture and after having sat on Autonomics I proposed five years ago, I do not care if we do it with Open Connect and add NLP and then ML, which is perfectly sound plan I have been working on for years now or we just let Accenture do it, assuming they can and we can afford it, I agree with you Tom, PRA is passé, let’s get on with Intelligent Automation!

  3. Thanks Mike and John,

    A team led by Lee is currently doing some sterling work to provide more clarity on nomenclature. But the lack of nomenclature is probably only secondary. More than anything, to John’s point, we need a switch in mindset and embrace the broad notion of Intelligent Automation. While the building blocks and technologies are interchangeable, the integration of those as well as the integration of data and insights will truly create value and offer differentiation.

    Doing the research for our inaugural Blueprint on Intelligent Automation at the moment, there are many shades of grey among the service providers in embracing that broad notion. Many are stuck in the old organizational models and thus in the "old" mindset: We have to stop retrofitting all the innovation like Intelligent Automation into the old organizational models. Only few providers have a clear and bold vision that is based on the fundamental assumption, that knowledge work will be disrupted and that many of the current jobs and activities will be a thing of the past. As an industry we urgently need a transparent and honest discussion on exactly that.

Continue Reading