Robotic process automation (RPA) and artificial intelligence (AI) appear to be at the peak of their hype cycle at the moment.  Software vendors, analysts, consultants, and the technology media have embraced the hyperbole as a great push for their products and services.  As I speak with business and IT leaders on the topic, one question everyone seems to be grappling with is how to sort through the buzz words and figure out which technology capabilities should be used for what.  I’d like to offer a simple paradigm for making sense of it all.

If we think about process automation occurring across a spectrum of work that spans from highly structured tasks to highly unstructured tasks, it is easier to understand why different types of tasks require different technologies.

• Deterministic Tasks – Non-invasive robotic process automation (RPA) solutions from the likes of Automation Anywhere, Blue Prism, UI Path or Assist Edge provide software robots to automate deterministic clerical tasks that can be expressed with a set of business rules. These solutions can be augmented with optical character recognition or handwriting recognition technologies to extend beyond machine readable data to include printed materials or hand writing.  More invasive BPM (Business Processes Management) solutions from the likes of Pega and IBM’s BPM provide robust process orchestration capabilities.
• Predictive Tasks – Our ability to automate predictive tasks has advanced significantly with the convergence of several different technologies. Low-cost sensors have spawned a flurry of activity around IOT (internet of things), with Gartner predicting 20+ billion connected devices by 2020.  Big data solutions like Hadoop allow companies to process massive data volumes to feed predictive models, from statistical analytics solutions like SAS Advanced Analytics, IBM SPSS, Mathworks Matlab, and The R Project for statistical computing.
• Cognitive Tasks – Recent break-throughs in artificial intelligence have brought automation of cognitive tasks mainstream with solutions like Alexa and Siri. These leverage natural language processing for task-based AI, and solutions like IBM Watson and Infosys NIA that leverage machine learning to identify patterns and automate more complex cognitive tasks.
• Social Tasks – The future of automation is around impacting human behavior through social AI. Research from Dr. Cynthia Breazeal, chief scientist of Jibo Inc. and associate professor at the Massachusetts Institute of Technology, has highlighted the power that robots programmed to provide non-verbal cues can have on influencing human behavior.  Her research has shown that augmenting a standard set of weight loss recommendations with a friendly, supportive presence to help people make the right decisions and form healthy eating habits, was able to sustain engagement with a weight loss program 40% longer than the same PC-based program.

Each of these technologies build on top of one other.  The information stored in big data platforms is used to train machine learning algorithms to make recommendations that can be automatically acted on through RPA.

(click image to enlarge)

Automation across the Spectrum of Work

As business leaders we must think smartly and swiftly how to leverage these emerging disruptive technologies to improve efficiency, reduce risk, accelerate cycle time, increase revenues, and accelerate working capital for our organizations.

With the vendor landscape rapidly evolving, it isn’t clear which solutions will emerge as clear winners, so it is important for decision makers to keep two keep principles in mind, in my view.

First, maintain a relatively short time horizon for payback/ROI.  And, consider the broader spectrum of work when making technology selection decisions, to avoid defaulting to a best-of-breed approach that could result in a sub-optimal solution.   This is the core of my guidance I offer to CIOs when engaging in these discussions!   Thanks for reading, and stay tuned for more insights from me in the weeks ahead.

This blog is the first in our three-part series on Automation.

Read Part 2: Picking the Right Automation Tool for Your Job

Read Part 3: Designing the Right Operating Model for Your Center of Excellence