A Podcast Series

Podcast: The Role of AI in Business Process Automation


Companies have long been interested in enhancing business process automation using artificial intelligence. The disruptions caused by the COVID-19 pandemic have raised the stakes, amplifying the need for AI’s use.

In this RTInsights Real-Time Talk podcast, Joe McKendrick, industry analyst at RTInsights and Connie Moore, Vice President and Principal Analyst at Deep Analysis, discuss the challenges and general benefits of using AI in business process automation that looks at how process automation can help companies during the current pandemic, and then look at how the technology can be used post-pandemic to deliver even more benefits to an organization.

Listen to the podcast now:

Download podcast

Subscribe to Real-Time Talk on your favorite platform:

About Connie Moore:

Connie Moore is Vice President and Principal Analyst at Deep Analysis. She joined the firm after four years as Senior Vice President, Research, at Digital Clarity Group, and more than twenty years as Research Director and Vice President at Forrester Research. Connie is a widely acclaimed speaker, advisor, consultant, and expert in digital process automation, customer experience management, digital experience platforms, and content services. In 2014 Connie received the Workflow Management Coalition’s globally recognized Marvin Manheim Award for influence, contribution, and distinction based on standout contributions to the field of workflow and business process management.

Read the transcript:

Joe McKendrick: Hello, and welcome to the RTInsights Real Time Talk podcast series. I’m Joe McKendrick, industry analysts at RT Insights, and your host for today’s discussion about real-time AI, analytics, IoT trends, and solutions that provide significant value to your business.

And today I’m really pleased to be joined by Connie Moore, Vice President and Principal Analyst at Deep Analysis. Welcome to our podcast, Connie. Glad to have you here.

Connie Moore: Thank you.

Joe: I’d love to learn more about what you do, your work. Could you give us a brief background on what brought you to Deep Analysis and your areas of interest?

Connie: Oh sure. When people ask me my profession, I always kind of stop because I’m really a business technology research analyst and management consultant. That’s a lot to say. I joined Deep Analysis a year ago. And prior to that, I was senior vice president of research at Digital Clarity Group where we specialized in customer experience. But I also was forced to research for more than 20 years. And I managed the chains that focus on enterprise applications, business process management, data management, and content management.

I’m interested now, this moment in time, in the intersection between customer journeys, which are outward, and operational processes, which are inside the organization. And I’m also creating eBooks and infographics about low-code automation, citizen developers, process mining, and the intelligent business automation platform we’re going to talk about later. I’m also particularly interested in helping women forge their careers in technology. And so I write about that on a fairly constant basis.

Joe: Fantastic, fantastic. Wow. And you’re really on the cutting edge of many things there. You’ve been there it sounds like, for some time. And now it looks like the area of focus is digital process automation. And that really sounds like when we talk about artificial intelligence and machine learning, we’re talking about where the rubber meets the road. Where the actual work takes place, the actual implementation work. Can you discuss that? What do you see as the opportunities for AI and machine learning as it relates to digital process automation?

Connie: Well, it’s interesting because AI and machine learning is already pervasive and we just don’t realize it. Some examples in different processes are detecting fraud, recognizing a facial pattern, recognizing voice, and sentiment analysis where the software looks at content and determines the tone, opinions, and attitudes in that content.

So a great example of AI and machine learning in use right now within digital processing is next-best action. And that’s where the contact center reps are guided how to respond to customers in real time. Another example is natural language processing that recognizes incoming unstructured content, automatically assigns that to a case for process automation.

Or one of my favorites is how natural language processing allows individuals designing a business process to do this in plain English to write text. And then the software turns that into standard workflow actions and builds the skeleton of the business process. So there’s many, many ways AI and machine learning can be applied.

Joe: And that whole area, natural language processing, we’re seeing it as consumers with Google Home and Amazon Alexa and so forth. And it sounds like it’s taking it even a step further. Not just simply answering questions back to you or providing information back to you, but actually putting what you seek into action. The process you seek to engage with.

Connie: Absolutely.

Joe: What about the vendor landscape? How quickly do you see digital process automation vendors adopting artificial intelligence and machine learning into their product lines? Is this something that’s happening out there or is there something holding the vendors back at this point?

Connie: So the examples I mentioned in the previous answer, are examples of machine learning in use in businesses. But if you ask the vendors about their AI and machine language strategy, you’ll find they’re interested, but they don’t really have much to say in terms of what’s in their products today, what’s on the product roadmap, or what their strategy is. And there’s some real reasons for that. First of all, they don’t know what customers want and when they want it. That’s very uncertain to them. But it’s also hard work because we see AI being used in the consumer landscape and that’s running on platforms like IBM, Salesforce, Google, Microsoft, and Amazon. But for the process automation vendors to get to machine learning in their own software products, they have to build on top of those platforms, adding their secret sauce and then adding their own data and training their product to use that data.

But the problem is that no two AI applications are alike. So each place in the software where they want to add intelligence may require its own unique solution. They all have unique data and unique learning experiences. So you could possibly use similar building blocks throughout your software product, but then the data’s train-down might vary and be quite different. So that’s some of the things holding them back.

Let me just go into the Holy Grail. The Holy Grail in a process automation software product is that an event can come in from dynamic case management system, and that automatically triggers human actions and automated systems working on that machine advice, and then creates new processes and new actions on the fly. So we translate all those words into an example, a system that users’ intelligence could detect an intruder based on inputs from IoT devices and applications.

And this automated system could dispatch a guard to look into the situation while at the same time updating other people’s systems and devices for further action. That’s an example of AI in use within a process. Or I can give you a real-world example in healthcare. The intelligence system predicts whether the patient will be sent home, should be sent to the emergency room, or sent immediately to the operating room, and makes recommendations based on those predictions. And then that also affects what probabilities are of missing a deadline and the financial impact of various outcomes. So it advises healthcare administrators on the next best action for the case. So those are some real-world examples, but it’s really slow for vendors to implement in product.

Joe: And for our audience Connie has authored an article on the deepanalysis.net site, deep-analysis.net, entitled AI and Digital Process Management is Complicated, that discusses a lot of these points. The fact that AI itself depends on who’s implementing it, right? It’s not a general type of application.

Connie: Exactly. That is the issue. And when we see it in the consumer world, we assume it’s an easy thing to do. But when you apply that to business processes it’s actually quite complicated.

Joe: And Connie, you also recently co-authored a report, the crisis—we’re talking about the COVID crisis—crisis paves the way for supply chains to go digital. And you wrote that many supply chains still rely on paper documentation, which I was actually kind of surprised to hear. I guess I’m not deep enough in the supply chain world to have witnessed that these days. But there’s still a lot of paper flying around. Why have organizations resisted digitizing this vital part of their business for so long? And is COVID kind of bringing that all to light? Is COVID kind of forcing the issue? Forcing digitalization.

Connie: It is forcing it. But it’s once again, a complicated thing. So the reason it seems like they’re resisting it is that companies in the supply chain have focused on inside their own operations. And the weak point in the system—besides being paper laden and manual and so forth—the weak links are the handoffs from company to company. So there are many suppliers, there are many partners, there are many participants in the chain. And every time you have a handoff across one enterprise to another that’s where really it’s very vulnerable because there hasn’t been enough attention on that.

Another reason why it’s very paper intensive, and very manual, and handing real folders from one desk to another, is that there’s been a lack of trust because first-generation capture technology, so when you send content from one business to another, the automated capture technologies weren’t sufficiently accurate or sufficiently efficient for people to trust.

And so they stayed with the tried-and-true approach—a paper document with a physical signature. But that’s really changing now because of IoT devices that are networking. Blockchain is really important from a security point of view. Process automation and AI oriented toward intelligent capture. That is shifting the landscape.

And so now that it’s more resilient, you don’t need that signed delivery note as proof that the goods have arrived and they’re in the right location and you’re in the right condition because the automation is more trustworthy. It’s sort of like in our everyday lives, at least for me, when I go to the bank and I withdraw money I don’t ask for a receipt. Or I fill my car up with gas, I don’t need the receipt. So that trust level has risen.

But it’s going to take a significant amount of investment, and resources, and capital to tackle these still heavily manual processes that are paper intensive. That’s why I think RPA has such fast acceptance. But on the other hand I have to say that it’s more of a band-aid solution. So companies are really going to have to look at the process and the handoffs, the technologies involved and not just stick on a band-aid solution of RPA.

Joe: You recently talked about in another report the rise of the intelligent business automation platform. It sounds like that might be what you’re talking about going beyond that band-aid approach. But how does that differ from what we’re seeing out there nowadays? Salesforce.com has this incredible platform. And SAP also has this very sophisticated ERP-and-beyond–type of platform. Can you discuss what you mean by intelligent business automation platform and how that differs from what we’ve typically seen on the scene?

Connie: So the enterprise applications vendors—Salesforce, SAP, and so forth—have really some impressive and very, very usable technologies within their platforms. But they’re targeted at the work that’s being automated by those enterprise applications. So for example, in SAP you can get business process automation, you can get RPA, but it’s within the confines or within the actual enterprise application itself. So what companies are finding they have problems with is all the work that got left out of those applications. And those activities are either manual or semi-automated or poorly automated and disconnected or not integrated. They have a lot of problems associated with them. And that’s where the digital process automation vendors and the low-code vendors are focusing is all the work that sits outside the enterprise applications. And that’s why RPA once again, is so hot. Because companies can solve some of the manual—really menial—activities. But that’s a tactical view.

A better view is to look at these processes in their entirety. They may be operational in focus. They may be customer focused. They may be ideally in the end focus, and look at using a number of technologies for the solutions. So we believe that as you look towards the future, the intelligent business automation platform is going to not only focus on process, but focus on machine learning, focus on cognitive services, focus on customer engagement, which so often gets left out when you’re looking at operational processes inside the organization, looking at contracts and pull it together. Content, collaboration, the actual packaged business app that’s process centric. And pull together the customer journey map and integrate that with process models and process mining. We haven’t talked enough about process mining. I think that that’s going to be core to the intelligent and business automation platform as well as what do you do with this data? So some of the vendors are looking at adding data lakes for storing the process data that the machine learning system uses.

The process will remain incredibly important, but there are all of these other technologies that enhance the process, make the process better. And I think that we’re going to pull back and look at this from a big picture from the platform perspective rather than focusing more on point solutions.

Now, this platform I’m talking about could be a collection of best-of-breed tools, but I think a lot of these technologies belong together and work better when they’re integrated, and work better when you have that end-to-end view of the customer engagement across all aspects of the company. So we think it’s going to be a platform that vendors work towards creating. And some of them are well on their way in this effort.

Joe: Okay. Fantastic. Fantastic. Well, Connie, I want to thank you for joining us today. It’s been great having you on our program. And again, we’ve been talking to Connie Moore, vice president, principal analyst with Deep Analysis. And the website again is deep-analysis.net. Thank you very much for joining us today, Connie. This has been great.

Connie: Thank you.


About Joe McKendrick

Joe McKendrick is RTInsights Industry Editor and industry analyst focusing on artificial intelligence, digital, cloud and Big Data topics. His work also appears in Forbes an Harvard Business Review. Over the last three years, he served as co-chair for the AI Summit in New York, as well as on the organizing committee for IEEE's International Conferences on Edge Computing. (full bio). Follow him on Twitter @joemckendrick.

Leave a Reply

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