When I hear about failures that happen in RPA projects or automation projects, it’s because the technology that was chosen to automate the very broad surface area was not good across all the service areas.
“We want to take the robot out of the human,†Param Kahlon, chief product officer of UiPath, tells Toolbox’s Neha Pradhan in a video interaction. Speaking about how UiPath became a leader in the global robotic process automation market, Kahlon talks about ‘attended automation’ where there is a need for humans to be involved, to guide and lead automation tasks.
In this edition of Tech Talk, Kahlon shares how enterprises can avoid automation fatigue and what impact business leaders should expect when there’s a delay in delivering RPA projects. He also puts forth recommendations to enterprises to satisfy automation maintenance needs. Kahlon further adds how companies who are looking at implementing RPA in the next 2 years can build teams who are attuned to RPA implementation.
Key Takeaways on How To Improve the Cost Savings and Avoid Automation Failures:
-
- Understand whether the automation is optimal and meeting expectations.
- Choose a platform with built-in resilience.
- Leverage automation to deliver faster digital transformation.
You can watch the complete conversation with Param Kahlon, CPO, UiPath in this video:
Here are the edited excerpts from our interview with Param Kahlon:
1. Can you describe how UiPath built a go-to-market product and a market strategy on the road to becoming Romania’s first $1 billion unicorn and subsequently a leader in the global RPA market?
Our journey started a long time ago. We focused initially on computer vision, and a core technology that helped our customers understand what is on the user interface. And just like human can understand it, our robots could understand it. And the ability to interact with those, UI elements that appeared on the user interface. Our business in March 2018 was primarily focused essentially on our ability to automate simple tasks using our ability to understand and interact with the user interface.
What we’ve done since then is we’ve expanded the scope of the platform from a product in robotic process automation, product focus company to an end-to-end automation platform. By an automation platform we mean that we can not only understand aspects of user interface, but we can also understand what is on a document, just like a human can. Also, ability to extract information, classify a document, a certain document type, and then extract information from it so that it can be posted to the line of business systems.
One of the key things we we’ve focused since March 2018, is ‘attended automation’ or ‘human and robot orchestration’ to get more work done within the enterprise.
Attended automation gives the ability to create engaging experiences that make sure that robots can automate what they can automate. But then a lot of times, humans need to take over and guide the automation or help complete certain tasks. And then robots can take over and do more with them.
So, we’ve created experiences that involve humans in the business of automation to make sure we can do more automations. So, with all these capabilities, what we’ve now created is an end-to-end automation platform that can look anonymous. Simple mundane tasks, but more complex automations can be done by incorporating AI into the platform.
2. What steps can be taken to overcome automation fatigue that could inhibit scale and impact ROI?
Yeah, absolutely! I think one of the things that I always tell customers is that you should not rush into automating any and every process within the enterprise.
The key thing is to first understand whether the process is optimal. Is this the way you want business to execute? If the process is ineffective and not optimal and you automate it, you only exacerbate or make the problem worse. This is because you’re automating a process that isn’t delivering on the results and the metrics that you want to achieve within the enterprise.
One of the key things we’ve tried to incorporate through our process mining product, which essentially can look at a process and understand what variances exist in the execution of the process. Are there bottlenecks that exist in the process? Is there a better way of doing certain things and then analyze as opposed to trying to automate?
Incorporating the ability to measure the impact that automation brings looking at ROI. This is not from the operational metrics of how many robots or what the utilization of robots is, but from a business metrics. Like, how have we improved the customer experience in automating the process? How have we reduced the lead times for processing certain documents and transactions within the enterprise?
Also read: Robotic Process Automation and Cloud Technology—Challenges and Opportunities
3. Can you tell us the business impact of emerging machine learning models on enterprises who have relied on manual processes?
Absolutely, great questions. Let’s look at machine learning first, what impact machine learning can have in the automation initiatives companies have I’ll walk you through sort of a use case. If you’re applying for a mortgage on a home that process is extremely manual, and it takes a company an average of about 4 to 8 weeks to process a home mortgage.
In loan originations process, majority of the time is spent on being able to look at documents, you have sent, extract information from it, put it into a system and then click a button so they can move it to the next step. You lose sometimes days especially in these economic circumstances, when so many people are applying for mortgages. There’s a huge backlog of mortgage applications that people have to process through.
Now, when you apply automation, you’re not only able to make those decisions and click those buttons, you can also automate some of the human capacity constraints while processing applications. In addition, you’re also able to apply machine learning to read documents, just like a human can.
Today’s AI and machine learning technology can basically say that the underwriting decision is based on the database of what kind of loan applications were either approved or denied in the past. Predictive decisions are based on past patterns of data that exists in the enterprise. You can accelerate the 4 to 6 or 4 to 8 weeks loan origination cycle, maybe down to a few days. You can be more responsive to customers and that’s the impact of AI and machine learning.
4. PwC claims and case studies show that conducting an RPA pilot project often takes 4-6 months instead of the expected 4-6 weeks. So, what should business leaders do to quicken the pace of their pilot RPA projects?
Yeah, this is important. Like most large scale projects, there is a lot of upfront work that companies should plan to understand how work is getting done in the enterprise. That’s not easy, because the process expert will give you one perspective of here’s how we do things. And then you’ll go and talk to the people that are doing it. And they may give you a completely different perspective on how work is getting done. I think this discovery process of really finding out how work is happening in the enterprise is not completely considered by somebody that’s taken on the RPF project.
That is why, we introduced products in our automation platform that observe things happening in the system at a large scale. So, using event logs, we can follow processes, we can follow documents and we can understand how a typical purchase requisition is flowing through the system, how an order is going through the end-to-end fulfillment process.
We’ve also introduced capabilities for task mining. Sometimes the system logs don’t tell you the whole story because you are maybe using applications where even logs aren’t generated.
If you looked at what people are doing on their computer screens, we can record the interactions that humans are having with their computer and then create that observation over a period of time. So, you can decide how you automate before you automate the end-to-end process.
Also read: Robotic Process Automation: How It will Help Small Businesses to Excel?
5. Fewer than 40% of RPA projects have not lived up to the expectations in terms of implementation time, cost, and analytics. Why implementation and tech readiness challenges occur and how business leaders can overcome them?
Banking and financial services are our prime use cases for automation projects because there’s a large volume of repetitive transactions that happen in that environment. However, the technologies used by banking and financial services customers are very customized. You will see that the technology that’s generally used in building those applications is also diverse. You will see applications built in the main frame. A business process generally is going through all these different systems to be able to tie everything together.
When I hear about failures that happen in RPA projects or automation projects, it’s because the technology that was chosen to automate the very broad surface area was not good across all the service areas. There are a lot of automation platforms that work well on web-based applications and do an okay job, but they’re not suited to do mainframe-based automations. They’re not suited to do automations that are maybe running on a virtual desktop. They need computer vision capabilities to automate certain contexts.
A lot of times, you know, customers automate things that can be automated but are not 100% autonomous. As in, they’re trying to take the human completely out of the loop. Autonomous also leads to a lot of issues with automations that people are trying to create.
Also read: RPA: A Must-Have for Digital Transformation
6. What are your recommendations to enterprises to satisfy RPA maintenance needs when they do not dedicate enough resources and management attention?
I think my recommendation is choose a platform that has resilience built into it. Look for a platform that can make sure that when you’re automating something, it is going to be responsive to change.
A lot of times companies choose platforms that are available for like test automation and those don’t have resilience as a design consideration. So, you want to choose a platform where resilience, the ability to react and respond to changes is sort of a sign, when building the software itself. The other thing you want to introduce is the concept of testing.
Change management should be tested within the enterprise. It should not be that you’re going to wait for things to fail. That should be a continual process. So, make sure to create an associated plan and script around it before continually testing it, similar to what we do in the application system.
7. According to WSJ, Amelia and other ‘white-collar robots’ are changing the humans’ cognitive monopoly. So, should organizations continue to focus on AI augmenting human capabilities or will RPA unseat human intelligence?
Yeah, good question. What we have seen is that whether it’s AI or whether it’s RPA, the intelligence we apply is very narrow through technology. The human cognitive monopoly on broader sort of scenarios and the ability to incorporate a lot of data sets to make decisions and react isn’t going away. So, our premise has always been that automation is successful. We’re seeing more and more use cases there where basically robots and humans are working together.
There are aspects of human job, which are very rote, mundane, repetitive that humans don’t enjoy in fact. I think we will look back at some of the things that we make people do today 10- 20 years from now, they’ll say it was completely inhuman to make people do those kinds of things.
We’ve always said that automation is augmentation of human capabilities within the enterprise so that people can do more. People can focus on what they enjoy, what they’re good at, and what adds more value to the organization they can take away.
There is a tagline we use, which is, “we want to take the robot out of the human.†It means take the repetitive, mundane, non-value type of activities that we make people do and let robots do that so humans can focus on creativity.
8. What advice do you have for enterprises who are looking at implementing RPA in the next 2 years, especially in terms of building automation teams from the ground up?
Yeah, it’s a great question. You should put automation at the center of digital transformation initiative within your organization.
You’ll look at your processes as you look at what you’re trying to achieve in terms of experiences you’re delivering to your customers. The question is are automations making the company more operationally efficient? Automation is a core pillar for enabling some of those initiatives within the enterprise.
And when you look at initiatives at that level, you will realize what automation is. It is not just an add on to things that you have done but becomes the center of transformation you’re driving within the organization. Automation helps to get there faster and quicker, and it does that in a very agile way.
We believe automation will add tremendous value to companies who are looking at digital transformation. By leveraging automation they will get there faster and in a more agile and efficient way.
About Param KahlonOpens a new window
Param Kahlon is the chief product officer at UiPath. In his role at UiPath, he is responsible for product strategy and roadmap. Prior to UiPath, Kahlon was a general manager at Microsoft, responsible for Dynamics 365 customer engagement apps (sales, customer service, and field service). Prior to Microsoft, Kahlon managed customer service and order management products for SAP® and Siebel/Oracle.
About UiPathOpens a new window
UiPath offers an end-to-end platform for automation, combining the leading Robotic Process Automation (RPA) solution with a full suite of capabilities and techs like AI, Process Mining, and Cloud to enable every organization to rapidly scale digital business operations. The company believes in using the transformative power of automation to liberate the boundless potential of people. A world with a robot for every person is their goal. Based in New York City, the UiPath’s presence extends to over 40 offices around the world.
About Tech Talk
Tech Talk is an interview series that features notable CTOs and senior technology executives from around the world. Join us as we talk to these technology and IT leaders who share their insights and research on data, analytics, and emerging technologies. If you are a tech expert and wish to share your thoughts, write to [email protected]Opens a new window .
Why automation projects fail and how companies can save costs and get better returns? Share your thoughts with us on LinkedInOpens a new window , FacebookOpens a new window , and TwitterOpens a new window .