This was the year automation technologies would “become the tip of the digital transformation spear,” according to a Forrester Research prediction.

Automation technologies refers to a diverse set of software from robotic process automation (RPA) to business process management (BPM) to artificial intelligence (AI) and beyond.

According to Forrester, they are becoming a key investment for CIOs and their enterprises. These technologies are also reshaping the jobs of human employees as the software is increasingly performing rote, repetitive tasks.

While organizations are making progress in their automation efforts, the vast majority are struggling to scale their RPA initiatives, according to Leslie Joseph, principal analyst at Forrester.

SEE: Digital transformation: A guide for CXOs (TechRepublic Premium)

This is due to “challenges around developing business cases, governance, organizational alignment and the lack of a cultural framework for managing automation,” Joseph said.

Intelligent process automation software provider Bizagi is looking to change that with the launch this week of a platform to natively allow organizations to combine AI/machine learning, digital and robotic automation in the cloud.

Bizagi said it is enabling digital and robotic processes from Blue Prism, UiPath and Automation Anywhere to natively take advantage of Microsoft Azure’s Cognitive Services.

“From an AI perspective, a critical new capability is that users can now simply upload a CSV file and select which attributes they want to predict,” said Gustavo Gómez, CEO of Bizagi. “Bizagi will then choose the most appropriate machine-learning algorithm. With this you can unlock the power of AI without the help of a data scientist.”

The timing of the release is good, observed Craig LeClair, a vice president and principal analyst at Forrester, because “the RPA market is very hot.”

There is a trend toward intelligent process automation platforms that “combine RPA, which does simple task automation without learning comprehension, with some of the practical AI components like text analytics that can rip through emails or with a chatbot being able to have a conversation with humans,” LeClair said.

A lot of major tech companies have announced IPA or automation platforms including Virtusa, Genpact and Tech Mahindra, as well as software vendors like Kofax and Hello Systems, said LeClair, who also authored the book, “Invisible Robots in the Quiet of the Night.”

He estimates that one-third of all tech services companies will be coming out with an automation portfolio that includes RPA and various AI building blocks platforms.

“What Bizagi is doing is positive, just not revolutionary,” LeClair said. “They’re an example of a digital process automation or BPM vendor that is embracing RPA in a positive way adding AI capabilities and putting that in the cloud.”

Automation hurdles

Currently, automation initiatives solely focused on RPA slow down because a bot can only handle simple processes with low variation, said LeClair. For example, it can handle an 18-minute reconciliation in accounts receivable that happens 1,100 times a month.

Forrester recently noted that “more than 50% of companies with RPA initiatives have less than 10 bots in production.”
But “to be more transformative and create new customer experiences…you need to add an analytics layer to add context and do more interesting things,” LeClair said, such as enabling the bot handling inbound emails to address complaints and other customer service issues in an efficient way.

RPA technology can only handle structured, tagged information and can’t, for example, go through 1,000 emails and extract words and then make judgments about which bucket that email should go into for resolution, he said.

Adding in analytics capabilities like natural language processing that can understand words and normalize information becomes a game changer, LeClair said.

Gómez agreed, saying that one of its clients, Deutsche Post DHL, automated its duty billing process so that employees can correctly bill the appropriate taxes when transporting goods in almost every country across the globe. “This is an end-to-end Bizagi process that combines human work with bots drawing on and inputting data in legacy systems,” after the bot takes data from the system and creates cases, he said.

Orchestration capabilities are critical for making this happen, LeClair added, because “you’re trying to combine these diverse technologies to align with a particular use case.”

Most of RPA is occurring in operations area like finance, accounting, and HR. Right now, the major inhibitor to scaling bots is not the technology–but coming up with the right processes, he said. LeClair advised businesses to follow Forrester’s “Rule of Five,” three simple rules all involving the number five to apply RPA.

Here are the three rules:

  • Five decisions made. RPA has limited rules capability. This is why combining RPA with a platform that includes BPM and machine learning, as in the Bizagi release, makes sense, he said.
  • Five applications accessed. Applications change and RPA bots manipulate applications directly, so the fewer involved the better.
  • Five hundred clicks, i.e. keystrokes, clicks, and mouse movements. Keeping the clicks below 500 is a good rule of thumb to keep tasks simple and repeatable.

Image: Getty Images/iStockphoto