Intelligent automation technology is maturing to the point where the discussion is no longer about automating discreet tasks, but full-on end-to-end processes including unstructured content. As with earlier automation efforts, however, it’s key to pick the right projects – and that is no longer just the proverbial “low-hanging fruit.”
These were a couple of key insights from Munish Arora, Associate Director of Advanced Analytics at the large global financial services and life insurance company Sun Life, during the second installment of Unstructured Unlocked, the Indico Data podcast focused on intelligent automation for unstructured data.
Intelligent automation technology is maturing
In a previous post, I detailed the multi-faceted approach Arora took to automating processes in the his contact center, including handling email and phone requests. Another topic we discussed was the how the intelligent automation industry has matured in the last few years.
“Organizations have matured at least by five or 10 points in last five years or so,” Arora said. “At that point in time people were focused primarily on automating specific tasks.”
Automating tasks typically involved structured data, which he said was the low-hanging fruit. Today he sees companies focusing more on end-to-end process automation, which requires artificial intelligence technology.
“Now I see more AI and [addressing] problems like unstructured data,” he said. “These are the things we could not solve in prior years.”
Listen to the full podcast here: Unstructured Unlocked episode 2 with Munish Arora
Gradual automation rollouts are key to success
Automating processes involving unstructured data doesn’t happen all at once, and companies need to be methodical in how they roll out intelligent automation technology. “It takes time for people to adapt to change,” as Arora said.
His company uses an agile development process, which helps introduce intelligent document processing technology gradually to a select group of users, maybe 10 out of the 100 who will ultimately use it. The company routinely builds prototypes and shows the select business users new features and functions every week or two. The approach reduces the surprise factor because at least some users have seen and tested each feature before Arora’s team rolls out the finished product.
“Those 10 users who have tested your product will become your brand ambassadors, to convince your [other] 90 users that they need to use it,” he said.
Automation project selection: “Fail fast”
Arora also had some sound advice on how to go about selecting the best use cases for intelligent automation projects.
One major factor is finding projects that will have the most impact on the client experience. That may mean first addressing processes having to do with voice calls from clients, where a live customer is on the line – such as the attended bots he discussed in this previous post.
Another consideration is process to process maturity, when solving a particular issue will enable a process to be automated end-to-end; here again, the ultimate goal is to improve the client experience.
“Third, I may want to look at my profitability versus my cost in building the project. What will be my payback period?” Arora said. “I don’t really prefer to pick up a three-year or five-year project with the payback of another three years. You want to have a process which is cheaper and faster to build, and can give ROI in a shorter time frame.”
I was tempted to go on a rant about how we can’t always expect payback in a single fiscal year from intelligent document processing projects. I’m glad I didn’t because Arora later clarified that’s not what he expects, at least not from more mature organizations that have already addressed low-hanging fruit projects involving structured content. When embarking on projects involving unstructured content, “It’s better to spend more time,” he said.
That said, the agile dev process also helps ensure you don’t embark on projects that have no chance of success.
“I believe we should fail fast and fail cheap,” he said. While no process automation solution will ever be 100% perfect, if it’s not going to work at all it’s better to find that out sooner rather than later. “Fail fast, learn from your mistakes, and then move to a different solution.”
For unstructured content, “The time has come”
Finally, I asked Arora where he sees unstructured content fitting in the current scheme of things. Is it the next big horizon for automation centers of excellence, or still early days?
“Businesses really want to have end-to-end [process] automations,” he said. “If you’re going to try to automate end-to-end processes, you’re going to get unstructured data, be it images, voice or some other form. If you really want to enhance your client experience, you need to solve that problem.
“The time has come,” he said. “I don’t see it as a futuristic thing. It’s very much in the present.”
Check out the full Unstructured Unlocked podcast on your favorite platform, including:
Read the full transcript of the podcast here.
