Automation is moving to the forefront of many discussions in the IT/BPO space. Companies on all continents are experimenting, investing, and working with new, high-tech solutions to the industry’s age-old problem: providing the best service possible.
Abdul Razack, SVP and head of platforms, big data and analytics, at Infosys, recently sat down with Nearshore Americas to discuss what his company has been doing in the area of automation and artificial intelligence.
With the company starting to see progress, he believes the major breakthroughs in scalable productivity are just over the horizon. But he is steadfast in his belief that new technology is a job creator — not a job destroyer — and that automation, at its best, should amplify and never replace the human element.
Nearshore Americas: What is Infosys doing in the arena of automation and artificial intelligence
Abdul Razack: The objective of this platform division that I run is two-fold. Objective number one surrounds the advances that are happening outside in the world today. You look and see things like driverless cars and things of that nature — it’s an extreme form of automation. But there are a lot of enterprise-wide jobs that this technology can be applied to. So objective number one is to identify new areas that you can apply this technology to.
Objective number two is to take existing things that we do and make them better. Make them more efficient. Make them perform more. The idea is to be able to use technology and help manage large IT landscapes. A lot this can be automated, and that’s what we’re doing.
We also have AI-based techniques and technologies that can observe the resolution to learn that and look at it from a self-healing perspective. So the next time the same pattern happens or the next time the same issue occurs, the system has learned from it. That is the way we are applying it to the Infosys service line, thereby gaining more efficiency.
Nearshore Americas: So while the solutions to common problems are often the same, the inputs — or the ways the issues are presented — vary widely? And the idea is to get the system to interpret the inputs better in order to reach the same solution?
Abdul Razack: Yes — and to contextualize that. So, for example, if you are in the financial services industry, the system can recommend the next, best action to the person on call so that the experience is better.
So the platform has the ability to do three things together: data processing, automation, and adding intelligence to the resolution. But contextualizing that to different industries is the key. Then for the experience to the end user — to the human using that technology — the value is derived there.
Nearshore Americas: So now are we at a point with AI where some things can be fully automated while others still require a person to interpret it?
Abdul Razack: Absolutely. And that will continue to exist. The closest analogy I can relate it to is the automotive industry. In the 1980s, in a lot of cases, the assembly and the manufacturing was manual. Today, 85% to 90% of that is automated — but there is still human intervention that happens in making sense of everything and making decisions. And then there is a system that learns from that.
So the idea is to automate the repeatable and mundane tasks and have the humans perform more intelligent and more value-driven work. The idea is not to replace the human — the idea is to amplify the human to perform that task better. That will always be there. I don’t see that going away.
Nearshore Americas: How would you categorize where Infosys is in this evolution compared to the rest of the industry?
Abdul Razack: We feel that we are ahead. I wouldn’t say we are far ahead, but I think we are ahead. We have put together these things, and from a thought leadership perspective we are out there.
In some cases, people will argue that it cannibalizes our business. But I believe that it’s a natural thing to do. If there are repeatable tasks, this will be eventually used — For accuracy and because, if you are 100% of the time doing mundane tasks every day that don’t use a lot of brain power, if I am that person, I’d be pulling my hair out. So it’s natural that these things will be automated — and should be automated — and we are facing it head on.
Nearshore Americas: I have heard that Infosys expected to be a little further along by now as far as your actual, measurable productivity gains. Can you explain the reasons behind that? Why maybe have you been a little bit slower to see those gains than you maybe originally expected?
Abdul Razack: Some of the things are complex. We have applied it into our own practice — the Mana services practice — and we have released about 3,000 engineers who do these repeatable tasks. Now those 3,000 engineers are doing more value-added tasks — and other tasks. You free them up to do additional projects. So we see that.
It’s only part of the company. You can’t apply it to the entire company. You can only apply it to the parts of the company where you can see these patterns. But now with the Mana platform we are applying it to a larger part or the company — which is the application and development management — where there is additional scope for automation.
So where we have applied it to is a small part of the company, which is where you see the productivity improvement. That unit is about 15,000 strong and we have released about 3,000 people from there, so that is significant from a percentage standpoint. But if you apply it to the entire 200,000-person organization, from a percentage standpoint it fall short. Plus the nature of tasks of tasks is complex so we still need to iron out those patterns and make that more hardened, which is what the Mana platform is trying to do.
Nearshore Americas: So you just want to start small and get a more refined process to iron out those wrinkles before you scale it out to the rest of the company?
Abdul Razack: Exactly. That’s how we’re looking at the platform we just launched. Initially we worked with five customers to understand where we can apply it to and now we’re going to expand that to about 25 and then scale it out through the entire world.
There is a sense of momentum that we shouldn’t lose. But then if it falls, it’s hard to pick that back up. So we are treading that a little cautiously to make sure that the platform can handle the kind of patterns and the kind of use cases that are out in the enterprise. In a lot of cases, we have experience managing these landscapes. But we want to be doubly sure that that is the case before we scale it out.