An RPA investment by one of the world's largest immigration law firms is paying dividends on the client side and in the back office.
Vince DiMascio, the chief information and technology officer at Berry Appleman & Leiden, describes how robotic process automation deployments are providing round-the-clock service to clients, which include some of the world's largest tech companies. RPA software is also being used as the glue for systems integration work that hitherto was too expensive to tackle. He makes a convincing case why RPA trumps APIs as his system integration tool of choice.
This is the second installment of a video shot at the recent MIT Sloan CIO Symposium. Check out the first installment, "Law firm CIO successfully deploys RPA and AI to improve client service," to hear DiMascio describe the many types of AI the firm is using and some of the challenges he's encountered along the way.
Editor's note: This transcript has been edited for clarity.
How are you implementing RPA?
Vince DiMascio: With RPA we've deployed it in this way, where you start with the core capability you get out of the box, which is [for] deterministic processes. If you know there's a defined path, then you can dispatch an automated system to follow that path. When the outcome becomes probabilistic, it's then that you want to include human subjectivity. [That is done] either in the form of a model that has been trained and approved by humans or through introducing human review along the process chain.
We do both. We have some systems where the automation will take steps on its own, based on tried-and-true models that we've developed. We have others where we suggest and draft and propose what the next steps should be and then put that forward for human review in a manner that is much more efficient than a human doing what's required to develop such a draft.
What are some tangible benefits you're getting from your RPA investment and from machine learning?
DiMascio: We're already seeing material value from our RPA investment. It's not so much about revenue growth as it is about revenue growth as a byproduct of exceptional client service that comes out of this. We have sophisticated technology clients. Many of our clients are the big tech companies -- and they want a user experience that's phenomenal. And that means self-service, that means on-demand. If you have a robot available to do work, whether it's reviewing material that a foreign national would submit in connection with a visa application or looking up statuses of applications in progress, that happens in real time for us -- and that is something that is just not a sustainable model with human staff.
So, we're doing things that we couldn't do otherwise do, with robots, to deliver client service. Quality enhancements is another obvious [benefit]. Automated systems don't transpose digits when they're typing things into other systems. Speed is another [benefit]. The robot moves much more quickly than any human finger can move. Those are the value-adds we're getting from the RPA investment.
With machine learning, it's similar capabilities [that enhance] the user experience: For example, you can upload a passport, and the system will say, 'This looks like a passport. Here's the information. Is this right?' without having to ask a person to enter information. Likewise with other document types. That's a user experience that sets us apart, and that our clients love.
Also, on the back-office side -- HR, IT and facilities and those operations -- we have systems specific to each of those business units that should be integrated. But the cost of integrating them is really high and therefore doesn't happen. With RPA we can make that integration happen without having to spend the dollars it takes to enable an API and write the code to connect those systems. That's one example of a big lift you get from RPA acting as the glue between systems.
Likewise, the traditional processes that you hear about with RPA, we've made manifest in our organization -- employee onboarding, [turning off access to systems to comply with our policies around data privacy]. There's an operational efficiency angle and a risk management angle that are very important to us as a law firm.
Do you see RPA as a temporary or permanent solution for system integration?
DiMascio: It's a solution that's going to hold for us. The rate of change of API-based systems is not much slower than the rate of change of UI-based systems. A lot of the UI-based systems, including the systems we've built and maintain, are single page apps with RESTful endpoints, and therefore the UI may not even matter even if you're using HTTP to connect and do the work. We're finding that the scary risk [with RPA] that people have been talking a lot about -- the UI changes and everything breaks -- is not as scary as some make it out to be.
Are you also using bots in IT systems?
DiMascio: The deployment pipeline, our security operations center -- in IT specifically, we're looking at using bots in both of those places. For example, if a server goes down and an email notification occurs, then our bot will cordon off that system and do the diagnostics that an entry-level or associate-level security operations center employee might do.
The report will always be complete, [the bots] will fingerprint the system, they'll identify what symptoms they've seen. And then they'll lock down the network and forward the report to a human, who's then armed with reliable information that's already been very recently and accurately collected by the bot, so that person can take the next steps that require the subjective human element. That's big for us.