IBM’s Rometty Describes Enterprise-Driven Digital Reinvention

IBM THINK — If chapter one in the digital transformation – aka digital reinvention – journey is characterized by digital, AI, experimentation and customer-facing apps, chapter two will be enterprise-driven, meaning scaling digital and AI, including hybrid cloud and mission-critical applications — all underpinned by responsible stewardship.

That’s according to IBM’s chairman, president and CEO, Ginni Rometty, who kicked off IBM Think 2018 this week in San Francisco

To help conference attendees digest exactly what the digital-reinvention journey looks like, Rometty got a little bit of help from her friends, namely, premier IBM customers – such as AT&T Communications, Geico, Hyundai Card, and Kaiser Permanente – companies on the leading edge of digital reinvention, and way ahead the pack of mere mortal businesses. That said, IBM has already racked up about 20,000 AI engagements with customers around customer service, HR, predictable energy, and so on.

IBM’s Ginni Rometty

“Many of you have deployed new capabilities, new apps or new ways of reaching your customers, described by many of you as random acts of digital,” said Rometty in her keynote address, titled Building Smarter Businesses. “And, by the way, it’s not going to let up.”

With that, Rometty shared five lessons learned when it comes to what it will take to scale digital and AI. The first two lessons have to do with the approach taken — outside in or inside out. The outside-in approach means offering services around the customer experience. Inside out, that’s about workflow, the idea of data driving change, which requires modernization of core applications, resulting in a flexible architecture.

The third lesson is that you’re going to need a platform to put things together.

“It will be fueled by data and have AI infused in its workflows, and people will feel like technology really empowers them,” said Rometty.

The fourth lesson learned is the need for an AI platform to keep track of its life cycle. The fifth lesson is that you’ll never have AI without an information architecture.

IBM made some product announcements this week to support chapter two, including Watson Anywhere, and IBM Automation with Watson. Watson Anywhere makes Watson portable across any cloud – public or private – or on-premises. Rometty describes IBM Business Automation with Watson as workflow with moments of intelligence baked in, or digital and AI with the guardrails of business rules.

Before sitting down for individual fireside chats with top executives from AT&T, Geico, Hyundai Card, and Kaiser Permanente, Rometty shed some light on where IBM is headed, and that’s toward chapter two, or scaling digital and AI.

While still in research – with products not too far away – IBM is working in three areas: core AI, trusted AI, and scalable AI. Core AI is about how to get AI to learn with less data. Trusted AI is about developing and applying tools to wire AI systems for trust. Scalable AI is AI to automate AI.

To get a sense of what a digital-reinvention journey looks like, there’s Geico, founded in 1936 and one of the largest auto insurers in the U.S — also, one of the first insurance companies to tackle a major digital transformation.

Greg Kalinsky, executive vice president and CIO at Geico, pioneered the company’s journey with Watson. He spoke about …

… digital and AI.

Decades ago, automation eliminated a lot of manual processes at Geico. After that came mobile.

“Then we realized that we had to be more elegant, but we were frustrated because we couldn’t get there trying to leverage what is really a monolithic sales and service application,” he said.

Note that Geico has over 8 million customers, one-quarter using self-service for their policies on a mobile application.

Given, the mobile application was easy to use, but Geico wanted the app to be more intuitive and dynamic to reflect what the customer needed.

While taking big risks might reap big rewards, it’s not easy being a pioneer.

“I was nervous that people weren’t ready for AI and was worried that in our zeal to find a more dynamic customer experience, we didn’t want to alienate people and steer customers away,” said Kalinsky.

The team doing the AI development work was measured, literally, reviewing early customer interactions, or the verbatim transcripts.

“We wanted to see how customers were interacting in a conversation with Watson — and they didn’t know it was Watson,” he said.

The proof-of-concept phase proved to be successful. So what next?

Given the amount of time and effort that went into this process, the next step was the value-add. Or, as Kalinsky described, not only do no harm but drive more sales.

The first attempt that focused on auto-insurance policies didn’t spike sales, even though it did no harm. When the same concept was rolled out for renter’s insurance policies, which when compared to an auto policy was less cumbersome from the get-go, a reconfigured application did what the company hoped it would do.

“There was a 40 percent increase in closure rate on those policies — a tremendous success,” said Kalinsky.

The development team took another shot at the auto-insurance policy with a new vision rather than trying to replicate an old structured application.

“What we realized was holding us back was, in order to take advantage of the dynamism that Watson presented and also in recognition of the fact that while Watson is a slam dunk when it’s an expert-facing application, our customer and the average consumer wouldn’t know the next right question to ask,” he explained.

Currently in process, the Geico development team is taking what it knows about the customer in the form of data, breaking the application into microservices, and recompiling the application at different points in time — allowing Watson to lead the customer in a way that’s meaningful to them, individually.

Rometty wanted Kalinsky to share his story for a reason, one that ties back to the five lessons learned.

“To get the value of AI you can’t just layer it on your old way of working. You might get some benefit and do no harm — but that was outside-in, then you found yourself going inside out, then you had to break it into microservices because you wanted all that flexibility to go back in,” said Rometty.

To close, Kalinsky shared some transformation leadership lessons.

“The challenge of management and leadership is that …

… the opportunity may appear to you to be obvious and the right course, but folks don’t like change. You really have to work together to demonstrate how great the potential [is] to distinguish yourself by your customer experience, customer service or price, and how you can help your enterprise leverage technology to change the way you do things,” he said.

For Geico, one technology change example was to move entirely into the cloud. The company partnered with IBM to move all its disparate data stores into a private cloud — which entailed overcoming a lot of resistance.

“We had to figure out how to transition and transform ourselves from our traditional hardened perimeter to a defense in depth, and we had to do that while we were still on-prem in such a way that it wasn’t threatening,” said Kalinsky.

When it was complete – if security is ever complete – the defense-in-depth approach removed the last barrier to resistance from the individuals at the business who were worried.

Today, Geico is on the public cloud and IBM’s private cloud and is getting ready to move its mission-critical claims systems to the IBM Cloud.

“That’s act two, chapter two, mission-critical,” said Rometty.