In part two of this interview with Rajeev Ronanki, Principal and Leader of Deloitte’s Cognitive Computing practice, FEI Daily discusses where the industry is headed with artificial intelligence and cognitive computing.
FEI Daily: How will financial executives be using artificial intelligence in the future?
Rajeev Ronanki: I think a lot of it is going to have to do with sensing and monetization.
Think of, for example, Blockbuster was looking at Netflix, and they essentially ignored that entire trend. As a result, they went out of business. Kodak, similarly with the digital trend back when in the '70s, they created the first digital camera. They didn't really do anything with it because it didn't really fit their core business model. They ignored it, and put it on a shelf, and moved on. Meanwhile the industry evolved rapidly around that concept, and their business model was obsolete. That was a few years ago. Now the technology disruptors are even more rapidly growing, and the capabilities are doubling roughly per Moore's Law and even more accelerated in some cases. In that pace of change, how do enterprises figure out what's out there, what might disrupt them, and have a point of view on whether they need to react to something, proactively go after and identify an opportunity or be satisfied that it's not a threat or an opportunity.
What we think is that with cognitive computing, big data, and all these things reaching maturity, that you can deploy a sensing capability that would essentially tailor and personalize opportunity sets for enterprises. Using that, you can then systematically decide how to position your capital debts, whether that's acquisitions, or organic capability development, or some combination of the two. I think that whole process could be re-imagined and re-implemented. It's still in the research stage, but that's where the industry is headed.
FEI Daily: Why is cognitive a business imperative?
Ronanki: I think there's one stream of thought that says almost every business out there is a technology-enabled business, so without that there is no scale, and you can't really be successful. If you take on that mindset, then cognitive really is a game-changing technology because up until now technology had to reflect the requirements that were developed by humans. You describe the rules, you codify the requirements, and you program those rules into the system. The systems then either automate, or create insights, or provide the apps, or the portals, or whatever the function is, but they don't necessarily adapt and change. The minute that a system goes live in production is reflective of everything that has happened up until that point. From that minute forward you still have to update it and maintain it.
Cognitive, for the first time ever, offers the ability to adapt, and change, and learn much like we do. It grows, and morphs, and changes with input. What if a computing system could do just that? Rather than being a static rendering of a point in time set of requirements, what if it could actually change, and evolve and learn, much like humans do, to simulate the way that we think? If you introduce this technology and it continually learns, it has the potential to augment human intelligence. That's why I think it's a true business imperative. If you ignore it, I think you ignore it at your own risk because the collective learnings of the enterprises remain with the people that are part of that enterprise and then just go away once the people at the enterprise move on. Whereas a system that captures that knowledge is able to grow and serve that enterprise differently than the way regular technologies can.
FEI Daily: What would you say are the best practices for cognitive adoption?
Ronanki: Focus on not just what cognitive can do, but rather what are the key challenges that an enterprise has. What are the areas of highest cost, highest inefficiency? What are the pain points around consumer engagement and customer satisfaction. Prioritize areas that have the highest value. Then once you have that, use cognitive technologies to demonstrate proof of value. The technology works. There's no question that it's been proven in multiple settings and different environments. A lot of our clients start with the wrong question, which is "Hey, can this technology work?" They do a proof of technology, a proof of concept, and they figure out it works. After that they're stuck because they don't know where to go with it.
We suggest starting with a list of use cases, a list of business problems, and a list of challenges. Then intersect that with "Could it be solved better through the use of cognitive technologies?" and systematically demonstrate the value through the application of that technology, and show that it works. Once you have that, scale it, so start small with proof of value demonstration, and then have a plan of scaling rapidly, and then put the two and two together to introduce it into the fabric of every element of the company in a way that's consistent with enterprise business prioritization rather than a technology-driven exercise.
Change management, the training, and the talent acquisition are equally important. We can't just acquire the technology and not have the data science, and the technology, and the talent wrapped around it. Enterprises do have to think about how to create a center of excellence where this can be seeded and then grown over time. There's a huge war for acquiring people with AI cognitive computing expertise, so Facebook, Google, IBM, Deloitte, all of us are in the market trying to go after the same talent. It's going to be challenging for enterprises and clients to acquire that talent. They have to think about other ways in which they can find and retain this talent. Doing those two things together, I think, would be probably two of the top best practices in terms of adoption.
FEI Daily: How can organizations adapt their culture to a more cognitive computing mindset?
Ronanki: The culture is around thinking in terms of results and outcomes. With every technology wave there are some that don’t believe the hype and others that believe it's the best thing that's ever happened. The truth is always somewhere in the middle. There's some hype and some reality, so having a realistic view of what the technology can do and really how to frame it in terms of results and outcomes is number one. Two, culture change needs to be business-led. It's a game changing technology. Having c-suite, CEOs, COOs, CFOs sponsorship is really key and then challenging the enterprise to think of results, and outcomes, and demonstrations of value.
Introduce culture change from the outside-in. One of the hardest things to do is to create this notion of a cognitive computing-based culture in the heart of enterprise. A lot of times that's going to fail because the data-based business is sometimes at odds with the timing and the realization of value, and it may be viewed as a distraction. What we've found to be a better way is to actually introduce it outside of the core and rapidly scale the edges in order to create that change rather than trying to introduce the change into the heart of the enterprise and fighting an uphill battle from day one.
Read Part 1 of FEI Daily's Q&A with Ronanki here.