Sramana Mitra: In your estimate, how are we in getting to that kind of usability of being able to use AI in a highly-leveraged way without having rocket science capabilities in the individuals?
Mike Flannagan: I think we’re doing a great job on the consumer side. Amazing work has been done to provide those sorts of technologies and recommendation engines to consumers. The lagging factor is business applications. That’s where SAP is focusing on.
How do we begin to integrate those capabilities in every one of our business applications so advanced analytics becomes something that not just 10% but 90% of your workforce can take advantage of. I would challenge other business applications providers to think about whether they will be able to maintain relevance if they don’t incorporate those technologies into their business applications in the process.
Sramana Mitra: How conversant or familiar in the fray are you with startups who are working on business applications with AI-based engines?
Mike Flannagan: I spend probably four or five hours a week listening to pitches from young startups in the area of machine learning and analytics. I’m certainly not as conversant as someone who’s a venture capitalist.
Sramana Mitra: Based on that window into the startup world, can you name one or two that are particularly exciting? If so, which ones and why?
Mike Flannagan: Given the nature of my role in SAP, I would prefer to talk about individual companies. If it’s okay with you, what I would be very comfortable doing is categories. The ones that are most interesting to me are the ones who have given any pitch on horizontal, broad-based machine learning and take in a niche that they understand very well.
There’s one company in particular that is evaluating potential employees by gamifying the process of what we would have historically done with something like a Myers-Briggs type indicator. The delta between what they’re able to tell you about a potential candidate just by having that person play a game like Angry Birds and what you find out by having them take the Myers-Briggs is so narrow that it’s negligible.
That’s an incredibly impressive application of a whole lot of different technologies in the background. That kind of thing where I can get someone to play a game for five minutes and give you the kind of insight into their personality that the Myers-Briggs would is very interesting.
Sramana Mitra: That’s fascinating. We agree that domain knowledge is one area to look at. Which domain fascinates you? Which domain has less activity but high opportunity for entrepreneurs to dig into?
Mike Flannagan: I think natural language is an area where there’s so much more potential than what’s been tapped.
Sramana Mitra: But natural language is still horizontal, right?
Mike Flannagan: It is. I think that’s part of the reason why there’s so much untapped potential. Natural language is so horizontally-oriented today. If you look at natural language applications that have been developed that are particularly good in their space, they have taken a more vertical approach.
Sramana Mitra: It’s a lot easier to train natural language within a vertical than in a broad way. Absolutely.
Mike Flannagan: If you look at the 10 or 12 startups in that space, most of them are trying to be very horizontal as opposed to picking a vertical and really being amazing in that vertical area. If I was picking one space today, I’d pick a space where none of the existing horizontal payers have great use cases in natural language and go as deep as you can.
Sramana Mitra: Great answer. Very precise and exactly what I was looking for. Thank you for participating.