How to become an artificial Intelligence Product Manager from an entrepreneurial experience

Entrepreneurial Edge
21 followers
Nov 21, 2024
I'd like to look at your earlier years in the startup and one of the things it's interesting
to cover is this trend in consumers with images and so on. Can you tell us a little bit more?
Sure, so what we really had to do is we had to look at what people were doing with text,
right? So there's a whole business built around natural language processing and text analytics
and social media where they're trying to identify trends and topics of conversation,
understand sentiments around a particular campaign, tracking things like impressions
and engagement. So for example, one use case might be tracking the value of a sponsorship,
right? So you know at a sporting event how many people are in attendance and how many people are
watching on TV, but there are all these people posting images of the game where there might be
sponsorship and logos on, you know, let's say a basketball, a backboard, the floor and so there's
all this additional impression and engagement that's happening visually, but it's not being
captured via text. So let's say AT&T sponsors the NCAA Final Four, knowing that the game is
really talking about AT&T typically, but they're taking lots of pictures of heavily branded assets
there and because you have the ability with computer vision and AI to track that, we can
understand what was kind of the return on investment that they got from a social perspective
on that investment. So that would be one use case. And so that just for our viewers to understand,
so basically the customer here is a media agency that is telling to their customers,
here are our results and our results are more, if you like, extensive that you have
believed until today because now we can measure also the visual engagement of people.
Absolutely. So yeah, typically agencies and brands that were purchasing this technology. This may
not be something a lot of people want to hear, but I'm beginning to see that AI is becoming,
at least in some respects, sort of a commodity. It's becoming very, very easy to build.
You know, we have access to data, cheap GPUs, you can, you know, leverage neural networks from
Microsoft, Facebook, you know, almost completely for free. And so specifically as it related to
doing things like logo detection and object detection, recognition, and all those different
areas where a lot of the value now becomes is around the workflow and the outcomes and
basically integrating into what people are doing in their everyday lives. I think that just
delivering a technology is not one of the, it's not how you can extract true value, especially
as we see more people becoming adept at building these technologies. I think early on when I started
there weren't a lot of people that could do it. But I think with things like TensorFlow and
Microsoft for sure, which makes it very, very easy and a low barrier entry for people to go
and build technologies like these without, you know, having to be some sort of AI expert.
And so what's really going to be important is really understanding the business application
and integrating with an individual's everyday workflow, understanding what they do day-to-day,
and then trying to apply that technology to what they're doing day-to-day. So Crimson Hexagon,
when I was there, they were very well integrated into the workflow of brand analysts and media
strategists. So they understood the things that they wanted to track, they understood what the
outcomes were that they wanted to look at, the measures. And so it was figuring out how to
with that to figure out what the ultimate value is going to be,
because that's what people will pay for. They won't pay for a technology alone.
