In each episode of The GameDay Playbook presented by FanFood, Rob Cressy discusses how leaders are transforming the sports and live entertainment industry by leveraging technology to enhance the fan experience and operate gameday more efficiently.
Steve Wasick, President at InfoSentience, joins Rob Cressy to talk about using technology to create content. How is the fantasy sports industry using AI technology to mass produce content and deliver value for fans at the same time? What can be done to ensure that AI technology has a brand voice and doesn’t sound like a robot? Will AI technology replace writers when it comes to content creation? From a forward thinking perspective, what is the next thing for AI technology generated content?
Listen to the Gameday Playbook on:
Rob Cressy: (00:04)
Welcome to The Gameday Playbook presented by FanFood, a discussion around how leaders are transforming the sports and live entertainment industry by leveraging technology to enhance the fan experience and operate game day more efficiently. I'm your host Rob Cressy. And joining me today is Steve Wasick, president of InfoSentience. Steve, great to have you on the show.
Steve Wasick: (00:30)
Hey, thanks for having me, Rob.
Rob Cressy: (00:32)
Can you give a quick overview on who you are and what you do?
Steve Wasick: (00:36)
Sure. So I started this company InfoSentience about eight years ago and what we've done is we've created software that can analyze a data set, figure out what's interesting about that data set and then write up or basically communicate that to people using natural language. So our first product that we ever did was just doing fantasy reports for fantasy sports players on CBS, both for football and baseball. I originally just sort of got this idea because I was in a fantasy football league and I was pretty busy and I was, you know, sort of monitoring my own team, but I didn't really know what was going on with the rest of the teams in the league and I thought, man, wouldn't it be cool if there was like sort of the equivalent to like a sports side front page that told me about kind of like what were the interesting things that happened, close games, bad coaching decisions, that type of stuff. And so I thought of this idea and I rolled it out and CBS picked it up. But once I had it I was kinda like, Oh, actually this is of really valuable in lots of other places too. It's not just sports, just having this sort of ability to take data and figure out what's important and talk about it. It's actually kind of pretty cool. So that's what we've done and kind of rolled out into some other, other industries too.
Rob Cressy: (02:05)
So how do you or it know what is interesting or valuable?
Steve Wasick: (02:13)
Basically we just kind of like teach it. We have a system set up where we can say like, okay, you know, we come up in our own head like, well, what would we talk about if we were writing about this league, you know, or we were writing about this game. Then we basically just teach the system to understand that those relationships within the data are important and then, you know, teach it basically how to write, give it the tools that it needs to write, you know, any individuals sort of story or sub-component or a paragraph or any of these things. We have to sort of give it the tools to go to work and then it can go to work and create the finished product.
Rob Cressy: (02:51)
One of the biggest challenges I've seen with chat bots or AI type content is a lack of brand voice. So you and I, we may use slang and jargon and certainly if we're looking at the fantasy football world, we're so into it where it's like if Patrick Mahomes has an amazing day, how can the content that's being created be Patrick Mahomes was straight fire today and blown up for five touchdowns as opposed to Patrick Mahomes threw five touchdowns because that's what really separates one brand from another. Because a lot of brands in regular marketing, not even with this automated content world, they're so vanilla, but the ones that have personality are the ones that stand out.
Steve Wasick: (03:43)
Yeah. I mean we work really hard to give that sort of vibe to the content. Because again, it's just about teaching the system. So you could just put in, you know sort of say like, hey, look, if somebody throws five touchdowns, you got to treat that different than if somebody throws three and we have this, you know, throughout the system with football and baseball, it's like players having nicknames. It's like we can put those nicknames into the system. Now it gets tricky because you can't just refer to somebody by their nickname 10 times in a row. It gets repetitive. So you have to have rules in place for how you deal with that. So it's not just as simple as just plugging something in and letting it go to work. But, you can get that type of voice. And I think we've successfully done that. If you look at the content that we, that we write for CBS and fantasy or if you even look what we do, and we do a ton of live sports for CVS as well. So we're covering NFL football, NBA basketball, soccer, that like just tons of stuff. Yeah, if you have that sort of just super dry robotic, like X, Y, Z happened, then a, B, C you know, it's pretty obvious to a reader that this is just like sort of canned and there's no life to it. The thing that's difficult for us, honestly, Rob, is that we don't tend to get the credit for making it pop, because when you read something that has a sort of life to it, you're so used to reading writers when you're reading about sports, they just like, Oh yeah, that's what it's supposed to sound like. Of course it's supposed to have this pie, you know? So has it, like, it's kind of the book whereas if we weren't doing as good of a job, it'd be really obvious to people and I'd be like, this is can. But then when it's written right, it's just kind of like, Oh yeah, that's how you write. You know? It's like, yeah, but the computer's doing it, man. Like this isn't easy.
Rob Cressy: (05:31)
So taking a page from a joke from Chris Rock, taking credit for shit you're supposed to do.
Steve Wasick: (05:37)
Exactly. That's exactly what it is. A perfect way to put it. And it's just like, but yeah, but like you do it or something. I don't know, I want to say that it's easy, it's just like I said, it's because of how we already were attuned to reading a certain way.
Rob Cressy: (05:51)
So one thing that I really liked about this concept was the volume of content. So when I look at areas where brands need improvement, its volume is certainly one of them because brands are more regressive than progressive. When it comes to the volume of content specifically, they feel like there's this orbit around their brand. If they tweet one too many times or one too many Facebook posts, they can't do it. So the default for a lot of content is less but its also because of they don't have the staff and or mindset to create higher volumes of content. And actually I think that mindset is number one, then the staff is number two because if you, the more content you create, the more value you can deliver, the more awareness you're going to get, the more value plus awareness equals the more opportunities for fan engagement.
Steve Wasick: (06:50)
And then as you continue to trickle this down the road, there's just going to be a funnel to drive them to whatever it is. So talk about the volume side of things because it is such a crucial piece to this because you guys have been able to cover all of these sports in such a way that it's very value driven in so many areas. Yeah, I mean, I think that's our big value add is just that, hey, once we've got this thing set up a weekend, yeah, we can kind of deal with scale pretty easily. We delivered over a hundred million fantasy reports to people on CBS sports. We're at our peak when sort of all of our sports overlap. We're delivering about 5,000 articles to CBSsports.com every week. You know, covering all, you know, it's like when you have an article written, let's say for a preview for an upcoming game you know, it's a lot of work to then just like do an update, right?
Like if you have a writer on staff saying like, Hey, let's update it with the new information, the betting lines have changed or anything else, but it's like our system can do that automatically. So you can start thinking about about content in a very different way. Once you have it set up, you know in an automated system, the marginal costs is close to nothing to do more. So that's a different mentality and I think just like you said, a big reason why people don't do more is because they have constraints on it, Dude. You know, staffing or however. If you eliminate those, then the calculus really can change.
Rob Cressy: (08:32)
So with me having a background in the publishing side of things, having created thousands of original articles and videos and podcasts on my end, what immediately comes to mind for me is the way that this is replacing writers and so often we hear the intersection of technology in current jobs, where will technology go to make things more efficient and better. With what you're doing, you're correct. Any one of us couldn't publish 5,000 articles or what would it take for the manpower to do so? So what are your thoughts on where this can go moving forward? Because I understand the efficiency of it, I understand the volume of it, all of it makes complete sense. And so often when running any business, if you talk to a CEO or founder, they say, listen, at some point you're going to have to offload things and maybe your team can only get it 80% is good as it would be otherwise. So let's say what you do is only 80% as good as if I were to write that article myself. But if we're going to magnify that times 5,000 times infinity was all of that. That's a very exponential growth.
Steve Wasick: (09:46)
Yeah, I think what I'd say is this, I mean I think that on a personal level I definitely worry about technology, especially AI down the road. But honestly I think writers are the last people that are going to lose jobs from AI. Like literally the last people because just like you said, if you're a good sports writer, your article is going to be better than our computers. We've written some stuff that if you read it, there's no way you're going to tell that it was written by a computer. Like you can read some of it. Some of the things that we've done, I said there's actually, if you don't mind a quick digression, like I say that it's sort of like three levels of automated content. And the first one is you read it and you say, this is written by a computer. Like it reads fine, but it's written by computer. The second level is you read it and it seems like it's fine. Like you don't notice it's written by a computer, but if somebody tells you, Hey, this was written by a computer, you can go back and read it and say, yeah, I can see that. I can see that or you read five of them in a row and you can see that. And then there's a third level, which is you read it, somebody tells you it's written by a computer. You say no effing way. There's no way a computer redness or wrote this and we've dabbled into that third territory. You know, where it's like it really, really is good, but it's still never going to be as good as a human writer.
Steve Wasick: (11:08)
CBS is never going to use our system to write the preview for the super bowl. They're never going to do it. And it just, in general, if there is a piece of content that would be valuable if written by a human being, it's being written by a human being. There's somebody who is already doing it. There's plenty of human beings who are writers out there. So the work for writers is sort of, it's already there. And people who have jobs doing the things, they're not going to lose him. Our system is to come into places where it's just not economical to write. Like those, that's where we come into play and there's no, like with CBS, it's like there have been some articles that are, you know, been written by us that might have been written by a human being in the past, but it's like that human being is now just writing something that's high, higher, add something about trades, you know, highlights, interpersonal stuff. Like, like, you know, just thinking like in bigger terms, in more sophisticated terms. They're not like nobody that we've ever worked with has thought about using our system as a way to sort of put people out of their jobs because there's so much content that needs to be written up by human beings right now that if you have somebody who knows how to write, you can easily find other things for them to do and we're just taking the more like pro forma stuff off the table and the stuff that's like not economically viable.
Rob Cressy: (12:27)
So let's talk about automation in general for ways that we, or someone listening to the podcast right now may be able to implement it in their business. Because what I love about this general concept is one, the scalability of things. And I remember when I was working at CareerBuilder, way back in the day I was selling digital advertising, banner ads, text links, all that stuff. And we had one client who was spending $300,000 a month with us on text link advertising at the bottom of a page on Microsoft. And it was at that moment that I learned about scalability because what they did was they tested the offer. We were able to convert it at a positive ROI for them. And then they literally backed up the Brinks truck and said, let's full blow do this on and just like that I learned about scalability. So that's what I liked so much about what you're doing is what this allows. Essentially having one writer write 5,000 articles, not possible, but you go, boom, let me do this and find a way to help you scale this. So there's just so many possible applications for this. So help us understand where this might be more applicable in everyday business.
Steve Wasick: (13:54)
Yeah. I think the key thing is there's two keys to identifying something that's going to be good for automated content. The first is you've got to have structured data, right? So our computer can't look out the window and tell you what's going on. It's can only look at, you know, numbers in a spreadsheet or something like that. Now they can be from all over the place. It doesn't have to just be data that you have. I mean, there's publicly available data like weather data or stock market data that could be integrated into a report for you. But whatever those sources are, and they have to be structured, they have to be numbers or, or words or however. The second issue is you need something that is scalable. Like it's not just that that's a benefit. It's sort of a requirement because it takes us a lot of work to build this up. I compare it to, you know, it's like we're building a car factory for you. And once the fact that he's built, you know, you can produce a bunch of cars, but it costs a decent amount upfront to build that factory. So if you have something that meets those two criteria, which people have a lot of things that meet those criteria, right? That where they have a lot of data and they want to be able to communicate it and they need to do it a lot. Right? then you've got something where automated content to come and help and I do want to point out we've been talking a lot about sort of narrative reports, but our system is it's not just the long narratives.
Steve Wasick: (15:28)
I mean, we can do data visualizations, sort of walking you through your data that way and not necessarily like, Hey, you need a 300 word report, but maybe you need, you know, three different paragraphs with some charts, right? Telling you what's going on with your data and that's something that we can do too. So it's, so anything that meets those criteria we can, we can provide a report, whether that's narratives, charts, graphs, specific forms that you got to send off to the government or something like that. Like all of that can be, can be automated.
Rob Cressy: (16:00)
What is on your mindset from a forward thinking standpoint. So within this conversation right now, help us understand what we should be paying attention to that has your interest right now.
Steve Wasick: (16:13)
Oh my goodness. I might not have a good answer, that question because the only thing that has, I'm like just sort of overwhelmed with work. So it's like I don't know about general things that are going on. But I really do think that automated content is going to be a big deal in the future. Whether it's us or other companies because when you really start working through it like I've done like as I mentioned at the beginning of the podcast, I just thought this would be a cool idea for a fantasy league, right? Like I was just like, Hey, how about I just start doing this? Cause I think it'd be cool to know what else is going on in the league. I don't know. And then as I came and kept working through it, I was like, oh actually this is like a bigger deal. I could do this too to write sort of reports for other stuff. Right. And then as I kept working through it and as we sit now as a company, it's like, actually I think this is a pretty big step forward in artificial intelligence in general because it's a really complicated, complicated problem to say, hey, look at this set of information and figure out what's interesting and communicate it. You know, using natural language or charts or visualizations or what have you and especially to do that over a series of time. That's one thing that's really critical and very difficult is it's one thing to get a report one time, but then if you're writing that same report the next day, whether it's on a team or financial data or whatever, you got to take into account what you just wrote yesterday. You can't just, you know, if something's really interesting, you can't just write about the same thing 10 days in a row. It's not interesting anymore. You already mentioned it, you already mentioned that the markets, you know on a trend up, so that's where the complexity of sort of that, again, that's probably just kind of, if you're reading it, you don't even think about because again, that's how you write. Of course you take these things into account, but it's really really difficult. So I think that you know, what we're doing and what other companies in this space are doing, it's, it's really about understanding information and communicating it at a very sophisticated level. And that's something that's going to have, I think, really wide ranging effects. It's not just as simple as like, Hey, cool, we got a fantasy report on my league. It's really about understanding information and communicating it that, you know, that's a big deal I think.
Rob Cressy: (18:27)
In the name of improvement. What is the next thing that's on the horizon right here? Because like anything, what you've got now, you're, you're teaching it, you're constantly getting better and better and we're seeing that. All right, well if someone's going to write a piece of content that isn't going to be automated, there's the opportunity for that and we do this thing over here, but what sort of the next add on or bell and whistle to make this even more attractive?
Steve Wasick: (18:57)
I think for us, we're working through our data visualizations is a big thing. It's just kind of being able to, I mean, we've had this capability for a while, but like, even just something as simple as like if you're talking about a sporting event and you mentioned, Hey, such and such tight end has had three straight games of a hundred yards receiving, you just have a sidebar with their picture and here's the, here's the games and here's their yardage, or here's their stats. Stuff like that. I mean, we can do that right now and we're trying to integrate that with the CBS things, but really that's going to be our main focus this year is just kind of like really tying in the narrative and the visualizations in ways that allow the story to pop even more and give it that, that readability and accessibility in a way that just words by themselves don't.
Rob Cressy: (19:47)
One thing that I have not seen done successfully yet is this concept transitioned to video. I've seen where there's been some automation where it's like, Hey, let me take any scenario. So I think of I think it was the Japanese automated digital shorts of tiger woods scenario when he got in trouble with his wife and they like cartooned it all but a robot did the entire thing. What is the opportunity, if anything, to use this concept with video? Because from a content perspective on my end right now, sure written, but video is the next level of thing. It's more engaging and there's just so much more opportunity.
Steve Wasick: (20:33)
See, Rob, you got me talking man. I wasn't going to mention it, but yeah, we're working on it. We think that we can 100%. I mean, we've already got a lot of the goals in place to do it and that's gonna be something since you mentioned it, I wasn't going to say, but I suppose since you mentioned it, yes. All the pieces are there, because once you have a narrative, you have to understand. I mean, you know narrative's a narrative, whether it's written down or whether somebody speaking it and it's just a question of providing the right visuals, right to go correspond with that narrative for a true video experience because otherwise it just sort of text to speech, which also has value but you have to have the right sort of videos, animations, that type of thing. But we have a pretty good start on that right now and we have a good structural framework for that and that is absolutely something that we intend to start providing at a quality level. You know in the near term.
Rob Cressy: (21:33)
Cool. And I like that cause I'm just thinking about all right, how in the world would I, someone who doesn't do this attempt to do this? And on the visual side of things, I do believe that the bar is relatively low. From a standpoint of I'm envisioning Patrick Mahomes throws five touchdowns. And what would have video of him doing that look like? So sure. I can read about that, but maybe above that because you see this all the time and all the fantasy apps is they give you the variety of different things and I could see a static cutout image of Patrick Mahomes moving or just slightly gradually moving forward. And then here's the stat overlay and then your stat overlay could very easily go boom, boom, boom, here's his stats over the last three games and here's what it looked like this game and do it in such a generic yet simple way because that's what you really see on SportsCenter or any of these things. Their general graphics to help support the overall narrative.
Steve Wasick: (22:36)
Yes, exactly. So yeah, and that's basically what, what we can do and what we intend to do here, like I said in the somewhat near term.
Rob Cressy: (22:47)
Awesome. So as we wrap this up, I really enjoyed this conversation. Is there anything that I didn't ask you that you think would be beneficial for the audience to know regarding this conversation? We just had?
Steve Wasick: (23:02)
I don't think so. I guess the one thing that I didn't say explicitly is just because we've been focusing on sports is that the tech that we have as sort of content agnostic, we've been doing a lot of sports, but it can apply to anything. It's not, you know, there's nothing special about sports and then that's just where our original contracts came from. So we focused on that. But this tech, you know, a big reason why I think it is going to be so transformative is that it's not limited to sports. You know, we're working on finance stuff, medical stuff you know, all sorts of things. Anybody that has structured data, it's like this system, can come along and help. So I guess but mostly you got me to say more than I even thought I was going to say. So I'd say it was pretty comprehensive.
Rob Cressy: (23:51)
Good job. Complete. So, Steve, where can everybody connect with you?
Rob Cressy: (23:58)
I'm a little bit of a social media hermit, but I am on LinkedIn. So you can contact me there but that's it unfortunately. As always, I would love to hear from you about this episode. I'm curious, do you think you would notice content, specifically video that is created by AI as opposed to if it was not, you can hit up FanFood on Twitter and @fanfoodondemand on Instagram @fanfoodapp or on LinkedIn. And as always, you can hit me up on all social media platforms at Rob Cressy.