You Ask, I Answer: Monetizing Data for Generative AI?

Today's episode of You Ask I Answer was recorded in front of a live studio audience at the Digital Now conference in Denver, Colorado, in November 2023. The session title was appropriately, You Ask, I Answer Live. Generative AI QA. Enjoy. Great, thank you.

All right, so like I was announced, this is completely informal. This it's it's a wild west, anything goes. You have a question about anything you've seen or heard today, I'll do my best to answer it. Uh if not, I will make something up and uh and tell you that too. Uh so who has a question that they would like to tap?

Yes, the person in green. Um, gotta wait for the microphone. Yes. I'm hearing loud in real life. That's okay, I am too.

So I am already just feeling a lot of annoyance and pressure from a deep software stack that I would like to slip down. So the thought of in the making that software stack even figure uh with AI, something that's intimidating to me. Um, and I'd just like to hear your thoughts on keeping your your software stack manageable while also incorporating AI into what you're doing. Yeah, so in terms of managing your your Martech stack, uh a big question is gonna be who your stack is centered on, because pretty much, like if you if you're a Microsoft shop, if you're Google shop, if you're an Amazon shop, you've already got a vendor that's probably already been through the vendor approval process and stuff. And so that would be the place I would start.

Pretty much every tech company has an AI stack. So Amazon has AWS Bedrock and SageMaker, Google has Duet and GCP, Google Cloud, uh, Microsoft, of course, has Azure. So everybody's got something. So if you want to try to at least avoid adding vendors, um, that would be the place I would start. Um that's also probably a good place to start because you're gonna have less um issues with compliance if you're working within your existing tech stack vendor, because if they've got an offering, that chances are in somewhere in the terms of service, there's also like saying, Here's how we protect your data and so on and so forth.

That that's the first place. It will grow. So the Hey Gen video generation. That's a point solution right now. There are ways to engineer a similar system, but you've got to be super technical to do it.

So you'll probably have to use that vendor until one of the big tech companies offers that internally and stuff. So yep. Next question. Right there on the front. Right, I guess.

So I was in uh session on practical applications, and we were talking about monetization. Yes. And so we're all you know, association uh type organizations. Seems like there's a lot of ways to leverage AI for process efficiency, uh improved, you know, uh products and services, uh user experience. Are there specific things that associations should be looking at that could potentially generate revenue uh either directly or maybe indirectly from applications of AI?

How do you make money right now? Uh membership, uh publications, meetings, uh uh that kind of stuff. Right. So immediately you think about publications, right? So publications are locked into a format, right?

It's a thing. Um you can use the tools that exist right now to broaden the scope of what those things are. So, like if you want, uh there's uh Google Cloud has a really good spee uh text-to-speech API. Uh, one of the things that I've done with my past books I've written is I've said, okay, drop this in, turn it into an audiobook, and you get uh you know machine read stuff, but it is it is now in a format that is that you can listen to instead. So even just taking the existing content library you have and making it multimedia is a way to increase the value of the product, and you can either sell it separately or you can bundle in as a value add.

Um that would that's the first thing that comes to mind. The second thing that comes to mind is if you're using the tools for content creation, by default, you can just create a lot more stuff with their assistance. Um the of all the tools that are out there for long-form content creation, probably the best one right now is Anthropics Claude, because it can handle very has a very large context window, aka memory. Um it can process 60-ish thousand words at a time. So you could give it, say, a journal publication and say, hey, I want you to come up with four more articles based on this theme from you know, you copying this style with these style of citations and just amplify the amount of stuff you can create.

Um the third area, which is a lot more risky, is having uh enabling conversational interfaces to your content. Right. So someone, if you if you build that that fine-tuned model, someone could have a conversation about your what your association does. That I'm still a little, I would be careful with that just because the models themselves can still say really funky things at times. Um depending on the model you use, it can hallucinate.

It can just outright make shit up. Um, but yeah, so monetization, and like I said during the the the whatever thing, um turning your data into one of those plugins that you can then sell access to that as a service is another option. That's a more technical option too, but it's a way of leveraging all the data you have, particularly if you have a lot of public-facing data. You can make a conversation agent or a plugin for language models that specifically talks about that thing. Um and it can either be access to the data set itself, it can be access to the machine readable version of that, or it can be a plug-in for a model.

If you enjoyed this video, please hit the like button. Subscribe to my channel if you haven't already, and if you want to know when new videos are available, hit the bell button to be notified as soon as new content is live.