Aviso AI’s Post

Transcript

On this episode of AI's Corner Office with the Visa, we're joined by Brian Gumbel, newly appointed president and COO at data miner, and Aviso AI CEO Trevor Rodriguez Templar Ryan Trevor. Stanford University's new 2024 I index had some interesting findings. The global survey on Responsible AI highlighted that company's top AI related concerns include privacy, security, and reliability. The survey shows that organizations are beginning to take steps to mitigate these risks. So Trevor. Tell me, what ways have you seen companies manage risk when using AI internally at their businesses? So we didn't go mainstream, especially with the advent of Chad GBT and open the eye and you know, you're looking at infinite possibilities from infinite amounts of data. But I think top of mind for most of our clients who happen to be, you know, folks in the security business like like Brian's old company Armis, folks like GitHub, you know, Microsoft, Microsoft, yeah, get up at Microsoft does use us. Lenovo just went with us over ChatGPT. So we've got we've got large companies and small companies, but I think the unlike team is they want to ensure that the protecting their IP. They want to make sure that it stays within the four walls of the enterprise, right? So when you look at Shaggy, BT and we look at Mama too, or in the open source models, they're really exciting, right? Because at the end of the day, what it is, is these models, I think like human beings based on neural networks. So we love that, right? But the fact of the matter is you want to make sure that your IP stays within your company. So what we've gone with the strategy is a hybrid model. And what I mean by a hybrid model is we leverage the strength of large language models in what we call small selected models that are trained. On your IP and your data and stay within the enterprise. So you get the best of both worlds. You get a context, which is really important because anything without context results in what we call hallucinations, you know, and other kinds of stuff, right? Or stochastic, what they call stochastic parrot where it sounds really authoritative, but really it's, it's not really answering the question, right? It's false information. So with that, with that in mind, we've seen tremendous response. We've got about 93% adoption of Mickey, which is our machine intelligence and knowledge user interface. And you know, we're very excited. We've now embarked on doing what what I call AI automated workflows, where we use multiple AI agents. I came to The Matrix to go in and automate about 90% of the work you do on a daily basis. And what that does for you is it allows you to liberate your inner Einstein or your inner. I'm not gonna say Elon Musk, but maybe you're in a Steve Jobs, you know, and, and really focus on on being a better version of yourself and, and being more productive. And, you know, I think Brian can speak, speak more to it, right? So, yeah, I think I think any organization right now is concerned about the use and misuse of AI within their organizations. And I think it's very important for companies to recognize that and start enablement programs and train their employees on how best to be able to work with AI as it's coming into organizations, whether it's. Than their own product or external products that are being used. So I think enabling companies and their employees is really top of mind and that's what should be done right now.