📚 node [[toby watson s use cases]]

Toby - Watson's use cases

Go to [[Module 4 - Watson Use Cases and Resources List]] or the [[Main AI Page]]

Transcript

Toby, tell me tell me about some successful use cases. Yeah. Sure. So today, those customers that are most successful their use cases fall into three broad categories. Those categories are going to be expanding rapidly. But the first one is really enhancing your customer interactions, and these are internal and external customers. The second one is how do I deal with business processes that may have regulatory, compliance or legal ramifications for my business? These are large complex documents, large complex regulations. Then thirdly is how do I enhance my development and operations by DevOps Team and bring it into my organization? So if you look at those three categories and the first one is really where AI is pervasive today. How do I enhance my customer experience? How do I add value to both my internal customers and my external customers? A prime example of that is one of our customers down in Brazil, Bradesco which is one of the largest banks in Brazil, and one of the first adopters, and really the first adopter in Brazil of AI. They chose an internal use case. When I say an internal use case this is for their advisors, talking to their customers. When a customer asks a question, and their first use case is about a retail banking products specifically, that internal advisor would have to go find the answer, whether it will be research on knowledge base or finding another human.

Today, Watson can answer 85 to 90 percent of those questions on the spot for that advisor, so they can be more responsive to their customer. That drives a lot of value for that customer, and also drives a conversation between that advisor and that customer that's less about finding information, and more about adding value. Interesting. For their specific problem. Absolutely. Toby, you mentioned the second way that a lot of people are getting started as through compliance. That's right. So when a compliance or large complex topics regulation, so we're really going to see a lot more of as we go forward is, how do I take these large complex documents, and allow make my organization to do things more quickly, in a more valuable way such as, we have customer of ours Cisco that uses Watson to interrogate statements of work? Statements of work are contractual instruments between an organization and a service provider, that layout typical contractual terms but also obligations, pricing, requirements, all those things. Anybody who has ever been in a world has dealt with statements of work, one of the challenges is, how do I understand what my obligations are as a part of the statement of work? Guess what? Watson now can draw out those obligations, but also how do I potentially compare what a provider of mine is asking of me to what my standard terms are? That would require lawyers and a lot of reading. Now Watson can do both the drawing out of obligations, and also the comparison and we're doing that with Cisco today. Okay. So far just a quick recap. We have looked at conversational assistance as one way to- That's right. -look at a use case, then compliance for your client-oriented business processes. Then third, I think you were talking about using this within your organization, and that's like a Dev shop. That's right. How do they bring it into their Dev shop? How do I bring it in and make sure this isn't really a capability that I can use on an ongoing basis? So we just recently announced and launched Watson Studio. Watson Studio is exactly that facility to allow organizations to do that. It's a single place for builders and developers, to go to build, and will have pre-trained models. We talked about those already a little bit and that's going to be in there, but the real value is going to be, how can I build my own models? How can I build my own models and deploy those to the organization? There's going to be deep learning models and machine learning models. But how can I build those within Watson Studio, but maybe even also leverage some open-source capabilities and technologies that are out there today that I already know very well in my organization? TensorFlow, Caffe, PyTorch, are some of the ones that we support, and there's others that we can get you information to as well. Then how do I develop that to a single experience? So my builders- and I've used the word builders a couple of different times because we can talk about developers, but Studio is focused on not necessarily deeply skilled. You don't have to be a deeply skilled AI practitioner, this deskills it and allows you to have more capability, more value for your organization. Therefore, it brings it to more of the organization to do so.

📖 stoas
⥱ context