Generating interest in generative AI!

Here in our Automation org, we’re really excited about how generative AI (gen AI) has disrupted the industry. Everyone is quickly trying to integrate gen AI aspects into their offerings. Clearly, we’re at a tipping point in what clients expect out of their software thanks to this (and that’s not a bad thing, for clients, or the people producing that software). While trying to “win” this race, our org is looking at trying to find spaces where we can take watsonx and make gen AI work for us in a productive and meaningful way for our clients.

That brings us to our documentation. For anyone who’s not familiar, our doc ecosystem is huge. With thousands of products, that’s probably to be expected. Our content teams have done a great job trying to facilitate finding content, but with the traction that gen AI is getting, we have to ask ourselves, “Are we doing enough?”. Content strategy and retrievability kind of go by the wayside when everyone is used to just asking your browser a question. Who hasn’t asked a gen AI solution a question these days? Whether you’re a huge fan of the technology, or fearful of what’s to come, a lot of people have tested it out, and reset how our industry views information retrieval.

So we’re trying to do something about this! In collaboration with our CIO office, Research, our Unified Search team, our IBM Docs tools team we’ve started to come up with a way to use gen AI in our content, training from our docs ecosystem, and produce hallucination-free results whenever someone wants to know something about an IBM product. One of the biggest obstacles our industry faces is the amount of “non-approved” information out there. When you ask other gen AI agents a question, you get back an answer from 2021 flavored by whatever was out on the internet. We’re trying to come up with a way of using verified, trusted sources and providing you with reliable answers!

Stay tuned!