AI DocBot app
NetApp's extensive technical documentation, spanning thousands of pages in multiple formats, made it difficult for internal teams to efficiently locate answers to complex questions. Navigating the system was cumbersome, often leading to unanswered inquiries and wasted time. The product team was tasked with improving efficiency and streamlining access to critical information.
AI to the rescue
The product team developed a DocBot, a custom AI chat bot using a fine-tuned LLM trained using thousands of NetApp's technical documentations from across the enterprise. This solution enables analysts, product owners, and engineers to efficiently search and navigate thousands of technical documents using natural language, through a single source of truth.
Wireframes
Wireframes of the UI structure and app user flow were presented to stakeholders for buy-in and the engineering team to discuss the technical realities of our design concepts.
Hi-fidelity concept mocks
Hi-fidelity Figmocks were used to garner feedback from user groups and define MVP priorities.
Code generation & 'document chatting'
Again, leveraging user interviews and surveys the design team introduced use-case specific features such as generating real-time code for easier API integrations and focusing the chat experience around specific documents to help frame the context of the responses.
While the product design portion of this product has been handed off, The application is still in production while the SageMaker LLM is being trained, validated and tested. Once the engineering teams and API teams have wired up the UI our team will reengage and begin usability testing.