In a Nutshell
We read assorted documents grouped into logical entities, apply tenant-specific prompts to perform tenant-specific extractions, and expose collected insights via agentic conversational search, traditional faceted search, as well as via an API . We read documents of various formats, extract and interpret embedded images, slice presentations into slides.
You get access to the knowledge extracted out of the documents in the form of entity-level aggregated summaries (example - scope summary from an SOW, solution summary from an RFP response). You can prompt for visuals (slides and images depicting architectures, business processes, strategies). More importantly - you ask your questions in natural language just like you would ask a human or the infamous WFT Project Managers DL and get back structured answers with relevant projects, opportunities, case studies, and professionals who have relevant experience.
You can create knowledge bubbles by pinning relevant information to a collection. Collections allow you to have intelligent conversations scoped to the subset of relevant past experience that you yourself have selected. You can add more content by uploading your own artifacts that we do not have access to. You can share your collection with others and collaborate.
What you do with DI is up to you. We hope that you don’t just come because you are curious. It’s easy to ask a simple question, do not get a satisfactory answer, and then dismiss the whole concept as incapable. That would be a mistake. Instead, come to DI when you are looking for where and when we did X, how did we approach solving a problem Y, who was the client that we did Z for and who are the people who have the knowledge relevant to what you are working on.