AI search for internal SharePoint knowledge

For Energiekontor, we worked on the architecture of an AI-powered search across internal SharePoint content. The focus was on how an OpenAI-based assistant could make existing company knowledge accessible through Microsoft Graph.

The technical concept combined a chatbot interface with a connection to private SharePoint content. Requirements for handling internal data were part of the architecture work from the start.

Our contribution covered software architecture and the technical foundation for hosting and automated deployment on Microsoft Azure. The knowledge source, access through Microsoft Graph and processing by the assistant were considered as a connected workflow.

In 2023, we used this architecture to lay the foundation for conversational search across company knowledge.

SharePoint Online provides the knowledge source, with Microsoft Graph as the intended means of accessing the content. A chatbot interface and OpenAI connect research with conversational interaction.

The architecture work takes Microsoft Azure into account for hosting and the server environment, along with automated deployment processes.

Period: 2023 · Architecture and concept.