What is Insight AI?
Insight AI is an AI suite for teams that turns fragmented information into accessible knowledge. It connects to where your documents already live, indexes their content into a searchable knowledge layer, and lets you ask questions in plain language — answered by AI models that can draw on your own content.
You use Insight AI in two places:
- The dashboard (dashboard.insightai.pro) — where you connect data sources, run ingestion pipelines, organize your team, and manage AI models.
- Chat — a dedicated conversation workspace (opened from the dashboard) where you talk to AI models and get answers informed by your indexed documents.
The core concepts
Section titled “The core concepts”Four building blocks connect your files to your answers:
| Concept | What it is |
|---|---|
| Data source | A connection to where your files live — a network (SMB) file share, or a Google Drive account. |
| Collection | A searchable index. When documents are ingested, their content is stored here so it can be found by meaning, not just keywords. Each collection belongs to a group, which controls who can search it. |
| Pipeline | The link between a data source and a collection. Running a pipeline reads supported files from the source and indexes them into the collection. |
| Chat | A conversation with an AI model. When you ask a question, Insight AI automatically searches the collections you have access to and gives the model relevant excerpts — and the answer shows you exactly which documents each passage came from. |
The flow is always the same:
Data source → Pipeline → Collection → Chat
In practice, the Create wizard handles the middle for you: connect a data source and the collection and pipeline are provisioned automatically. Then chat — your questions are enriched with the most relevant passages from your indexed documents.
What Insight AI indexes
Section titled “What Insight AI indexes”Pipelines extract and index text from common document formats: PDF, Word
(.doc, .docx), PowerPoint (.ppt, .pptx), plain text, Markdown, and CSV.
See supported file types & limits for details.
Who does what?
Section titled “Who does what?”Every signed-in user works in the same dashboard. Teams are organized through groups: data sources and collections are owned by a group, and only that group’s members can search or manage them. Groups carry a permission level and can be nested to mirror your team structure. Models and groups themselves are workspace-wide, so coordinate with your team before deleting anything shared.