What is Talvi?
Understand the centralised knowledge layer behind datasets, AI agents, learning, and connected tools.
Your organisation’s knowledge, ready to use
Talvi is a centralised knowledge layer that connects organisational content with the people and AI tools that need it. It works alongside your existing knowledge and learning systems, helping you make maintained source material useful in everyday work. This is the focus of Talvi’s Trusted Enterprise Knowledge overview.
For an employee, that means asking a question and following the answer back to its evidence. For a content team, it means maintaining the knowledge behind those answers and turning useful material into learning people can reuse. For administrators, it means controlling the datasets, agents, and audiences involved.
How the pieces fit together
| Part of Talvi | What it does |
|---|---|
| Source content | Supplies the documents, audio, video, and learning packages your organisation makes available. |
| Dataset | Organises that material into a knowledge base. Several agents can use the same dataset. |
| AI agent | Combines a dataset with a role, mission, and any enabled tools to support a particular audience or task. |
| Access point | Gives people a way to use the agent: Get Answers, an embedded website bot, or a supported integration. |
| Administration | Keeps content availability, agent access, and learning publication under your team’s control. |
A shared knowledge layer does not mean everyone sees everything. Users work through the agents and content made available to them. Connected tools and external providers can also have their own permission requirements.
Build value across the knowledge lifecycle
Bring useful content together
Start with a dataset and maintain it through Data Import. Supported material includes documents, recordings, and SCORM learning content. External import sources depend on the integrations configured for your installation; see supported file types and import options.
Keep people in control of the sources
Use File Management to review availability and Talvi Data Cleaner to compare similar files before deciding what to keep. Maintaining shared knowledge helps multiple agents benefit from the same content work. Content owners still decide which version is appropriate and which material should be removed.
Answer a question and show the evidence
In Get Answers, people ask in their own words and inspect the sources behind a response. Where supported, they can open the relevant PDF page, audio or video timestamp, or SCORM section. The original material remains available for context and checking.
Make learning reusable
Use Talvi Coach to explore a source and Talvi Nuggets to create useful formats such as checklists, guides, and quizzes. Administrators can review and distribute those Nuggets. Optional Mentor with Memory agents add journey guidance developed with your team.
Improve the knowledge over time
Review questions and activity, feedback, and Nugget analysis reports to identify areas worth improving. Use the findings to prioritise source updates, clearer guidance, or better distribution, then check the effect with users.
One foundation, different experiences
Your organisation can use the same dataset through agents with different purposes—for example, an employee support assistant and a learning mentor. The role, mission, audience, and enabled capabilities determine the experience each agent provides.
Where configured, Talvi MCP Server makes authorised Talvi knowledge available to compatible AI assistants. A Talvi Embeddable AI Bot brings a configured agent to a website. These routes extend where knowledge can be used; client support determines how tools and interactive sources appear.
Features vary by installation, agent, permissions, and available content. Ask your administrator which routes are enabled for your organisation.
Choose your next step
- Use Talvi: Ask your first question and check its sources.
- Maintain knowledge: follow Talvi Recipes for imports, cleaning, access, and distribution.
- Build an experience: open AI Agent Studio or explore MCP connections.
For broader product context, visit Talvi for learning teams and Talvi for IT teams. To explore an additional capability for your installation, talk to Talvi.