AI answers from your
internal knowledge base
Give every employee instant, cited answers from your runbooks, policies, IT knowledge base, and internal wikis — in Slack, Teams, or any surface. Reduce internal helpdesk volume and get new hires productive faster.
Built for internal knowledge ops
Every feature designed around how enterprise teams actually work with internal knowledge
Runbook & Policy AI Search
Index your runbooks, policies, SOPs, and IT knowledge base. Employees get precise answers grounded in your actual internal documentation — not hallucinated advice.
IT Helpdesk Automation
Deflect repetitive IT tickets before they reach your service desk. Kelu answers common requests — password resets, VPN access, software provisioning questions — instantly.
Confluence & OneDrive Connector
Connect your existing Confluence spaces, OneDrive sites, and Notion workspaces. Content stays in sync as your team updates it — no manual re-indexing needed.
Microsoft Teams Bot
Employees ask questions in Teams channels and DMs. The Kelu bot replies with cited answers from your internal knowledge — right where work already happens.
Role-based Knowledge Access
Engineering answers come from engineering runbooks. HR answers from HR policies. Collections and SSO group permissions ensure each team sees only the knowledge scoped to them.
Onboarding Acceleration
New hires get accurate answers to process, tooling, and policy questions from day one — reducing time-to-productivity without burdening senior staff with repetitive questions.
How it works
- 1Connect internal sources — Confluence, OneDrive, Notion, Google Drive, and internal wikis. Kelu indexes everything and keeps it current with scheduled syncs.
- 2Configure access controls by mapping SSO groups to knowledge collections, so each department sees only the content scoped to its role.
- 3Deploy the Teams or Slack bot in your workspace so employees can ask questions in the channels they already use.
- 4Employees get instant, cited answers grounded in your runbooks, policies, and wikis — low-confidence questions fall back to your helpdesk or the right team channel.
- 5Analytics surface the most common questions and knowledge gaps, so you can close gaps and watch internal ticket volume drop.
What you get
Trusted by enterprise IT and ops teams
“Kelu powers our internal knowledge assistant. 200 engineers ask questions daily and get answers grounded in our runbooks and architecture docs.”
“We connected it to our Slack workspace and Confluence in an afternoon. Our on-call teams now get accurate answers to internal process questions instantly.”
Frequently asked questions
Can Kelu index our internal OneDrive and Confluence sites?
Yes. Kelu has native connectors for both OneDrive and Confluence, and a website source reaches an internal wiki that neither covers. Content is indexed on a schedule and updated incrementally as pages are edited. Access-controlled content can be gated by SSO group membership so employees only see knowledge they are authorized for.
How does access control work for sensitive internal content?
You organize knowledge into collections (Engineering, HR, Finance, IT, etc.) and map each collection to SSO group membership via OIDC or SAML. An employee in the HR group can query the HR collection but not the Engineering collection. Access is enforced at the API level on every query.
Can we deploy the Teams bot for all employees without IT bottlenecks?
Yes. The Teams integration is installed once by an admin and becomes available across all channels and DMs in your tenant. Employees can @mention the bot or use a slash command. No per-user setup is required.
What happens when the bot cannot answer a question?
When confidence falls below your configured threshold, Kelu returns a transparent fallback message pointing the employee to the correct internal team or escalation path. You control the fallback text — typically pointing to your IT helpdesk or the relevant team channel.
Do you support on-premise or VPN-gated internal content?
Yes. Enterprise customers can deploy a private crawler agent inside their network to index content behind a corporate VPN. Kelu also supports IP whitelisting for the crawler. Contact us to discuss your infrastructure requirements.
Is our internal content used to train AI models?
No. Your indexed content, embeddings, and conversations are isolated to your workspace and are only used to answer your own employees' questions. Content is never shared across customers, and data is encrypted at rest and in transit.
How does content stay up to date as our wikis change?
Sources sync on a schedule and are re-indexed incrementally, so edits to Confluence pages, OneDrive documents, or Notion pages flow into the knowledge index automatically. There is no manual re-upload or re-training step.
Ready to build your internal knowledge assistant?
Connect your first knowledge source free — no credit card required.
Get Started Free