AI TRAINING
Knowledge Base to AI Assistant in Two Days
Turn your existing internal documents into a working, staff-facing AI assistant without writing code.
What it covers
A two-day hands-on workshop that guides operations and IT leads through the full journey of connecting existing document sources — Notion, SharePoint, Google Drive, printed manuals — to an AI assistant employees can actually query. Participants configure at least one live tool (Notion AI, Glean, Dust, or Microsoft Copilot) against their own content during the session. By the end, teams leave with a functioning prototype, a maintenance playbook, and a clear understanding of data governance basics required to deploy safely. Format is instructor-led with a 70% hands-on ratio.
What you'll be able to do
- Configure a document connector in Notion AI, Glean, Dust, or Microsoft Copilot against your own internal content
- Apply a structured document audit to identify which sources are AI-ready and which need cleaning
- Write and validate prompt templates that return accurate, grounded answers from your knowledge base
- Set role-based access controls so only authorised staff can query sensitive content
- Produce a one-page maintenance playbook covering content refresh cadence and GDPR compliance checkpoints
Topics covered
- Auditing and structuring existing docs for AI ingestion (Notion, SharePoint, Google Drive, PDFs)
- Choosing the right tool: Notion AI vs Glean vs Dust vs Microsoft Copilot for your stack
- Connecting document sources and configuring retrieval settings
- Writing and testing effective internal queries and prompt templates
- Setting access controls and user permissions for staff-facing deployment
- GDPR basics: what data should and should not enter the knowledge base
- Building a content maintenance and update workflow
- Measuring adoption: usage metrics and feedback loops
Delivery
Delivered in-person or via live virtual classroom over two consecutive days (6-8 hours each). Participants must bring access credentials for their own document repositories on day one. The session is 70% hands-on labs and 30% facilitated instruction. A shared Slack or Teams channel is set up before the workshop for async Q&A and remains active for 30 days post-training. Printed quick-reference cards and a digital configuration checklist are provided. Remote delivery requires each participant to have screen-sharing capability and admin rights to at least one document source.
What makes it work
- Nominate one internal 'knowledge base owner' before the workshop who will be accountable for content refresh
- Start with a single, high-traffic use case (e.g. HR policy Q&A) to prove value before expanding scope
- Run a staff feedback loop in the first two weeks to catch hallucinations or gaps in source material
- Document GDPR decisions made during configuration so future contributors follow the same rules
Common mistakes
- Dumping all company files into the AI tool without first auditing quality — garbage in, garbage out answers erode trust fast
- Skipping permission mapping and exposing HR or financial documents to all staff by default
- Choosing a tool based on brand familiarity rather than compatibility with the existing document stack
- Treating the workshop as a one-time setup rather than establishing a content maintenance routine
When NOT to take this
If the organisation has fewer than 50 internal documents or its staff already relies on a well-maintained wiki that everyone uses, the overhead of configuring an AI layer adds complexity without meaningful time savings — a better-organised wiki is the right fix first.
Providers to consider
Sources
This training is part of a Data & AI catalog built for leaders serious about execution. Take the free diagnostic to see which trainings your team needs.