The company knowledge that doesn't rot in the head of the senior

Create wikis per team, department or project, every page with Markdown editor, inline images, code blocks and attachments. Every change is versioned, you see who changed what when and roll back to any previous state. Permissions per page or area, read and write separately controllable, inherited along the hierarchy with explicit override option. Hierarchical navigation with drag-and-drop, cross-linking between pages recognised automatically and maintained, link breakage on move is avoided. Stale detection marks content not updated for twelve months, the responsible person gets a notification and can release, revise or confirm the page as 'current and valid' without content change. On employee change, responsibilities are moved automatically, no wiki entry remains orphaned. Approval workflow for sensitive areas optional, four-eyes principle before publication with electronic release.

Ask questions in natural language, the AI helper searches wikis, attachments and PDFs and delivers the answer with source reference and verbatim quote. Works across all modules, including contract attachments from the HR module, audit reports from the ISMS module or accounting documents from the finance module. Vector embeddings are stored multi-tenant isolated in Postgres with pgvector, your knowledge never leaves your tenant. Mistral-based, EU-hosted at Mistral in France, no data outflow to the USA. No training on your data, contractually assured via DPA addenda. Permission awareness built in: only content the user is allowed to see flows into the answer, others remain invisible even as implicit background information. Answers cite concrete passages with reference to wiki page, section or page number in PDF, you click directly to the original source.

Recurring questions from the support ticket system are automatically clustered into FAQ suggestions. The AI helper suggests question and answer based on the actual solution responses of your support team in comparable cases, you edit and publish in one click. Published FAQs are linked directly in the helpdesk, new tickets get matching suggestions shown before they even reach support, the ticket creator often finds the answer themselves. Ticket load typically drops by 20 to 30 percent in the first three months, documented with before-and-after comparison in the dashboard. Clusters are recalculated monthly, new accumulations the system recognises automatically and reports as candidates. Multilingual FAQs (German and English) are maintained in sync, translation suggestions included with validation by native-speaker workflow. Statistics in the dashboard: which FAQs reduce the most tickets, which topics are underrepresented, where are open knowledge gaps that block the support team and where is optimisation potential.

Define structured onboarding paths per role, e.g. frontend developer, accounting, sales or management. Tasks, mandatory reading, training and tool access are presented in order, the new employee clicks through and ticks off steps, with progress indicator and intermediate stages for motivation. Supervisors see progress in the dashboard, can send reminders and mark critical steps as mandatory with hard deadline and escalation on default. Mandatory training for GDPR or ISMS is automatically embedded, completion is documented audit-proof with time stamp and content version, violations against mandatory training are visible in the compliance report. For recurring training (annual refresher) the path is automatically reassigned, you don't have to maintain anything manually, the system recognises deadlines itself. Onboarding time measurable, typical reduction 30 to 40 percent documented with before-and-after comparison. Paths are versioned and comparable with diff view, you improve over time without data loss.

Every employee has a skill profile with competencies, certificates and project experience. Search for employees who know Postgres performance, have experience with DATEV migration or hold an ISO 27001 auditor certificate. Knowledge no longer just rots in the head of the senior, but is documented and findable for project staffing, substitute arrangements and knowledge risk control. In case of illness or resignation you access documented procedures, not assumed experiential knowledge. Skill ratings are self- and supervisor-validated, the AI helper suggests skill updates from project participation. Anonymous search function for sensitive topics available, e.g. who has experience with compliance incidents. Skill risk heatmap shows critical single-person skills and prioritises substitute build-up.

Migrate existing Confluence spaces via one-click import including hierarchy, attachments and version history. Cross-links are remapped automatically, images embedded, tables formatted. Notion workspaces, Markdown repos from GitHub or GitLab and SharePoint knowledge bases are also imported with structure preservation. Migration of the last twelve months typically runs in one working day, anything older optionally against additional effort. Migration with mapping preview before import: you see what goes where, which conflicts (same titles, duplicate URLs) would arise and how the system would resolve them. Smoke test after import generates report with link breakage detection, image verification, permission mapping and full-text search indexing. Non-imported content is documented in the report with reasoning.