Notifications
Clear all
Topic starter
08/07/2026 6:50 pm
TL;DR: Fragmented knowledge across SharePoint, Confluence and file servers forces employees into workarounds and drives repetitive support demand, while Matrix42 says its AI assistant returns answers from verified internal sources and reports a 60% drop in direct support questions. The governance issue is access, not information volume: knowledge only helps when it is governed, findable and trustworthy.
NHIMG editorial — based on content published by Efecte: AI assistant for knowledge discovery
By the numbers:
- Matrix42 says it produced a 60 % minskning av direkta frågor till supportteamen in production environments.
Questions worth separating out
Q: How should organisations govern AI assistants that retrieve internal knowledge?
A: Treat the assistant as a governed access path, not just a user interface.
Q: Why do fragmented document systems create support and governance problems?
A: Fragmented systems make it hard to find the authoritative version of a policy or procedure, which leads to duplicate tickets, inconsistent guidance and slower decisions.
Q: What do security and IAM teams get wrong about knowledge search tools?
A: They often treat search as a convenience feature rather than a controlled access channel.
Practitioner guidance
- Map authoritative knowledge sources Identify the systems that contain official policy, SOP and HR content, then designate one source of truth for each topic area.
- Constrain retrieval to approved repositories Limit assistant access to the internal systems that have review, retention and access controls in place.
- Add provenance to answers and escalations Require the assistant to surface where each answer came from and to escalate when sources conflict or are stale.
What's in the full article
Efecte's full article covers the operational detail this post intentionally leaves for the source:
- A walkthrough of the employee self-service workflow and how the assistant fits into existing support channels.
- Specific examples of how SharePoint and Confluence content is retrieved and surfaced in natural language.
- The production efficiency figures behind the reported reduction in direct support questions and form submissions.
- The article's practical framing for HR and IT teams that need to decide where to start with knowledge discovery.
👉 Read Efecte's analysis of AI assistant knowledge discovery and support deflection →
AI knowledge discovery for support teams: what changes for IAM?
Explore further
08/07/2026 7:11 pm
Knowledge discovery is an identity and access problem disguised as a productivity problem. The article describes a familiar friction pattern, but the operational failure is governance: users cannot reliably reach authoritative knowledge fast enough, so they create tickets or work around the process. That is not a content-volume failure, it is a controlled-access failure across human workflows and service channels. Practitioners should treat internal knowledge retrieval as part of the access model, not a separate convenience layer.
A few things that frame the scale:
- The average estimated time to remediate a leaked secret is 27 days, despite 75% of organisations expressing strong confidence in their secrets management capabilities, according to The State of Secrets in AppSec.
- Organisations maintain an average of 6 distinct secrets manager instances, creating fragmentation that undermines centralised control, according to The State of Secrets in AppSec.
A question worth separating out:
Q: How can teams tell whether AI self-service is actually reducing operational load?
A: Measure whether repeated requests, duplicate tickets and manual escalation volume fall after deployment, then check whether answer quality remains stable across teams and regions. A useful signal is that users can resolve routine questions without leaving the workflow, while support staff spend more time on exceptions rather than document hunting.
👉 Read our full editorial: AI knowledge discovery exposes the limits of fragmented access
