What is Verification?

Verification is Guru's system for marking content as trusted and current. It helps your team know which information they can rely on—and helps your AI tools ground answers in accurate knowledge.

What is Verification?

Verification is a trust signal that tells users whether content has been reviewed and confirmed as accurate. Every piece of content in Guru (Cards & connected sources) can have one of three verification states:

• Verified (green checkmark): Content has been reviewed and confirmed as accurate and current
• Unverified (gray indicator): Content has been flagged as potentially outdated or inaccurate
• None (no badge): Content hasn't been reviewed yet—neutral state with no indicator shown

You'll see these verification badges throughout Guru: in search results, in AI-generated answers, on Cards, and next to source documents.

Why Verification Matters

Most organizations face a fundamental knowledge quality problem: content decays faster than teams can review it. Manual verification processes typically reach only 8-12% of organizational content, leaving the rest to slowly become outdated.

The stakes are high. When AI gets inaccurate information, it confidently distributes that misinformation at scale. What previously affected one employee-customer interaction can affect thousands.

The cost of wrong information:

• Compliance violations from outdated policies
• Customer churn from incorrect support answers
• Productivity loss from employees following old procedures
• Erosion of trust in your knowledge systems

Verification creates accountability for content quality without creating bottlenecks.

Two Approaches to Verification

Manual Verification: Designated people directly review and approve content, setting expiration dates or marking content as evergreen. Ideal for critical content that requires expert judgment.

Automated Verification: Knowledge Agents continuously verify content based on usage signals, engagement patterns, and AI analysis. Handles verification at scale without manual intervention.

These work together. Use manual verification for compliance-critical documents while letting automated verification handle thousands of other documents. Override automated decisions manually whenever needed.

Manual Verification

Verification puts humans in direct control. Designated verifiers review content and explicitly mark it as verified or unverified.

Who can verify content:

• Members of verification groups assigned to Collections
• Collection owners and authors
• Users with "KA Source Records Verifier" permission (for external sources)

What you can verify:

• Guru Cards: Verify individual Cards with expiration dates or mark them as "Does not expire" for evergreen content
• External sources: Verify documents in Google Drive, Confluence, SharePoint, or any other connected source

Manual verification works best when content requires expert judgment, has compliance or legal sensitivity, or involves a manageable volume of high-stakes information.

Automated Verification with Knowledge Agents

Automated verification sets Guru apart from traditional knowledge management and enterprise search tools. Knowledge Agents verify content continuously and autonomously.

How it works:

Knowledge Agents run daily verification checks on content that's been used in Answers. Using three types of rules—behavioral (usage data), content-based (AI analysis), and analytical (engagement patterns)—the system makes smart decisions about what should be verified or unverified.

For example:

• Verify a document that's been thumbs-upped multiple times
• Verify evergreen content that remains consistently accurate
• Unverify time-bound content that's fallen outside its valid date range
• Unverify content that's been flagged as inaccurate multiple times

The system only reviews content that's actually being used. A 7-day review window prevents unnecessary churn.

What gets verified:

Knowledge Agents verify all content across your connected ecosystem, not just Guru Cards. Whether it's a policy document in Google Drive, a product spec in Confluence, or a process guide in SharePoint, Knowledge Agents continuously analyze and verify content everywhere.

Default settings:

• Auto-verify: ON by default for all Knowledge Agents
• Auto-unverify: OFF by default for all Knowledge Agents

Key benefits:

Automated verification eliminates 80% or more of manual review work while providing complete coverage across all your connected sources. It delivers real-time accuracy and creates a network effect where every verification improves knowledge for your entire organization and all your AI tools.

Transparency and control:

Every decision is logged in the Quality Log. You'll see exactly why content was verified or unverified, with clear reasoning. You can review and override any decision at any time.

Choosing Your Approach

Use manual verification when:

• Content requires domain expert review
• You're dealing with compliance or legal sensitivity
• Volume is manageable and stakes are high

Use automated verification when:

• Managing content at scale (hundreds or thousands of documents)
• Usage patterns are a reliable quality signal
• You want continuous, real-time updates without manual intervention

Combine both approaches:

• Let automated verification handle most content at scale
• Manually verify your most critical 10-20% of content
• Override automated decisions when you disagree
• Configure Knowledge Agents to route critical content to experts for manual review

How Verification Affects Your Knowledge

Verification status directly impacts how content is used throughout Guru:

• In Knowledge Agents: Configure to use only verified sources, verified and neutral sources, or all sources
• In search results and answers: Verified content is visually marked and can be filtered out of search results.

Important: When multiple Knowledge Agents reference the same external source, verification status is shared across all of them. These "disagreements" surface sources that might need human review.