AI Risk Narrative [Command+]
The AI Risk Narrative is an AI-generated security briefing that distills your live integration data into plain English. Once a day, CMDash analyzes your connected data sources and produces a concise summary of your top security exposures — with specific recommendations for what to do about them. The Narrative is available on Command and Enterprise subscriptions.
What It Is
The AI Risk Narrative removes the step between raw security data and actionable insight, surfacing a synthesized read of your posture directly on your Dashboard instead of requiring you to piece it together from several cards.
Each Narrative identifies:
- A short executive summary of your current security posture
- Your top exposure areas right now, each with a severity rating
- Specific recommendations, each tagged by priority (immediate / short-term / long-term)
- A short trend analysis of where your posture is heading
The Narrative is written in plain English, without technical jargon. This makes it equally useful for a CISO reviewing it personally and for sharing a summary with a non-technical executive.
Where to Find It
The AI Risk Narrative is displayed on your Dashboard, in the "AI Risk Narrative" card. There is no separate full-page Narrative view — the Dashboard card is the only place it appears in the product today.
You can hide or re-show the card from your own Dashboard: click Edit layout, then remove or use + Add widget to bring back AI Risk Narrative alongside the other Dashboard widgets. This preference is per-user; it doesn't affect whether the Narrative is generated for your organization, only whether your own Dashboard displays it.
What the Narrative Includes
Each Narrative has four parts:
1. Summary
A 2–3 sentence executive summary of your organization's current security posture.
2. Top Exposures
Up to three exposure areas, each with a title, a severity rating (Critical / High / Medium / Low), and a one-to-two sentence description of the exposure and its business impact.
3. Recommended Actions
Up to three specific, prioritized recommendations, each tagged Immediate / Short-Term / Long-Term with a short rationale for why it matters.
4. Trend Analysis
A 2–3 sentence analysis of your security posture's trajectory.
The model is given your current posture score and signal breakdown, your compliance framework progress, your integration coverage and health, and your highest-scored open risks — so it may reference compliance frameworks or specific risks by name where relevant, but there isn't a separate, guaranteed "compliance impact" field independent of the summary/exposures/actions above.
How It Works
When CMDash generates the Narrative, the process follows these steps:
- Data aggregation: CMDash collects the current state of your connected integrations, posture score, compliance framework progress, and highest-priority open risks.
- AI generation: The aggregated data is passed to an AI model (Anthropic Claude), which generates the natural-language Narrative.
- Delivery: The Narrative appears in your Dashboard's Narrative card.
See the Privacy section below for exactly what data this includes.
Refresh Cadence
The Narrative is regenerated automatically once a day, at a fixed time (6:00 AM UTC) — not adjusted per workspace timezone.
A manual refresh is available: click Regenerate on the Narrative card. Manual refreshes are restricted to Owner and Admin roles and are rate-limited to one per hour per organization to bound cost. Other roles can view the current Narrative but cannot trigger a regeneration.
A manual refresh does not discard prior Narratives: click History on the Narrative card to open Narrative History, which lists up to the 10 most recent Narratives (each timestamped and tagged Scheduled or On-demand) and lets you expand any of them to read the full briefing.
Customizing the Narrative
The Narrative always analyzes all of your connected integration data, current posture, compliance progress, and open risks — there is currently no way to scope it to a specific framework, risk tag, or asset group.
The only customization available today is showing or hiding the card on your own Dashboard (see "Where to Find It" above).
Using the Narrative in Board Reports
The Narrative can be included in CMDash's board reporting tools:
- Board Deck (PPTX): includes an "AI Risk Narrative" slide with the top exposures and recommended actions.
- Custom reports (Report Builder): you can add an "AI Narrative" section to a custom report.
Both always pull the most recently generated Narrative at export time — there isn't a way to pick an older Narrative from a prior day for a report that covers an earlier date.
To generate a board report: Dashboard → Board Report → Export PDF or Export Deck.
Accuracy and Limitations
The AI Risk Narrative is only as current and accurate as your connected integrations.
- Stale integrations: If an integration hasn't synced recently, the Narrative may reflect outdated data. Check Integrations → Sync Status if a Narrative seems out of date.
- Coverage gaps: The Narrative can only analyze what CMDash can see. Assets, systems, or users not covered by any connected integration won't appear in it.
- AI limitations: The Narrative is generated by an AI model and may occasionally mischaracterize the severity or cause of an exposure. Always verify Narrative recommendations against the source data in your integrations before acting. Treat the Narrative as a starting point for investigation, not a definitive diagnosis.
Privacy
Narrative generation uses Anthropic Claude. The following applies to what's sent to it:
- What's included: your organization name, and business-context data relevant to producing the Narrative — posture score and signal breakdown, compliance framework names and progress, connected integration providers and health, and your highest-scored open risks' titles. This is real tenant business data, not fully anonymized metrics.
- What's excluded: no employee names, email addresses, or other end-user PII are included in Narrative generation inputs.
- No training on customer data: CMDash relies on Anthropic's standard commercial API terms, which exclude customer data from model training. CMDash does not have a separate, CMDash-specific zero-retention agreement with Anthropic at this time.
- Data processed within CMDash's infrastructure: the aggregation step described above occurs within CMDash's cloud environment before any data is sent to the AI model.
Anthropic is listed as a subprocessor at /trust/subprocessors. For questions about data handling, refer to CMDash's Privacy Policy or contact your account manager.
Common Questions
Can I turn off the Narrative?
You can hide the card from your own Dashboard at any time (see "Customizing the Narrative" above) — this doesn't require any particular role. There is currently no organization-wide setting to stop the Narrative from being generated on its daily schedule short of downgrading below the Command tier; hiding the card only affects what your own Dashboard displays.
Each daily Narrative uses one of your plan's monthly AI runs (see Billing). Because nobody starts it, it is automatic AI: it runs only while less than half of the month's runs are used, so it never spends the runs your people need. Past that, the daily Narrative is skipped until the allowance resets on the 1st (UTC); the most recent Narrative stays on your Dashboard.
How is the Narrative different from Insights cards?
Insights cards (on the Dashboard) are static, metric-based widgets — they display a specific number or chart without interpretation. The AI Narrative synthesizes data across multiple signals, identifies the most significant exposures, explains the trend, and recommends specific actions. The Narrative is interpretive; Insights cards are informational.
Can I share the Narrative externally?
There is no direct share link or copy-to-clipboard action on the Narrative card. For external sharing, use the board report export (PPTX or a custom PDF via Report Builder), which packages the Narrative alongside your other posture metrics.