Communication Policy
Purpose
This policy sets how Rygen Technologies communicates about AI systems and the AI Management System (AIMS) with internal and external stakeholders.
Scope
This policy applies to all AI-related communications, including new features, incidents, compliance requirements, and stakeholder feedback, across all Rygen platforms and internal operations.
Communication Framework
For each stakeholder group we communicate with, this section sets what we communicate about AI, when, and how.
Internal Communications
CEO & Board
- AI strategic initiatives and certification status: as needed, via executive briefings
- Major AI investments: before launch, via executive briefings
- Significant risks or incidents: immediately by email
CTO
- AIMS performance metrics: every four months in the management review report (AI-012), and as needed in 1:1 meetings
- AI system health: every four months in the management review report, and immediately in Slack for incidents
- Risk status: every four months in the management review report, and immediately in Slack for urgent changes
- Development progress: as needed, via 1:1 meetings and Jira
- Resource needs: as needed in 1:1 meetings, and every four months in the management review report
Engineering Teams
- AI development standards: immediately on process changes, via team meetings, the AIMS Handbook, and Slack
- New AI tools and capabilities: as available, via team meetings and tech talks
- Process updates: immediately on process changes, via team meetings and the AIMS Handbook
All Staff
- AI Policy updates and compliance requirements: when approved, via Slack and the AIMS Handbook
- New AI features launched: at major launches, via Slack announcements
- Available training and lessons learned: as available, via Leapsome
External Communications
Clients
- New AI features: at launch, via release notes and email
- Known limitations: at launch, via release notes and help documentation
- Incidents affecting service: the client’s account manager notifies the client by email
- Feedback requests: annual satisfaction surveys by email and on customer success calls
End Users
- AI feature availability: at launch, via in-app notifications and help documentation
- How to use AI features: at launch and on documentation updates, via help documentation, training videos, and the support portal
- How to report issues: at launch and on documentation updates, via in-app notifications and the support portal
Auditors/Compliance
- AIMS documentation: on request and for scheduled audits, via secure document sharing and formal reports
- Audit evidence: on request and for scheduled audits, via secure document sharing and audit meetings
- Incident reports: on request and at annual reviews, via formal reports and audit meetings
- Compliance status: at annual reviews and scheduled audits, via formal reports and audit meetings
Model/API Providers
- Usage compliance: per contract terms, via vendor portals and email
- Incident reports: per SLA, via vendor portals, email, and support tickets
- Integration issues: as needed, via support tickets and email
Communication Triggers and Protocols
Mandatory Communications
New AI System/Feature Launch
- Internal announcement: in #releases when the release is requested and completed, and in the internal release notes
- Client notification: at launch, in release notes and the user guide
- Documentation update: before launch
- User guidance: available at launch in the help documentation and release notes
AI Incidents
AI-023 Incident Response Procedure defines severity classification, the operational response, and AIMS integration. Incident communication follows Rygen’s Engineering Incident Response Procedure, including its external-notification timing. The table identifies the AIMS recipients and channels.
| Severity | Internal Notification | Client Notification | End User Notification |
|---|---|---|---|
| SEV1 (Critical) | CEO: Immediate via email; CTO: Immediate via Slack | By the account manager by email | Via Beamer or the support channel |
| SEV2 (High) | CTO: Immediate via Slack | By the account manager by email | Via Beamer or the support channel (if user-facing) |
| SEV3 (Medium) | CTO: Next business day via Slack | As needed based on impact | As needed via help documentation or support portal |
Post-incident reports follow AI-023, Post-Incident Activities.
Policy/Process Changes
- Internal notification: in #ai-governance when the change is approved
- Training, if needed: before the effective date
- Acknowledgment: at hire and annually in Leapsome; contractors on the Microsoft 365 acknowledgment form
Compliance Updates
- Certification achieved: announced to all staff and publicly on the website and LinkedIn
- Audit findings: as AI-015 sets in Report Distribution and Appendix A.3, Escalation Criteria and Re-Audit Requirements
- Required actions: affected teams immediately
Regular Communications
Every four months
- Management review report to top management (AI-012)
Annually
- Compliance training notification
- Stakeholder feedback request
Feedback Channels
Internal Channels
- Slack #ai-governance: process improvements to AI Governance Committee members, as needed
External Channels
- ai-admin@rygen.com: general AI feedback, read by the Principal AI Engineer and the Information Security Officer, answered within 2 business days
- Support tickets: AI feature issues to the support team, per SLA
Responsibilities
- Principal AI Engineer: overall policy implementation and technical communications
- Product Managers: client feature communications
- Marketing: external announcements and newsletters
- Support Team: incident communications and user guidance
- Team Leads: cascade information to their teams and confirm understanding
- Communications Lead (incident role): manages incident communications under this policy during active incidents
Policy Review
This policy is reviewed annually or when stakeholder needs or communication requirements change significantly.
Annex A: #ai-governance Channel Protocol
Purpose
The #ai-governance Slack channel is an auditable AIMS communication log and one part of the AIMS record that an auditor may review. It is a chronological record of governance information, deliberation, decisions, and questions about Rygen’s AI systems and the AIMS. The channel records the discussion and the act of deciding. Material decisions are documented in the appropriate artifact, such as a GitLab merge request, decision record, or meeting minutes, which a post may link to. The channel is reserved for AIMS governance, and posts are treated as records.
Posting Format (BLUF)
Every top-level post is one of four types and uses Bottom-Line-Up-Front (BLUF) structure:
- Heading line:
TYPE - <subject>, for exampleINFO - Update to Communication Policy. - Bottom line: one or two sentences stating the essential point first.
- Background: two to six bullet points giving the context needed to understand or act on the post.
- Recommendation: for
DECISIONposts only, the proposed course of action.
Message Types
| Type | Use when | Required elements |
|---|---|---|
| INFO | Sharing something the team should know; no response expected. | Heading, bottom line, background |
| DISCUSSION | Opening a topic for input where no single decision is yet framed. | Heading, bottom line, background |
| DECISION | Proposing or recording a governance decision. | Heading, bottom line, background, recommendation |
| QUESTION | Requesting a specific answer from the team or a named person. | Heading, bottom line (the question), background |
Threading
- Each post is a self-contained top-level message, so the channel reads as a chronological log of distinct items.
- All replies, follow-ups, and deliberation go in the post’s thread, never as new top-level messages.
- For
DISCUSSION,DECISION, andQUESTIONposts, the original poster should open the thread, for example with the first reply, to keep related discussion in one place.
Examples
INFO - Communication Policy updated to v1.2
AI-007 Communication Policy has been updated to v1.2, adding Annex A (this channel protocol). See <link>.
Background:
- Defines #ai-governance as the auditable communication log for the AIMS.
- Establishes the BLUF posting format and four message types.
- No change to existing stakeholder or incident communication requirements.
DECISION - Review the Risk Dashboard at the management review
Proposing the Risk Dashboard be reviewed at each management review rather than separately. Seeking agreement by <date>.
Background:
- The management review already reviews the risk register (AI-008, AI-012).
- A separate review repeats that work.
Recommendation: Review the Risk Dashboard at each management review, retaining ad-hoc review on any material change in risk.
Revision History
| Version | Date | Author | Summary of Change |
|---|---|---|---|
| 1.0 | 2025-06-05 | Field Bradley | Initial draft. |
| 1.1 | 2025-09-02 | Field Bradley | Migrated to markdown and gitlab |
| 1.2 | 2026-06-05 | Field Bradley | Added Annex A: #ai-governance channel protocol (BLUF format, message types, threading). |
| 1.3 | 2026-07-14 | Field Bradley | Updated illustrative Risk Dashboard decision example to the four-monthly cadence and Jan/May/Sep risk-review dates for consistency (AI-1759) |
| 1.4 | 2026-11-04 | Field Bradley | Removed the monthly AIMS status report and monthly performance metrics, which are not done; CTO updates come from the management review report, Slack, 1:1 meetings, and Jira. Process improvements are raised in #ai-governance, as the committee holds no regular meetings (AI-2173) The illustrative DECISION post no longer refers to monthly reviews. |
| 1.5 | 2026-11-09 | Field Bradley | Shortened the policy, cites AI-015 and AI-023 for reporting timelines, and removes communication commitments that are not carried out (AI-2173, AI-2131). States the channels and timing used for launches, policy changes, acknowledgments and certification, and removes communications that are not made. Incident communication follows the Engineering Incident Response Procedure; client and end-user cells, user guidance at launch and the Slack channel reference updated. AI team communications use the AIMS Handbook in place of Confluence. |