Incident response
Version 1.1 · October 2026
This document describes how Solicia detects, responds to and communicates information security incidents. It is the customer-facing summary of our internal incident response runbook, which contains infrastructure-specific procedures and commands and is retained internally. We are happy to walk a customer's security contact through any part of that runbook on request.
1. Detection
Detection is automated.
- Alert rules evaluate the production log stream continuously and notify us by email. Critical conditions are evaluated every five minutes. Notification delivery has been tested. Current rules cover: attempted authentication bypass; sustained failed authentication beyond the rate limit; requests attempting to reach our origin servers directly rather than through the protected edge; failure to record an audit event; any production security control failing open; and abnormal application error rates. - Cloud-native threat detection runs continuously across both infrastructure providers: Microsoft Defender for Cloud and AWS GuardDuty. - Application audit trail. All access to customer data is logged at the application layer across more than sixty event types, including authentication, administrative actions, document delivery, data purges, and any cross-tenant access. Records are tenant-scoped and exportable. - Customer reports. Customers can report a suspected incident to admin@solicia.ai, acknowledged the same business day.
2. Severity classification
| Severity | Definition |
|---|---|
| Sev-0 | Customer data confirmed exposed outside its tenant boundary or to a third party; or the service unavailable with no recovery path underway |
| Sev-1 | Credible suspicion of data exposure, contained; a compromised credential; service unavailable with recovery underway |
| Sev-2 | Degraded service with no confidentiality impact |
| Sev-3 | Minor defect with no customer impact |
Where classification is uncertain we classify higher and downgrade later.
3. Response sequence
- Contain before diagnosing for any Sev-0 or Sev-1. We accept a longer outage in preference to continued exposure. Containment options include invalidating all active sessions, isolating a workload, and suspending background processing. Inbound email queues at our mail provider during isolation and is not lost.
- Preserve evidence before remediating. Logs and audit records are captured first, because redeploying destroys container state.
- Establish blast radius: which customers, which records, and whether data left the system boundary.
- Notify affected customers (see section 4).
- Remediate and verify the root cause.
- Post-incident review in writing.
We maintain specific playbooks for suspected cross-tenant data exposure, compromised administrative credentials, compromised workloads, model output failures, provider outages, urgent customer security requests, and law enforcement or subpoena requests.
4. Notification commitments
| Audience | Trigger | Commitment |
|---|---|---|
| Affected customers | Any incident that may affect the security of their data, including suspected incidents | Within 24 hours of becoming aware |
| Affected customers | Written post-incident review for any Sev-0 or Sev-1 | Within 10 business days |
| OAIC | Eligible data breach under Privacy Act 1988 (Cth) §26WE | As soon as practicable, and within 30 days of awareness |
| Affected individuals | Eligible data breach | With the OAIC notification |
Our 24-hour commitment is deliberately tighter than the Privacy Act requires, and it applies to suspected as well as confirmed incidents. We do not wait for certainty before telling a customer.
Notification includes the known facts, the categories and volume of data affected, which of the customer's users or matters are involved, the timeline, containment and remediation actions taken, and our assessment of likely impact. We state explicitly what remains unknown. We do not withhold detail on grounds of commercial sensitivity.
5. Customer access to evidence
On request during or after an incident we provide the relevant application audit records and logs so the customer can conduct their own investigation, and sufficient information to respond to the Privacy Commissioner or another regulator. We will engage directly with a regulator if asked to.
6. Responsibilities
Solicia is responsible for detection, containment, investigation, remediation, and recovery of the service and customer data. The customer is responsible for actions within their own systems and for any onward notification to their own clients.
7. Law enforcement and legal requests
We have never received a government or law enforcement request for customer information. If we did: we acknowledge receipt only, take legal advice before responding on substance, preserve the data requested, narrow the scope where it is overbroad, produce only what we are legally obliged to, and log exactly what was produced. We notify the affected customer unless legally prohibited, and as soon as any prohibition lapses.
8. Testing
Restore capability has been verified in practice: on 25 July 2026 the complete production database was restored into new infrastructure with automated verification of schema and record counts, followed by a confirmed live transaction.