The incident lifecycle
One signal is the start, not the verdict.
Your team controls the review and response. Tuteliq brings the signals, queue and decision history into one workspace.
- 01 Detect Integrate Tuteliq's proprietary behavioural safety models through the API or SDKs. Detection returns safety signals to inform review, not a final legal or moderation verdict.
Detect
Integrate Tuteliq's proprietary behavioural safety models through the API or SDKs. Detection returns safety signals to inform review, not a final legal or moderation verdict.
- 02 Record Use the built-in incident store to bring flagged interactions into the moderation dashboard. Set up your customer-held private encryption key when creating a Tuteliq API key; incident content is end-to-end encrypted.
Record
Use the built-in incident store to bring flagged interactions into the moderation dashboard. Set up your customer-held private encryption key when creating a Tuteliq API key; incident content is end-to-end encrypted.
- 03 Triage Filter the incident queue by status, category and severity. Inspect the available context and distinguish test runs from real incidents before prioritising your team's work.
Triage
Filter the incident queue by status, category and severity. Inspect the available context and distinguish test runs from real incidents before prioritising your team's work.
- 04 Review An authorised reviewer inspects the incident and decides whether to confirm, dismiss or escalate it. Difficult conversations need contextual judgement; a detection signal alone does not establish abuse.
Review
An authorised reviewer inspects the incident and decides whether to confirm, dismiss or escalate it. Difficult conversations need contextual judgement; a detection signal alone does not establish abuse.
- 05 Act and escalate Record the review decision and, where relevant, a structured platform action. Your organisation decides any safeguarding referral or external report through its own authorised procedures.
Act and escalate
Record the review decision and, where relevant, a structured platform action. Your organisation decides any safeguarding referral or external report through its own authorised procedures.
- 06 Track and report Follow incident status and action history. Use reporting tools and aggregate Trust Center evidence where enabled. Public reporting does not expose private conversations; daily scores are not a live queue.
Track and report
Follow incident status and action history. Use reporting tools and aggregate Trust Center evidence where enabled. Public reporting does not expose private conversations; daily scores are not a live queue.
Workflow examples
Different teams. Clear responsibility.
Illustrative scenarios, not customer testimonials or claims of connected reporting services. Apply your organisation's policies and applicable law.
Platform safeguarding
Private custody conversations between separated guardians generate repeated alerts.
- Open the flagged incidents and review the context available to your authorised team.
- Distinguish difficult but non-abusive exchanges from threats, coercion or child-safety concerns. Dismiss false alarms and record appropriate decisions.
- Escalate genuine concerns through your safeguarding procedure. Keep test incidents out of operational scores.
Tuteliq supports review, not custody judgements or a determination that an offence occurred.
Hotline analyst workflow
A hotline team assesses a suspected child sexual abuse report received through its own intake system.
- Keep the intake reference in your own system and use authorised Tuteliq analysis to inform triage, where permitted by your procedures.
- An appropriately trained analyst reviews available signals and context, applies the hotline's assessment criteria and records the decision.
- If a referral is required, an authorised person submits it through the hotline's established reporting channel and keeps the referral record in that system.
Illustrative workflow, not a deployed hotline integration. IWF intelligence in detection does not automatically submit reports to IWF or any hotline.
Law-enforcement handover
A platform reviewer identifies a credible threat or suspected exploitation requiring specialist escalation.
- Escalate the incident to the organisation's designated safeguarding or legal lead and document the review decision.
- The lead determines urgency, reporting obligations and the authorised recipient under applicable law. Immediate danger uses the relevant emergency channel.
- Share only the necessary information through the recipient's approved secure channel, following your evidence-preservation and access procedures.
Illustrative handover, not automatic police or NCMEC reporting. Tuteliq does not determine criminal liability or guarantee evidential admissibility.
Private incidents. Public accountability.
Your private encryption key remains customer-held. Tuteliq cannot decrypt incident content or recover that key. Hosted inference processes raw content transiently and deletes it immediately after inference.
Trust Centers publish aggregate evidence where enabled, not incident content. Marked test incidents are excluded from operational safety scores. No new retention period or legal guarantee is implied by this overview.
Explore the platform
- Incident storage controls
- Agent-assisted moderation
- MCP tools and permissions
- For Trust & Safety teams
- Safety Classification methodology