Skip to main content
When you record an incident in Signal, AI works alongside you to ensure the record is complete and correctly classified. It analyses your description and suggests categories, locations, and events — saving time and helping you capture everything that matters.
The AI Recommendations panel in the incident form, showing suggested next steps you can edit or remove

AI suggestions when creating incidents

When you have written the description and click Next, Signal analyses what you have written and suggests:
  • Categories — the types of concern that best match what you have described, such as Physical, Emotional, Neglect, or Online Safety.
  • Location — the place the concern relates to, based on your description.
  • Events — relevant events to record alongside the incident, such as external agency involvement (for example Police or CAMHS).
You review each suggestion and choose whether to accept, adjust, or dismiss it. The AI helps ensure nothing is missed, but you always have the final say.

Short term actions

Signal can also draft suggested next steps — the actions a member of staff should take to follow the concern up. You can generate these while you are creating the incident, without saving first. On the review step, click Run Analysis in the AI Recommendations panel. Signal drafts a list of suggested actions from three sources:
  • Any active action plan for the student. Where a plan exists it takes precedence over policy, so the suggested actions stay consistent with what has already been agreed for that child.
  • Your school’s own uploaded policies, used as supplementary guidance or where there is no action plan — so suggestions follow your agreed procedure rather than generic advice.
  • The student’s own incident history, so a repeat pattern is taken into account.
Suggestions are deliberately limited to immediate, short-term actions. Signal will not propose creating an action plan or a long-term strategy — those are decisions for the safeguarding team. Each suggestion is fully editable before you save:
  • Reword any suggestion, or @mention a colleague in it so they are notified.
  • Delete any suggestion you do not want with the bin icon, and restore it if you change your mind.
  • Anything you keep is saved as a comment on the incident, so the actions become part of the record.
Running the analysis is optional and never automatic — nothing is generated until you click Run Analysis, and nothing is saved until you save the incident.
You can also do this later on a saved incident, using Create Short Term Actions on the incident card. It works the same way, and you click Save as Comments to keep the results.

Policies matched after saving

Policies are the one part that happens after you save. Once the incident is stored, Signal matches it against your school’s safeguarding policy documents and attaches the relevant ones, so staff have the right guidance to hand. This runs in the background and takes a few moments.
The review step showing the AI-suggested category as a chip you can accept or change, with the suggested location and Family Incident setting beside it
Write detailed, descriptive accounts of what happened. The more context you provide, the more accurate the AI suggestions will be.

AI Summary

Every incident gets an AI Summary — a short summary of that one incident, so you can grasp what happened without reading the full record. The student’s previous incidents are used as background context when the summary is written. That means the summary can point out how this incident sits against what has gone before — for example, that it repeats a concern raised earlier in the term. It is a summary of the single incident, not a summary of the student’s whole history.

What a summary includes

  • What happened — a concise account of the concern recorded on this incident.
  • Patterns worth noticing — where this incident echoes or escalates earlier concerns for the same student.

Where summaries appear

The AI Summary appears on the incident itself, so you see it wherever you open an incident — from a student profile, an incident list, or a search result. This is especially useful when:
  • You are triaging several new incidents and need the gist of each one quickly.
  • A new DSL is taking over and needs to get up to speed on a case.
  • You want to know whether a concern is a one-off or part of a repeating pattern.
An incident showing its AI Summary

Document analysis within incidents

When you attach files to an incident — photos, scanned letters, reports — Signal’s AI automatically analyses them. It extracts text and generates a summary, so you can understand the contents without opening every attachment individually. Learn more about how this works in Document Analysis.

AI is a tool, not a decision-maker

All AI suggestions are recommendations. They are designed to support your professional judgement, not replace it. You review every suggestion and make the final decision on categories, locations, events, and next steps.