AI-Assisted Incident Command Training for UK Fire and Rescue
WatchBrief is a digital training tool designed for station-based incident command development. It enables any watch officer to deliver a structured, standards-aligned scenario session to a whole watch — using a projected group display and their own phone as a private leader panel — without the preparation burden traditionally associated with scenario-based learning. Each session generates a unique, operationally realistic scenario grounded in NFCC National Operational Guidance, complete with narrated phase progressions, multiple-choice discussion questions, answer keys, and a facilitated debrief.
The tool is built to align directly with the competency framework for Incident Command at Level 1 and Level 2, the NFCC National Operational Guidance, JESIP principles, and radio procedures. It supports consistent, repeatable training delivery across a service — removing the dependency on any single individual's preparation time or scenario knowledge — and provides session leaders with immediate, NFCC-grounded answer keys they can use to lead structured post-phase discussions.
Station-based incident command training is one of the most valuable development activities available to watch-level firefighters and junior officers. Scenario-based exercises build the mental models that commanders rely on in genuine emergencies — but they are disproportionately dependent on the time and expertise of the officer leading them.
Building a credible scenario from scratch requires an officer to research the incident type, construct a realistic progression, prepare discussion questions, source answer keys grounded in NOG, and organise supporting materials such as approach maps or building information. This process can take hours. When the same officer leads sessions week after week, the watch encounters the same scenarios repeatedly — reducing engagement and limiting the breadth of learning.
The result is a gap between the intent of structured scenario training and its practical delivery: sessions are infrequent, quality is inconsistent across different session leaders, and the depth of NOG-grounded discussion is limited by what a single officer can carry in their head on a given shift.
WatchBrief is designed to close that gap: it handles scenario construction, answer key generation, and session delivery, so the session leader can focus on what only they can provide — experienced facilitation and operational context.
A WatchBrief session is designed to be delivered in a watch room or training room with a projected display. The session leader configures and runs the session from their phone or tablet while the watch engages with the group screen.
The session leader opens WatchBrief and selects their fire service, incident type (from dwelling fire to aircraft and hazardous materials incidents), terrain, time of day, weather conditions, and the IC competency level being targeted (Level 1 or Level 2). They also choose a session depth — Quick, Standard, or Deep — which sets how many phases the scenario runs to and roughly how long the session will take. Optional configuration allows the session leader to specify local risk features, geographic constraints, or a session theme. The tool then generates a complete, unique scenario — taking around two to three minutes — before presenting a QR code the watch can use to follow along.
Each scenario unfolds across a series of structured phases on the projected screen. Every phase carries a voice-narrated scene update, a discussion prompt, and a question — usually multiple-choice, but some phases use true/false or multi-select formats to suit the competency being tested. The number of phases is set by the session depth chosen at generation:
Rather than a fixed script, each session is assembled from a library of NOG-mapped phases, every one tied to a specific National Operational Guidance control measure. The phases follow the chronological arc of a real incident — opening, build, escalation, and debrief — with the escalation injects chosen to fit the configured incident type:
Standard and Deep sessions include a dedicated hazard-identification phase — a dynamic risk assessment exercise presented differently from the multiple-choice phases. The projected screen shows a developed image of the incident scene alongside a grid of hazard tiles. The watch is asked what hazards they can identify from the scene, and the session leader then reveals which tiles are genuine hazards. It mirrors the dynamic risk assessment an incident commander performs on arrival, and is grounded in the NOG control measure for risk assessment at an incident.
On the session leader's phone, a private leader panel displays the full answer key for the current phase, the reasoning behind the correct answer, and controls to advance through the session, reveal the correct answer to the group, or trigger the optional complication phase. The session leader never needs to prepare this content — it is generated alongside the scenario and is immediately available as they run the session.
After the group has responded to each question, the session leader taps to reveal the answer on the projected screen. The correct option is highlighted and a rationale panel appears below the question, explaining why the correct answer aligns with NFCC National Operational Guidance. This provides a consistent, authoritative basis for officer-led discussion regardless of the session leader's personal familiarity with the specific incident type.
Completed sessions are saved to a shared library accessible from the setup screen. Session leaders can reload a previous session to repeat it with a different group, or browse past scenarios to plan a training programme.
WatchBrief uses large language model AI to generate all scenario content dynamically — meaning no two sessions are identical. The AI does not retrieve pre-written scenarios from a database; it constructs each scenario from scratch based on the session leader's configuration, producing a unique incident with a realistic UK address, specific building type, operationally plausible conditions, and a consistent narrative across every phase.
The AI operates under strict instructional constraints. Every answer key and rationale it generates is required to be grounded in the relevant NFCC National Operational Guidance for the scenario's incident type. The system draws directly on NOG content when constructing discussion answers — not general knowledge — so the training material reflects the doctrine that UK fire services actually use.
The AI is instructed to produce UK-specific terminology exclusively. It uses "breathing apparatus," "incident commander," "make pumps," "cordon," and "fire appliance" — not American equivalents. All radio messages are structured to established UK FRS radio procedures, using HAULEM format for informative messages and the correct phonetic tactical mode prefix for Oscar (Offensive) and Delta (Defensive) mode declarations. Resource counts are validated against Pre-Determined Attendance figures for the configured incident type.
Each generation uses a variation seed to ensure diversity across sessions — the same incident type will produce a different address, building, initial conditions, complication, and set of decision points every time. This eliminates the scenario-fatigue problem that affects fixed scenario libraries. At the same time, the AI's operational constraints prevent unrealistic or unsafe scenario elements from being generated.
Every correct answer cites the relevant NOG principle. Watch members receive doctrine-accurate feedback, not opinion.
Each session generates a unique scenario. A watch can run monthly sessions without repetition.
The AI is constrained to produce accurate UK procedure. No American terminology, no improvised radio formats.
Each phase narration is read aloud by a synthesised voice on the projected display, making the session more engaging for a group setting and reducing the cognitive load on the session leader. The narration is generated as part of the scenario content and is consistent with the written scenario facts throughout the session.
Generated scenarios pass through an automated validation stage before being presented. This checks structural completeness — ensuring every phase contains a valid answer key, well-formed question options, and a clearly-identified correct answer for its question type — before the session becomes available. Sessions that do not pass validation are regenerated automatically.
WatchBrief is designed to support — not circumvent — the established competency and guidance frameworks that govern UK incident command training. The table below maps key tool features to their corresponding standard or framework reference.
| Standard / Framework | WatchBrief Feature | How it aligns |
|---|---|---|
| NFCC National Operational Guidance (NOG) | Answer keys & rationale generation | All answer keys and correct-answer rationale are produced with explicit reference to the relevant NOG section for the scenario's incident type. The system is instructed not to draw on general knowledge where NOG guidance exists. |
| JESIP Joint Decision Model (JDM) | Phase structure & multi-agency phases | The phased session structure mirrors the cyclical nature of the JDM — gathering information, assessing risk, identifying options, and taking action. Build phases such as Information Management and the dedicated Multi-Agency / JESIP phase (included at Deep depth) specifically target multi-agency coordination, inter-operability, and shared situational awareness consistent with JESIP principles. |
| Radio Procedures | All AI-generated radio messages | Mobilisation calls, initial informative messages, priority messages, and stop messages are generated in full compliance with established UK FRS radio procedures. HAULEM field structure is enforced, tactical mode declarations use the mandatory Oscar/Delta phonetic prefix with rationale, and callsign formats match standard service prefixes. |
| SFJ Awards / National Occupational Standards — Incident Command L1 & L2 | IC Level session parameter; discussion question targeting | The IC competency level is a required configuration field. The AI uses this to calibrate the difficulty and focus of discussion questions, targeting the specific command competencies and decision-making expected at the selected level. |
| NFCC Dynamic Risk Assessment (DRA) Guidance | DRA prompts embedded in scenario phases | Phase questions and answer keys in the First Attendance, Hazard Identification, and Tactical Mode phases specifically address tactical mode decisions, entry control, and DRA processes consistent with NFCC DRA guidance. |
| NFCC Learning from Experience Framework | Debrief phase summary | The closing Debrief phase generates a 200–300 word ideal actions narrative describing what an exemplary incident commander would have done at each decision point. This provides a structured basis for post-exercise reflection aligned to the LfE cycle of identify, analyse, disseminate, and implement. |
| NFCC Pre-Determined Attendance (PDA) / Mobilisation | Resource count validation in scenarios | Appliance counts, specialist resources, and mobilisation sequences in generated scenarios are cross-referenced against Pre-Determined Attendance figures for the selected incident type and service, ensuring resource picture accuracy throughout the session. |
WatchBrief removes the preparation burden — it does not remove the officer leading the session. The value of station-based scenario training comes from experienced facilitation: challenging the group's assumptions, drawing on personal operational knowledge, redirecting discussion when it drifts, and helping junior colleagues connect abstract doctrine to real incidents they have attended.
The tool provides the materials — scenario narrative, discussion questions, answer keys, rationale, and debrief summary — so that the session leader can dedicate all of their attention to facilitation rather than recall. An officer using WatchBrief can run a structured, NOG-grounded session even on an incident type they have limited personal operational experience of, because the answer keys provide reliable doctrine-aligned guidance for discussion.
The session leader retains full control of session pacing. They choose when to advance each phase, whether to trigger the complication, when to reveal the answer, and how long to spend on post-phase discussion. The session is a tool in the officer's hands — not an automated replacement for their judgement.
WatchBrief is offered on a simple, transparent basis: a fixed pilot fee to prove the value at a small number of stations, a low per-scenario generation cost, and an annual platform agreement for full-service access. All prices exclude VAT.
WatchBrief is available as part of the CommandGridTools suite of specialist tools for UK Fire and Rescue training. All tools are designed and maintained by a practitioner with direct experience of UK FRS incident command training.
To access WatchBrief or learn more about the CommandGridTools suite, visit the address below.