Armoury is the digital checklist platform SCDF (Singapore Civil Defence Force) uses torun daily equipment readiness checks across all 23 fire stations. Firefighters, ambulancecrews and station commanders use it at the start of every shift to confirm that vehicles,medical equipment and protective gear are present and working.

This quarter we fixed the workflow problems frontliners had been raising for two quarters,and onboarded the last daily check that wasn’t yet tracked in Armoury.

• The Problem: Armoury accepted only one photo per checklist question. SCDF’s EmergencyMedical Services Department requires photo evidence for stretchers and drug equipment, soambulance crews had to manually combine several photos into a single image before theycould upload it. Every shift, every time.
• The Fix: The photo limit per question is now 6. Each photo uploads individually, insequence. No combining step.
• The Result: Satisfaction among EMTs (Emergency Medical Technicians, the ambulance crewrole most affected) recovered from 26.7% to 82.5% satisfied, with their average ratingmoving from 3.00 to 4.05 out of 5. The manual workaround is gone entirely. Paramedics atAlexandra Fire Station raised the speed of the fix unprompted during a site visit.

• The Problem: SCDF requires every firefighter to check their PPE (personal protectiveequipment: oxygen tanks, protective clothing) at the start of each shift. None of it wastracked in Armoury. Station commanders had no way to see who had completed the check,short of asking each person.
• The Fix: Daily PPE checks are now live at all 23 fire stations. Each checklist is builtaround a vehicle’s maximum manning, the full complement of personnel it can carry. Inpractice, not every seat is filled on every shift. Admins now configure maximum manningupfront, and frontliners mark all unfilled positions in a single tap instead of openingeach one individually.
• The Result: 12 taps down to 1 for handling unfilled positions. 247 new checklistslive across the force. Armoury is now the single platform for every daily shift readinesscheck at SCDF.

• The Problem: Frontliners run different types of checks at different times of day, butthe home screen showed all 50+ checklists at once with no filter. Finding the right 20meant scrolling past everything else, every single time. Raised as a friction point acrosstwo consecutive quarters.
• The Fix: Checks are now grouped by service type. The home screen shows each user onlythe checklists relevant to their role and assignment.
• The Result: 366 users took over 2,300 actions in the first week of release. Homescreen clutter is gone for all frontliners.

• The Problem: The PDF export listed every response in sequence, including every itemthat passed. Station commanders and HQ had to read the full submission to find the handfulof items that actually needed action.
• The Fix: Outstanding issues are now surfaced prominently alongside each response,visible at a glance.
• The Result: Issues can be acted on the moment the report is opened, rather than afterreading through the entire submission.

• The Problem: When connectivity or session problems occurred, frontliners saw a genericbrowser error page with no explanation of what had happened or what to do about it.
• The Fix: Plain-language error pages that say what went wrong and what to do next.
• The Result: Frontliners can resolve common issues themselves instead of raising asupport ticket.

• Security: An independent penetration test by external security firm Cure53 wascompleted. No critical findings. All items are closed: 1 High severity finding fixed, 5 of6 Medium findings fixed, and the remainder formally risk-accepted with documented rationale.
• Platform currency: Upgraded Node 20 → 22 and Next.js 14 → 16, clearing dependencies thatwere approaching end of life.
• Release quality: Automated integration and end-to-end tests now cover the critical userjourneys. An AI-assisted pipeline generates a test checklist scoped to what actually changedon each release, and a team member verifies it against staging before anything reachesproduction.

Built and tested through Q2, released on 6 July as the final feature change before Armourymoved into maintenance.

• The Problem: User satisfaction had fallen to 3.68 out of 5 in Q1, down from 4.3. Ausability audit confirmed what the score was signalling: navigation took too many steps,interaction patterns were inconsistent between screens, and a cluttered home screen made ithard to find the right check quickly. An independent usability benchmark (the SystemUsability Scale, scored out of 100) came in at 67, just under the 68 industry average, andpointed at the same problems.
• The Fix: A redesigned homepage, checklist view, history page, check-starting flow andchecks page. Larger yes/no buttons, chosen by user vote. Clearer grouping of related items.Outstanding issues shown before a check begins. A prompt to log an issue when a “no” answerhas none attached. Confetti on submission.
• The Result: Satisfaction reached 4.29 out of 5 (84.7% satisfied) across 937 SCDFrespondents, the highest in Armoury’s history and stable across two consecutivemeasurement pulls. Every role scored above the 70% industry benchmark. Logging an issueagainst a failed item is now the norm rather than an afterthought.

Armoury is now in maintenance. The platform keeps running, security issues get fixed, andsupport continues through each SCDF department’s designated contact. New feature developmenthas stopped.

This quarter, we gave SCDF’s senior leaders direct access to live station data for the first time — and resolved the four biggest pain points slowing down the admins who build and manage checklists.

• The Problem: Divisional Heads had no direct access to Armoury’s data. They had to request it manually, and decisions were made on monthly summaries — not live information.
• The Fix: A Division Dashboard embedded directly in the Armoury Admin Portal, accessible to all Divisional Heads and HQ departments.
• The Result: One division head used the dashboard within 24 hours to independently identify stations with missing submissions and investigate why — no prompting needed. Admin data usefulness jumped from 3.5 to 4.86 / 5.

• The Problem: Building a checklist with 100+ items required 6–8 clicks per item. Creating large checklists took hours of repetitive input.
• The Fix: Admins can now copy any item or page in 1–2 clicks. Paired with item name suggestions — Armoury surfaces names already used in other checklists, so admins can select a standardised name instead of typing it each time.
• The Result: 50% reduction in clicks per item created. Naming is now consistent across checklists, making cross-checklist auditing more accurate.

• The Problem: Assigning a checklist to a station required navigating away from the teams page to check what it already held — an 8-click process with no station context visible.
• The Fix: Checklist assignment and removal is now available directly from the teams page, with a full view of what the station currently holds.
• The Result: 8 assignments made in the first week, unprompted. Admins reported less mental load and fewer assignment errors from switching between pages.

• The Problem: Armoury’s search didn’t recognise section names — the headings that group items within a checklist. Admins had to scroll manually through entire templates.
• The Fix: Section names are now searchable. Admins can type any section name and jump directly to it within a checklist.
• The Result: Admins described the change as reducing mental load — surfacing exactly what they need instead of presenting an information dump.

• The Problem: Items couldn’t be reordered after creation. The only fix was to delete and rebuild — so most admins left templates out of order, creating discrepancies when cross-referencing against physical equipment.
• The Fix: A drag-and-drop interface lets admins reorder any item in 3 interactions: click, drag, drop.
• The Result: Checklist order now matches the physical layout of equipment during checks. Discrepancies that previously had to be mentally reconciled are eliminated.

• The Problem: Newer firefighters and EMTs often had to guess what correct equipment storage looked like, especially for unfamiliar appliances.
• The Fix: Photos can now be attached to any checklist item. Frontliners tap the item to see a reference image during a check.
• The Result: Reduced onboarding friction for new personnel. Users noted it helps newer EMTs confirm exactly what they’re supposed to look for.

• The Problem: Station commanders spent 1.5 days every month manually downloading reports for vehicle audits.
• The Fix: A “One-Click” export that bundles a month’s worth of data into a single file.
• The Result: Audit preparation time crashed from 24+ hours to under 5 minutes.

• The Problem: Headquarters had plenty of data but no “bird’s-eye view” to tell them which stations were actually ready for operations.
• The Fix: We iterated through three versions of data visualization to create a “Red/Green” status dashboard.
• The Result: Satisfaction from Divisional Leads jumped from 2.5 to 4.0/5 as they can now identify and fix readiness gaps instantly.

• The Problem: Armoury tracks if gear is working, but not how many are in stock. Logistics needed both to make buying decisions.
• The Fix: Instead of rebuilding an inventory system, we partnered with the SCDF’s TFTD Data Branch to merge Armoury’s “usage data” with existing inventory systems.
• The Result: A holistic view for HQ that shows both operational readiness and supply levels in one place.

• Top navigation & progress bar with upfront Issues/RemindersWhy: cut scrolling and page‑hopping.What changed: quicker locker switching; early data shows −18% on PL Gen 5 in median time taken to submit the checklist.

• Condensed View (denser cards + in‑place actions)Why: reduce taps/scrolls without rubber‑stamping.What changed: 10–30 min faster on long lists in pilots

• Checklist Transfer for Parade TasksWhy: onboard all daily vehicle checks with full history intact.What changed: 730+ checklists onboarded; coverage to 100%

• EMS daily checklist simplified (~300 → <40 items) —Why: lower cognitive load; focus on critical daily items.What changed: faster checks; satisfaction ~4.0/5 (prev ~2/5); depth moved to weekly.

• Live edits to published templates (versioned, audit‑logged)Why: avoid duplicate + reassign churn.What changed: Admins can update live templates safely with no downtime or data loss.

• AM/PM/Night shift windows for HealthcareWhy: align checks to nursing handovers.What changed: 204 nurses onboarded; shift‑level reminders & reporting.

• Made Armoury much faster at the moments users notice most — opening the homepage, opening a checklist, and downloading a report.Why: Reduce waiting and prepare for more teams and vehicles.What changed: ~2–10× faster overall (homepage now ~0.7s P50; checklist page ~0.4–1.0s; report download ~0.4–2.0s).

• Reports now keep the exact question wording you answered.Why: Audits need historical accuracy; edited questions shouldn’t rewrite history.What changed: Reports always show the original phrasing from time of submission.

• Report downloads are “snapshotted” after you submit.Why: Crews often re‑download the same report (avg ~6×); we also want to enable bulk export later.What changed: Downloads are quick and consistent because we serve a saved copy in our storage compared to generating on demand.

• Clear, in‑the‑moment error feedback during checks (with a forced refresh for breaking updates).Why: Bad network could silently fail saves; users only found out at submission.What changed: A banner tells you immediately, the UI stays truthful, and old app versions don’t linger → fewer submission errors.

• Mobile‑friendly invite links (2 taps to share).Why: Admins often onboard people mid‑shift; desktop steps were slow and error‑prone.What changed: Fewer invalid‑code logins and “send me today’s code” requests; sharing now takes seconds on a phone.

• Undo a skipped checklist (when enabled by admins).Why: Ground reality changes (e.g., stand‑by vehicles); accidental skips happen.What changed: Users can recover a skipped checklist themselves; no backend ticket needed.

• Launched the Open Checks module for admins and users, enabling checks to be started and submitted anytime, with no fixed end time. This was designed for checks that don’t need to track compliance levels, giving teams flexibility for ad-hoc or ongoing tasks.
  • Enhanced the Past Checks view to support Open Checks, so users can easily track and access historical submissions alongside scheduled checks.
• Introduced Draft Mode for checklist creation, empowering admins to build, save, and refine checklists before publishing. No more needing to finish everything in one sitting—admins can work iteratively to improve checklists. Happy Admins, Happy Users!
• Expanded Audit Logging to capture more detailed activity—now you’ll know who clicked what, and when, offering better accountability and transparency across check processes.
• Automated Invite Code Management with a new cronjob, streamlining code generation and lifecycle management for onboarding new users more efficiently.
• Rolled out the Unskip Check feature in response to user feedback—giving admins the flexibility to undo accidental skips or re-activate last-minute checks without hassle.
• Upgraded critical dependencies (React Query, Next.js, Prisma, Iron-Session) to ensure better security, improved performance, and long-term platform stability.
• Migrated our release process to semantic-release, automating versioning and streamlining deployments to reduce manual errors and speed up delivery.
• Boosted backend resilience with improvements like database indexing, smarter queries, and retry mechanisms, ensuring the platform stays robust under heavy use.
• Improved error logging and alerting so our team can detect, diagnose, and fix backend issues faster, reducing potential disruptions for users.
• Delivered a range of UI refinements across admin dashboards and user pages—fixing bugs, polishing interactions, and smoothing out workflows for a better overall user experience.

• Introduced scoring as a feature with user and admin flows, enabling comprehensive assessment of operational readiness.
• Released new input types, including Yes/No, Number, Dropdown, Textbox, Photo, and Date & Time, enhancing checklist flexibility and onboarding of complex workflows.
• Facilitated weekly EMS Hazmat and MRT Shelter checklist pilots, resulting in over 8,300 additional submissions.
• Enabled historical frequency support for weekly and twice-daily checks, expanding scheduling options to meet agency needs.
• Enhanced scheduling capabilities by allowing admins to pause, resume, or edit schedules after assignment, improving operational flexibility.
• Implemented real-time issue alerts via webhooks and Telegram, ensuring prompt action by Logs ICs on flagged issues.
• Enabled users to report and update issues outside of scheduled check windows, accelerating issue resolution and reducing downtime.
• Piloted dashboards highlighting appliances with <80% availability scores, improving visibility of critical issues for follow-up actions.
• Replaced static PDF reports with interactive dashboards, enabling admins to monitor readiness in real-time.
• Revamped the admin portal checklist builder with consistent component patterns, improved hierarchy, and fixed crashes for a smoother checklist creation experience.
• Addressed search functionality issues, ensuring users are directed to relevant questions during checklist searches.
• Standardized the color palette using Tokens Studio, enhancing UI consistency and user navigation.
• Conducted design workshops with FRS and Hazmat teams to identify onboarding pain points and inform future product improvements.
• Supported SCDF’s Automated Duty Officer Watch Room research and EMS pilots, reinforcing product adaptability for diverse use cases.
• Aligned product enhancements with the 2025 vision, including foundational work for onboarding CERT and Enforcement checklists.

• Increased compliance rate from 87% in June 2024 to 94% in September 2024, demonstrating enhanced adherence to operational standards.
• Facilitated a design workshop for Emergency Medical Services (EMS) to develop a new Table of Equipment for daily and weekly checks, improving checklist relevance and usability.
• Optimised issue reporting and management, enabling LogsICs to report and manage issues regardless of checklist status, resulting in faster issue resolution.
• Enhanced product polish by addressing UI janks, usability issues, and product gaps, leading to a more seamless and user-friendly experience.
• Implemented webhook support to notify LogsICs of new issue reports in real-time, ensuring timely awareness and action.
• Developed submission score metrics to track the operational readiness rate of each submission, providing actionable insights for continuous improvement.
• Standardized the color palette using Tokens Studio ensuring consistent visual design across the platform for better user navigation and aesthetics.
• Initiated efforts to generalize Armoury by building new input fields, including Yes/No questions, Photo uploads, and Dropdown selections, enhancing checklist flexibility.
• Launched Pulse Item Transaction Download allowing StoreICs to download item transactions for better tracking and auditing of inventory movements.
• Updated Itemsets behavior enhancing the management and categorization of items within checklists for improved efficiency.
• Introduced Standalone Photo Input Field enabling users to upload photos independently of other checklist input for better data accuracy and documentation.
• ImplementedPulse Last StockTake Date providing users with timely insights into inventory status and aiding in the planning of future stocktakes.
• Expanded Pulse History View to all users enhancing transparency and enabling better monitoring of store activities.

• Onboarded 51 ambulances for daily HandOverTakeOver (HOTO) Checks, enhancing operational readiness and compliance.
• Increased compliance rate of submissions from 79% in April to 87% in June.
• Introduced frequency checks to verify if officers have conducted their checks within their shifts.
• Added a “Skip Check” function to allow dual fleet stations to indicate which callsigns are not in use during a shift.
• Enhanced PDF reports to display the photos submitted as part of the checks.
• Launched a centralized issue tracker tool for Station Team Admins to monitor and address team issues.
• Extended report access from 30 days to 6 months, providing users with historical submission data.
• Introduced a “Past Check” feature to view historical submissions and identify non-compliant appliances.
• Added an email reminder feature to notify Team Admins 2 hours before checks are due.
• Facilitated discussions to improve compliance policies, aligning HQ and station expectations.
• Added a submission overview feature for users to see pending checklists, ensuring all submissions are completed.

• Onboarded 56 Ambulances for daily HandOverTakeOver (HOTO) Checks, we streamlined and digitized the checklist process, ensuring prompt and accurate assessments, thereby enhancing operational readiness and compliance.
• The release of Pulse for Marina Bay Fire Station Medical Store enabled Medical StoreICs to efficiently monitor inventory levels, track expiration dates, and manage inventory withdrawals in real-time, fostering better decision-making and adherence to quality standards.
• The introduction of search functionality, transaction, stocktake, withdrawal, and team management features for Pulse enhanced the platform’s capabilities, facilitating smoother inventory management processes, and empowering teams to make informed decisions swiftly.
• User interface and experience improvements, such as adding tab views for Status and About in the item drawer and implementing padding and text styling to match the design, enhanced user interaction, making the system more intuitive and user-friendly.

• We scaled Armoury to all 23 Physical Fire Stations and 3 Special Units (Dart, SRU, K9), extending the reach of our platform, ensuring comprehensive coverage and standardized processes across all units, thereby enhancing operational efficiency and compliance.
• The introduction of Interim Submission offered users the flexibility to submit interim reports, enabling timely updates and better monitoring of ongoing activities, ultimately facilitating more informed decision-making.
• The revamped Issue Reporting flow enhanced the efficiency and accuracy of issue identification and resolution, promoting proactive management of potential challenges and ensuring smoother operations.
• The addition of the ItemSet Select All feature streamlined the process of selecting multiple items, saving time and effort for users conducting checks or managing inventory.
• The ability to download reports provided users with convenient access to essential data and insights, empowering them to analyze trends, track performance, and make informed decisions.
• The Admin Portal, featuring functionalities such as Daily Submission tracking, Equipment Readiness Dashboard, and Team & User Management, offered administrators comprehensive tools to monitor activities, manage resources effectively, and ensure operational readiness.
• User interface and experience improvements across login screens, submission screens, and error screens enhanced usability and accessibility, ensuring a seamless and intuitive user experience throughout the platform.

• Revamped Armoury into a web application brought increased accessibility and flexibility to users, allowing them to access the platform from any device with internet connectivity, thereby improving efficiency and convenience in checklist management.
• The trial with Marina Bay Fire Station served as a pilot program to test the effectiveness and usability of Armoury in a real-world setting, gathering valuable feedback and insights to further refine the platform and ensure its suitability for broader deployment.
• Scaling Armoury to selected Fire Stations extended the platform’s reach and impact, enabling more units to benefit from its features and capabilities, thereby fostering standardization, efficiency, and compliance across multiple locations.