Acting on Your Insights — Manager Decision Guide

This guide is for Super Admins and shift managers who review performance data in GC Surge. It translates what you see in the dashboards into specific actions. For metric definitions, see Dashboard KPI Reference. Covers: APT Issues, Filter Performance, Staffing and Coverage, Alarm Volume Patterns.

APT Issues

APT is high across the whole team

What you see: Average processing time is above 90 seconds for most or all operators.

What to do:

1. Check alarm volume in Analytics — if daily alarms have increased recently, the team may be understaffed for the current load. Review staffing against the Real Alarm Density by Hour chart to identify peak windows.
2. Check the Filter Ratio in the Alarm Center Admin View. If NOVA99x is filtering less than 85%, more false alarms are reaching operators than expected. See NOVA99x AI Filtering for tuning steps.
3. Check if operators are using ZenMode. APT is only recorded for alarms handled through ZenMode. If operators are closing alarms elsewhere, the metric will not reflect real performance.

One operator has significantly higher APT than peers

What you see: A single operator's Avg. Processing time in the leaderboard is 2x or more above the team average.

What to do:

1. Check their Picked Sites count — if they are consistently on high-alarm-volume sites while peers handle lighter ones, rebalance site assignments.
2. Check their Responsive Rate — a low rate combined with high APT suggests they are overwhelmed at peaks, not slow at processing. Adjust their shift coverage.
3. If their Responsive Rate is normal but APT is high, they may be over-investigating before closing. Review their closure tag distribution to see if they are consistently marking ambiguous alarms.
4. For new operators, high APT in the first week is expected. Review week-over-week to confirm it is trending down.

APT has not improved after enabling NOVA99x

What you see: NOVA99x is active but the team's average APT has not dropped.

What to do:

1. Confirm the Filter Ratio in Analytics is above 85%. If it is low, NOVA99x is not yet tuned correctly for your camera types — see NOVA99x AI Filtering.
2. Compare APT before and after NOVA99x using the Last 30 days view in Analytics. If APT is unchanged, check whether operators are handling fewer alarms per shift (the load should have dropped) — stable APT on fewer alarms is still an improvement.
3. Verify ZenMode is being used for all closures. NOVA99x and ZenMode work together — filtering alone reduces volume, but ZenMode is what reduces per-alarm handling time.

Filter Performance

Filter Ratio dropped this week

What you see: NOVA99x is filtering a lower percentage of alarms than usual.

What to do:

1. Check if new cameras or sites were recently added. New cameras often send higher initial alarm volumes until NOVA99x learns their environment.
2. Check if any cameras changed location or angle. A repositioned camera generates unfamiliar patterns that NOVA99x treats as real events until recalibrated.
3. Check for seasonal or environmental changes — new construction near cameras, weather patterns, or changed lighting conditions can temporarily raise false alarm rates.

Filter Ratio is consistently below 70%

What you see: NOVA99x is removing fewer than 70% of incoming alarms over an extended period.

What to do:

1. Review your camera sensitivity settings. Overly sensitive motion detection sends events that NOVA99x was not trained on and cannot confidently classify.
2. Check whether multi-monitor mode is correctly configured. If the platform is receiving panel-triggered events rather than camera-generated video alarms, filter performance will be lower.
3. Contact support with your site configuration — some camera environments require NOVA99x tuning specific to your deployment.

Staffing and Coverage

SLA Breach rate is rising

What you see: More alarms are being acknowledged beyond the configured response-time threshold.

What to do:

1. Open Analytics and check the Real Alarm Density by Hour chart. Identify the hours where breaches concentrate — this is usually a staffing gap, not an individual operator problem.
2. Match your shift roster against the peak hours. If the highest-volume window has fewer operators than quieter periods, rebalance shift coverage.
3. Check if one operator's individual Responsive Rate is pulling down the team average. A single understaffed operator on a high-volume site will generate a disproportionate share of breaches.

Activity Rate is low across most operators

What you see: The majority of operators show low active time relative to their logged-in hours.

What to do:

1. This typically means alarm volume is lower than shift coverage. Consider reducing the number of concurrent operators during low-volume windows and shifting that coverage to peak hours.
2. Cross-check with the Alarm Density chart to confirm volume is genuinely low rather than alarms being missed or misrouted.

Alarm Volume Patterns

Alarm volume has grown but CPO has not improved

What you see: More cameras are active but cameras per operator has stayed flat or declined.

What to do:

1. Check whether NOVA99x is keeping pace with the increased volume. A growing fleet means more raw alarms — confirm the Filter Ratio has not dropped.
2. Check whether ZenMode is being used consistently across all operators. CPO improves when both NOVA99x (reduces volume) and ZenMode (reduces per-alarm time) are working together.
3. If both are working and CPO is still flat, the additional cameras may be in high-activity environments that generate more real alarms per camera than your existing fleet. Review the alarm breakdown by site in Analytics.

Alarm volume spikes at specific hours every day

What you see: The Real Alarm Density by Hour chart shows a consistent spike at the same time each day.

What to do:

1. Identify which sites are generating the spike using the site breakdown in Analytics.
2. If the spike is from false alarms, NOVA99x tuning for those specific sites will reduce it. See NOVA99x AI Filtering.
3. If the spike is real alarms (post-NOVA99x), ensure you have adequate operator coverage scheduled for that window. Predictable peaks are a staffing planning input, not a platform problem.

For the full definition of every metric referenced in this guide, see Dashboard KPI Reference. For the Alarm Center Admin View interface, see Alarm Center — Admin View.