Good morning, Sarah.
Here is what your AI co-agent found while you were away.
Monitoring 47 facilities - Last analysis: 6:12 AM
2.3M
Data points analyzed overnight
47
Facilities monitored
4
Findings require attention
Findings - Priority Ordered
CRITICAL ALARM
Booster Pump 4B - Alarm Pattern Anomaly Detected
ALARM INTELLIGENCE
ESCALATED: "Low Discharge Pressure" alarm fired at 11:47 PM last night. Night operator acknowledged in 4 seconds (typical for this alarm). But this time was different. Discharge pressure had been declining 2 psi/hour for the previous 6 hours before crossing the alarm threshold. This alarm has fired 312 times in the past year - average acknowledge time is 3.2 seconds. Operators treat it as noise. This is NOT noise.
Pattern analysis indicates early-stage cavitation or discharge check valve degradation. Suction pressure remained stable, ruling out supply issues. This matches the pattern from March 2025 that resulted in emergency pump replacement ($14,200 repair cost, 18 hours reduced capacity).
Recommended: Immediate investigation required. Check suction line for air leaks, inspect discharge check valve operation, verify impeller clearances. Do not dismiss this alarm as routine - the 6-hour pressure decline makes this actionable.
Alarm Intelligence Booster Station 4 Safety Critical
HIGH
Well Pump 7 Efficiency Decline
Discharge pressure dropped 8% over 12 days at constant speed. Runtime up 36%, production flat. Pattern consistent with impeller degradation or check valve obstruction.
Recommended: Operator should schedule a drawdown test and compare results to the last rehabilitation baseline.
Pump Station Predictive
WATCH
Effluent Ammonia Trending Toward Limit
Limit
Increased from 1.2 to 3.8 mg/L over 14 days. Permit limit: 5.0 mg/L. Estimated exceedance in 6-8 days at current rate. Correlated with 18% influent loading increase.
Recommended: Operator should increase aeration basin DO setpoint from 2.0 to 2.5 mg/L at the SCADA HMI. Monitor response over 48 hours.
Compliance Process
INSIGHT
Energy Optimization Opportunity - Main St Boosters
61% of electricity costs occur during peak-rate hours across three booster stations. Estimated $34K annual savings available through pump scheduling adjustment and VFD optimization.
Recommended: Review VFD setpoint optimization with time-of-use rate shifting. Operator to adjust pump schedules at SCADA workstation.
Energy Cost Savings
Alarm Intelligence LEARNING FROM 312 ALARMS

The AI monitors every SCADA alarm and learns operator response patterns. When an alarm fires that looks like routine noise but is actually different from the pattern, it escalates to ensure operators investigate before acknowledging.

218
Alarms last 7 days
189
Routine (acknowledged <5 sec)
1
Escalated (anomaly detected)

Escalated Alarms Requiring Investigation

ESCALATED Feb 6, 2026 11:47 PM

Booster Pump 4B - Low Discharge Pressure

This alarm has fired 312 times in the past year. Average operator acknowledge time: 3.2 seconds. Night operator acknowledged this instance in 4 seconds. But this time was different.

Why This Was Escalated:
  • Discharge pressure had been declining 2 psi/hour for 6 hours before crossing alarm threshold
  • Pattern does not match the 311 previous occurrences (which were momentary pressure dips)
  • Sustained pressure decline indicates equipment degradation, not transient condition
  • Similar pattern occurred March 2025 and resulted in emergency pump failure ($14,200 repair)
AI Recommendation:

Do not dismiss this alarm as routine. Investigate immediately: Check suction line for air leaks, inspect discharge check valve operation, verify impeller clearances. Early-stage cavitation or valve degradation likely.

Nuisance Alarm Candidates

These alarms fire frequently and are routinely acknowledged without investigation. Consider threshold adjustment or suppression.

Tank 7 - High Level Alarm

Fired 47 times (last 7 days) | Avg acknowledge: 2.1 seconds | 0 investigations

AI Analysis: Alarm threshold set at 95% full. Tank reaches 94.8-95.2% daily during normal operations. Threshold should be raised to 96% or alarm should be suppressed as informational only.

Booster 3 - Chlorine Residual Low

Fired 28 times (last 7 days) | Avg acknowledge: 3.8 seconds | 0 corrective actions

AI Analysis: Residual drops below 0.2 mg/L threshold twice per shift during peak demand. Returns to normal within 5 minutes without operator intervention. Likely flow surge transient. Consider time-delay on alarm or adjust threshold to 0.15 mg/L.

Zone 4 - Low Pressure Alert

Fired 34 times (last 7 days) | Avg acknowledge: 2.9 seconds | 2 investigations

AI Analysis: Triggers nightly between 2-4 AM when irrigation load drops and pressure control valve overshoots. Self-corrects within 3 minutes. Operators investigated twice but found no issue. Consider adding deadband to pressure control or suppress alarm during 2-4 AM window.

How Alarm Intelligence Works

Learns Normal: Tracks every alarm, operator acknowledge time, response actions, and outcomes over months
Identifies Noise: Alarms acknowledged in <5 seconds with no investigation = routine noise operators ignore
Detects Anomalies: When alarm looks like noise but data pattern is different, escalates with context
Recommends Fixes: Suggests threshold adjustments, suppression windows, or time-delays to eliminate nuisance alarms

Ask About Your Alarms

Have questions about alarm patterns, thresholds, or why something was escalated? Ask the AI analyst.

Why was the Pump 4B alarm escalated last night? We get that alarm all the time.
8:14 AM
ALARM ANALYST
Pump 4B Escalation Explanation
You are correct that "Low Discharge Pressure" at Pump 4B fires frequently - 312 times in the past year. Average operator acknowledge time is 3.2 seconds, which tells me operators treat it as routine noise.
But the instance at 11:47 PM last night was different:
  • Discharge pressure had been declining 2 psi/hour for 6 hours before crossing the alarm threshold
  • The 311 previous alarm instances were momentary pressure dips that self-corrected in under 30 seconds
  • A sustained 6-hour decline indicates equipment degradation, not a transient condition
  • I found a similar pattern in March 2025 that resulted in emergency pump failure ($14,200 repair cost)
The alarm itself looked routine, but the underlying data pattern was not. That is why I escalated it - to make sure an operator investigated before dismissing it as noise.
8:14 AM - Analysis completed in 1.8 seconds
Show me all the alarms at Booster Station 4 in the last 30 days. Which ones should I be worried about?
8:16 AM
ALARM ANALYST
Booster Station 4 - Alarm Summary (Last 30 Days)
I found 47 alarm events at Booster Station 4 in the past 30 days:
⚠️ Requires Attention (1 alarm)
  • Pump 4B Low Discharge Pressure - Escalated Feb 6 at 11:47 PM due to 6-hour declining trend (not typical pattern)
✓ Routine Noise (46 alarms)
  • Pump 4A VFD Communication Loss (12 times) - Momentary Modbus timeout, self-clears in 2-3 seconds
  • Pump 4B Low Discharge Pressure (18 times, excluding escalation) - Momentary dips during pump staging, avg 8 seconds duration
  • Station 4 High Pressure (16 times) - Pressure control valve overshoot during low-demand periods, self-corrects within 1 minute
My recommendation: The only alarm requiring attention is the Pump 4B escalation from last night. The other 46 alarms are routine operational transients. However, the "VFD Communication Loss" alarm firing 12 times in 30 days might indicate a Modbus wiring or termination issue worth investigating during scheduled maintenance.
8:16 AM - Analysis completed in 2.3 seconds
We get the Tank 7 high level alarm constantly. Should we just turn it off or raise the threshold?
8:18 AM
ALARM ANALYST
Tank 7 High Level Alarm - Threshold Analysis
I have been tracking Tank 7 high level alarm for the past 6 months. Here is what I see:
  • Current threshold: 95.0% full
  • Alarm frequency: 47 times in last 7 days (6.7 alarms per day)
  • Operator acknowledge time: avg 2.1 seconds (treated as noise)
  • Tank level during normal operations: 94.8% to 95.2% daily cycle
  • Actual overflow risk level: 98.5% full based on tank geometry and inlet flow rates
Recommendation: Raise Threshold to 96.5%

Do not disable the alarm entirely - tank overflow is a real operational risk. Instead, raise the threshold to 96.5%. This gives you:

  • Eliminates nuisance alarms during normal 94.8-95.2% cycling
  • Still provides 2% safety margin before actual overflow risk at 98.5%
  • Allows operators to respond when tank level is truly abnormal
Alternative option: If you want to keep the 95% threshold for informational purposes, change it to a "low priority" or "notification" alarm that does not require acknowledgment. Reserve "high priority" alarms for conditions that need immediate operator action.
8:18 AM - Analysis completed in 1.9 seconds
Show alarms for Zone 4 last week Why do we get so many chlorine residual alarms? Which alarms should we suppress? Compare alarm patterns month over month
Effluent Ammonia - Compliance Trend Analysis AI Analysis
WATCH - PERMIT EXCEEDANCE PROJECTED IN 6-8 DAYS
Effluent Ammonia (mg/L)
Influent Loading (% of design)
Permit Limit (5.0 mg/L)
Projected Trajectory
01.02.03.04.05.0+ Ammonia (mg/L) 60708090100110 Influent Load (%) Jan 232425262728293031Feb 12345 Permit: 5.0 mg/L Projected: ~Feb 12-14
Influent Correlation Detected - Jan 28
Influent loading began sustained increase from 82% to current 97% of design capacity. Seasonal pattern consistent with prior January-February periods. Ammonia rise rate correlates at 0.89 with influent loading change - strongly suggesting biological treatment capacity is being exceeded, not a process upset.
Rate Acceleration - Feb 2
Ammonia increase rate accelerated from 0.15 mg/L per day to 0.22 mg/L per day. Blower runtime and DO setpoints have not changed during this period. Without intervention, current trajectory projects permit exceedance (5.0 mg/L) between February 12-14.

AI Compliance Analysis Generated 6:12 AM

Effluent ammonia has increased from 1.2 mg/L to 3.8 mg/L over 14 days. The current reading is 76% of the 5.0 mg/L permit limit. At the current rate of increase (0.22 mg/L/day over the last 5 days), the permit limit will be reached between February 12-14 without intervention.
Root Cause Assessment:
The ammonia increase correlates strongly (r=0.89) with rising influent loading, which has increased from 82% to 97% of design capacity over the same period. This is consistent with seasonal patterns observed in prior years. The biological treatment process is not keeping up with the higher nitrogen load because aeration capacity has not been adjusted to match.
Recommended Actions (priority-ordered):
Immediate - Today
I recommend the operator increase the aeration basin DO setpoint from 2.0 mg/L to 2.5 mg/L at the SCADA HMI. This provides additional dissolved oxygen for nitrifying bacteria to process the higher ammonia load. Expected response: ammonia reduction of 0.3-0.5 mg/L within 48 hours based on historical DO-response data for this facility.
48-Hour Follow-Up
If ammonia has not stabilized below 3.5 mg/L after 48 hours, consider increasing DO setpoint to 3.0 mg/L and bringing Blower 3 online for additional aeration capacity. The operator should make these changes at the SCADA HMI and set effluent ammonia sampling to 4-hour intervals during this period.
Planning - This Week
I recommend reviewing the seasonal aeration strategy with operations staff. Influent loading increases in January-February have caused similar ammonia trends in 2024 and 2025. Consider implementing a seasonal DO setpoint schedule at the SCADA system that proactively increases aeration capacity when influent loading exceeds 85% of design.

Current Readings

Effluent Ammonia
3.8 mg/L
76% of 5.0 mg/L permit limit
Days to Projected Exceedance
6-8 days
At current rate without intervention

Correlated Parameters

Influent Loading
97% of design (+18%)
Aeration DO Setpoint
2.0 mg/L (unchanged)
Blower Runtime
Blowers 1 & 2 at 92%
Blower 3 offline (available)
Mixed Liquor (MLSS)
3,400 mg/L (normal)
Water Temperature
12.4 C (seasonal low)
Lower temp reduces nitrification rate
Historical Pattern
Similar trend Jan-Feb 2025
Resolved with DO increase to 2.5
Compliance Impact if Permit is Exceeded
$25-50K
Potential penalty per exceedance event
Mandatory
State reporting within 24 hours
Possible
Consent order or enforcement action
Operator Workflow
I have increased the DO setpoint to 2.5 mg/L on Aeration Basin 1 at the SCADA HMI. Confirmed the blowers ramped up to match. What should I watch for over the next few hours?
7:14 AM
AI ANALYST
DO Setpoint Change Acknowledged - Monitoring Initiated
Good. I can see the DO reading in Basin 1 has started responding - currently reading 2.18 mg/L and climbing toward the new 2.5 mg/L setpoint. Here is what to watch for:
Next 2-4 Hours
DO should stabilize at or near 2.5 mg/L. If it plateaus below 2.3 mg/L, the blower may not have sufficient capacity and you may need to consider bringing Blower 3 online.
12-24 Hours
Ammonia should begin to flatten or decrease slightly. The nitrifying bacteria need time to respond to the additional oxygen. If ammonia continues rising at the same rate (0.22 mg/L/day), the DO increase alone may not be sufficient.
48 Hours
Target: effluent ammonia below 3.5 mg/L. Based on historical data, this facility has responded within 36-48 hours to DO increases of this magnitude during similar loading conditions.
I will continue monitoring the DO level, ammonia readings, and blower performance. I will alert you if ammonia does not begin responding within 18 hours or if DO fails to stabilize at the new setpoint. No further action is needed from you right now unless you want to adjust the monitoring interval.
Monitoring Schedule (automated alerts)
Every 4 hours: Ammonia trend check and DO stability verification
At 18 hours: First response assessment - will alert if no improvement detected
At 48 hours: Full summary with recommendation for next steps
7:14 AM - Monitoring schedule configured
If I need to bring Blower 3 online, are there any concerns with the electrical capacity or anything I should check first?
7:16 AM
AI ANALYST
Blower 3 Readiness Assessment
Good question. I reviewed the available operational data for Blower 3:
Last Run Date
Dec 18, 2025 (49 days ago)
Last Maintenance
Oct 2, 2025 (bearing service)
MCC Breaker Status
Unknown (not monitored by SCADA)
Rated Capacity
75 HP - same as Blowers 1 & 2
Before starting Blower 3, I recommend the operator check the following at the blower building:
  1. Verify the MCC breaker for Blower 3 is in the ON position
  2. Check the inlet filter for debris or blockage (49 days idle)
  3. Confirm discharge valve is open and check valve is functioning
  4. Start in manual mode first and verify airflow before switching to auto
Running all three blowers simultaneously should be within the electrical panel capacity based on the original design records, but I do not have real-time electrical metering data to confirm current loading. If the plant has a main switchgear power meter, it would be worth checking that total demand stays below the panel rating when Blower 3 starts.
7:16 AM - Assessment completed in 2.1 seconds
AI Operations Analyst Reading from TimescaleDB - 2.3M data points (read-only)
How can I meet production targets but reduce energy costs at the three Main Street booster stations?
6:42 AM
AI ANALYST
Energy Analysis: Main Street Booster Stations 1, 2, 3
Analyzed 12 months of operational data (Jan 2025 - Jan 2026):
Monthly Energy Cost - Peak vs Off-Peak
JanFebMarAprMayJunJulAugSepOctNovDec Peak Off-Peak
  • Total electricity cost: $187,400 over 12 months
  • Peak-hour consumption: 61% of total cost ($114,300)
  • Average production: 4.2 MGD across three stations
  • Current efficiency: 1,480 kWh per million gallons
Recommendations
  1. Consider shifting 30% of peak-hour pumping to off-peak using storage tank float optimization. The operator would adjust pump schedules at the SCADA workstation. Estimated savings: $34,300/year.
  2. I recommend adjusting Station 2 VFD setpoint from 52 Hz to 48 Hz during low-demand periods (10 PM - 6 AM). This change should be made by the operator at the VFD controller or SCADA HMI. Estimated savings: $8,200/year.
  3. Station 3 Pump B is operating 18% less efficiently than Pump A at comparable loads. I recommend designating Pump A as lead during peak hours in the pump scheduling configuration. Estimated savings: $4,100/year.
Total Estimated Savings: $46,600/year
Data sources: TimescaleDB power metering tags, SCADA pump status, utility rate schedule (Rocky Mountain Power TOU-8)
6:43 AM - Analysis completed in 4.2 seconds
What about Station 3 Pump B - is that an efficiency problem or a maintenance issue?
6:45 AM
AI ANALYST
Station 3 Pump B - Performance Comparison
Efficiency (kWh/MG) vs Flow Rate (GPM) - Pump A vs Pump B
1200140016001800 4006008001000 Flow Rate (GPM) kWh/MG Pump A Pump B 800 GPM
Pump B is consuming 18% more energy than Pump A at comparable flow rates. The efficiency gap widens at higher flows (above 800 GPM), which suggests impeller wear rather than a controls issue. A controls problem would show consistent offset across all flow rates.
  • Pump B last maintenance: July 2025 (bearing replacement)
  • Pump B last impeller inspection: March 2024
Recommendation
  1. I recommend scheduling a vibration analysis and impeller inspection. The field crew should perform these tests and record findings in OCore. If impeller replacement is needed, estimated cost is $3,200 for parts and labor vs. $4,100/year in excess energy cost. Payback period: 9.4 months.
6:45 AM - Analysis completed in 2.8 seconds
Create a maintenance task for Pump B impeller inspection and assign it to the field crew.
6:47 AM
AI ANALYST
Task Created: Station 3 Pump B - Impeller Inspection
Task IDWO-2026-0847 PriorityMedium AssignedField Crew A - Morning Shift LocationMain St Booster Station 3 DueFeb 14, 2026 EquipmentBST3-PUMP-B (Goulds 3196)
Scope of Work (auto-generated from analysis):
  1. Perform vibration analysis on Pump B bearings and impeller assembly
  2. Compare vibration spectrum to Pump A baseline for same flow conditions
  3. Inspect impeller for erosion, pitting, or tip clearance wear
  4. If impeller replacement needed: estimated $3,200 parts + labor (see attached analysis)
  5. Record all measurements in OCore for AI trend tracking
Attached Context:
AI Efficiency Analysis (Feb 6) Pump Curve Comparison Goulds 3196 Service Manual
Task created and assigned to Field Crew A. The energy analysis, pump performance comparison, and equipment manual are attached to the work order so the crew has full context when they perform the inspection. Once the crew records their findings in OCore, I will incorporate the results into the efficiency model for Station 3.
6:47 AM - Task created in 1.1 seconds
Can you build me a summary report of this energy analysis for the board meeting next Tuesday? Include the savings projections and the Pump B maintenance recommendation.
6:49 AM
AI ANALYST
Board Report: Main Street Booster Stations Energy Optimization
Report Preview
Energy Optimization Analysis - Main Street Booster Stations
Prepared February 6, 2026 | Data Period: January 2025 - January 2026
Executive Summary
AI-driven analysis of 12 months of operational data across Main Street Booster Stations 1, 2, and 3 identified $46,600 in annual energy savings through three operational adjustments. An additional equipment maintenance issue at Station 3 is contributing to $4,100/year in excess energy consumption and poses reliability risk if unaddressed.
$187,400
Annual Energy Cost
$46,600
Projected Annual Savings
24.9%
Cost Reduction
Recommended Actions
1. Operators optimize pump scheduling for off-peak rate periods ($34,300/yr)
2. Operators adjust Station 2 VFD setpoints during low-demand windows ($8,200/yr)
3. Designate Pump A as lead pump at Station 3 ($4,100/yr)
4. Field crew to perform Station 3 Pump B impeller inspection ($3,200 investment, 9.4-month payback)
Full report includes the trend charts, pump performance comparison scatter plot, month-by-month cost breakdown, and the maintenance work order for Pump B. All data is sourced from your TimescaleDB operational records with the analysis methodology documented. Want me to adjust anything before export?
6:49 AM - Report generated in 3.6 seconds
If we move forward with the pump scheduling changes, what order should we implement them in? And what should the operators watch for after making each change?
6:51 AM
AI ANALYST
Recommended Implementation Sequence and Monitoring Plan
Here is the sequence I would recommend. Each step should be implemented by the operator at the SCADA workstation, with a stabilization period before moving to the next change.
Week 1 Designate Pump A as lead at Station 3
This is the lowest-risk change. The operator should update the lead/lag configuration in the pump controller or SCADA. No infrastructure changes required.
What to watch:
  • Verify Pump A starts first on demand calls for the first 48 hours
  • Confirm storage tank levels maintain target range during peak demand
  • Monitor Pump A runtime - it should not exceed 18 hrs/day to avoid premature wear
Week 2 Adjust Station 2 VFD setpoint for off-peak
Operator should reduce VFD frequency from 52 Hz to 48 Hz during the 10 PM - 6 AM window. This can be configured as a time-based schedule in most VFD controllers or through the SCADA system.
Risks to monitor:
  • Pressure at the furthest service connection should stay above 35 psi minimum
  • Storage tank should not draw down more than 6 feet during off-peak reduced-speed operation
  • If pressure complaints come in from the service area, revert to 52 Hz and we can re-evaluate
Week 3-4 Shift peak-hour pumping using storage optimization
This is the highest-impact change but requires adjusting storage tank float levels and pump start/stop setpoints. The operator should modify the high and low float setpoints to pre-fill tanks during off-peak hours (before 2 PM) and coast on stored water during peak rate hours (2 PM - 8 PM).
Critical checks:
  • Do not reduce storage below 50% capacity during peak demand periods
  • Fire flow requirements must be maintained at all times - verify with fire marshal if needed
  • Start with a 15% shift target for the first week, then increase to 30% after confirming system stability
After each step, I will track the energy metering data and compare against the baseline to measure actual savings vs. projections. I can generate a weekly comparison report so you can see the impact of each change before moving to the next one. All operational changes are made by you and your team - I will handle the data tracking and performance analysis.
Note: I have read-only access to SCADA and metering data. All setpoint changes, pump configurations, and VFD adjustments must be performed by the operator.
6:52 AM - Analysis completed in 3.4 seconds
How does our energy efficiency at these stations compare to industry benchmarks? Are we already doing well or are we behind?
6:54 AM
AI ANALYST
Energy Benchmarking: Main Street Boosters vs. Industry Standards
Energy Intensity Comparison (kWh per Million Gallons)
800 1000 1200 1400 1600 AWWA Top Quartile 1,050 AWWA Median 1,250 Your Stations (now) 1,480 Projected (after changes) 1,160 AWWA Bottom Quartile 1,580
Your Main Street booster stations are currently at 1,480 kWh/MG, which places you in the bottom third of comparable booster installations according to AWWA benchmarking data for small-to-medium distribution systems (1-5 MGD).
  • AWWA top quartile for your system size: 1,050 kWh/MG
  • AWWA median: 1,250 kWh/MG
  • Your current performance: 1,480 kWh/MG (18% above median)
  • Projected after recommended changes: ~1,160 kWh/MG (7% below median)
The good news is that the recommended changes would move you from the bottom third to slightly better than median - a significant improvement without capital investment. To reach top-quartile performance, you would likely need VFD retrofits at Stations 1 and 3 (which currently have across-the-line starters) and possibly a storage tank capacity study to enable more aggressive off-peak pumping. That would be a separate capital planning conversation worth having with the board.
Want me to track progress?
After the operator implements each change, I can generate weekly benchmarking updates showing where you stand relative to AWWA targets. This gives you a clear before-and-after story for the board and helps validate the savings projections in real time.
Benchmark data: AWWA Energy Management for Water & Wastewater Utilities (2024 update). System size comparison: 1-5 MGD booster stations, western US utilities.
6:55 AM - Analysis completed in 4.1 seconds
VFD retrofit ROI Escalated alarms analysis Identify nuisance alarms 3-year energy trend Find similar stations
AI Agent Settings

Configure how often your active AI agents run analysis and monitoring tasks. Sentinel agents run continuously on scheduled intervals. Analyst agents are available on-demand anytime.

Active Sentinel Agents

SCADA Alarm Intelligence & Prioritization

CRITICAL SAFETY

Monitors all SCADA alarms and operator response patterns. Learns which alarms are routinely acknowledged without investigation. Escalates critical alarms that deviate from normal patterns to prevent operators from missing dangerous conditions hidden in noise.

312 alarms monitored Active 1 escalation last 24h

Last 7 days: 218 alarms fired. 189 acknowledged in under 5 seconds (routine). 27 required investigation. 1 escalated due to anomalous pattern (Booster Pump 4B, Feb 6 11:47 PM). Learning: "Low Discharge Pressure" alarm at Booster 4B is typically nuisance, but 6-hour pressure decline before alarm threshold made this instance critical.

Equipment Performance Degradation Monitor

Analyzes pump efficiency, bearing vibration patterns, and runtime trends to predict maintenance needs 2-4 weeks before failure.

47 data sources Active

Compliance Boundary Monitor

Tracks effluent quality parameters against permit limits and alerts when trends approach regulatory boundaries.

12 data sources Active

Active Analyst Agents

Energy & Optimization Analyst

Answers operator questions about energy consumption, pump efficiency, VFD optimization, and time-of-use rate strategies.

On-demand Active

Analyst agents are available on-demand anytime through the AI Analyst interface. No scheduling configuration needed.

Custom Agent Builder DESIGN YOUR AGENT

Build a custom AI agent tailored to your specific operational needs. This guided builder helps you define what patterns to monitor, what data sources to analyze, and what recommendations the agent should provide. All agents are read-only - they analyze data but never control systems.

Example: Distribution System Pressure Optimizer

SENTINEL AGENT

Monitors distribution system pressure zones continuously to identify opportunities for reducing energy consumption while maintaining minimum pressure requirements. Analyzes pressure readings across 47 monitoring points, correlates with pump station operations, and recommends pressure setpoint reductions when system-wide pressure exceeds minimum requirements by more than 15 psi.

Data Sources:
Pressure sensors (47 points), pump station flow and pressure, tank levels, VFD speeds
Monitoring Pattern:
Continuous analysis of pressure vs. minimum requirements, flagged when system-wide cushion exceeds 15 psi for 4+ hours
Recommendations:
Specific VFD setpoint reductions at identified pump stations, estimated energy savings, pressure safety margin verification
Expected Impact:
10-15% reduction in distribution pumping energy, $30-50K annual savings for mid-sized system

Build Your Custom Agent

Sentinel Agent

Continuous monitoring. Watches data patterns 24/7 and proactively alerts operators when conditions warrant attention. Best for predictive maintenance, efficiency monitoring, and compliance tracking.

Analyst Agent

On-demand analysis. Operator asks questions, agent analyzes data and provides answers with supporting evidence. Best for optimization studies, root cause investigation, and report generation.

Choose which SCADA tags and operational data the agent should analyze. The agent will have read-only access to this data.

What specific actions should the agent recommend? Remember: the agent analyzes and recommends, the operator reviews and executes.

Next steps: After submission, the OCore team will configure your custom agent using your specifications. Setup typically takes 2-3 weeks depending on data source complexity and baseline data collection requirements. You will receive a test deployment for validation before the agent goes into production monitoring.