Sambhrama BH-3 · MODEL-VALIDATED PERFORMANCE

AI Performance & Benefit Evidence

Measured performance, AI-identified opportunity, actions and verified outcomes.

CURRENT FIELD STATUS

ONLINE

Last field telemetry: 2026-09-11T14:07:17

Current field telemetry is reachable and fresh.

HISTORICAL EVIDENCE AVAILABILITY

Controlled optimisation history: AVAILABLE

Evidence period:: 2026-09-05 to 2026-09-11
Last field telemetry:: 2026-09-11T14:07:17
Data coverage:: 431 operating samples across 7 days
Evidence class:: MODELLED

Authoritative Pilot 2 measured actual and modelled expected history remains available independently of current edge connectivity.

Pilot 1 controlled field evidence

Controlled optimisation aggregate unavailable

The presentation does not recreate or assume aggregate results when the authoritative history source is unavailable or incomplete.

OBSERVED PERFORMANCE

171.4 LPM average measured flow

AI OPPORTUNITY IDENTIFIED

17.2 LPM modelled shortfall

VERIFIED AI BENEFIT

NOT YET ESTABLISHED

Last 10 Days — Benefit Evidence

Up to 10 recent calendar days supported by authoritative history. Missing days are not fabricated.

Evidence period:: 2026-09-05 to 2026-09-11
Days with valid evidence:: 7/10
Coverage:: 70%

Date	Operating Hours	Observed Performance	Expected Performance	AI / Test Action	Outcome	Evidence Result
11 Sept 2026	—	171.4 LPM MEASURED Field flow telemetry aggregated by authoritative Pilot 2 intelligence	188.6 LPM MODELLED Pilot-2 learned healthy operating behaviour	Advisory review; no intervention executed NOT APPLIED · ADVISORY	17.2 LPM MODELLED Measured actual compared with modelled expected Modelled shortfall / potential water-delivery opportunity.	OPPORTUNITY EVIDENCE Why? / Evidence Actual flow is MEASURED field telemetry. Expected flow is MODELLED by Expected Flow AI. The difference is an opportunity, not a saving. No executed intervention or before/after outcome is supplied. AI recommendation Advisory review; no intervention executed Actually executed? No What happened afterward? No qualifying before/after outcome evidence Data quality 42 operating samples support the daily actual/expected comparison. Confidence basis Model: Expected Flow AI flow-v2 · Pilot-2 learned healthy operating behaviour · 42 daily operating samples
10 Sept 2026	—	235.4 LPM MEASURED Field flow telemetry aggregated by authoritative Pilot 2 intelligence	241.1 LPM MODELLED Pilot-2 learned healthy operating behaviour	Advisory review; no intervention executed NOT APPLIED · ADVISORY	5.7 LPM MODELLED Measured actual compared with modelled expected Modelled shortfall / potential water-delivery opportunity.	OPPORTUNITY EVIDENCE Why? / Evidence Actual flow is MEASURED field telemetry. Expected flow is MODELLED by Expected Flow AI. The difference is an opportunity, not a saving. No executed intervention or before/after outcome is supplied. AI recommendation Advisory review; no intervention executed Actually executed? No What happened afterward? No qualifying before/after outcome evidence Data quality 65 operating samples support the daily actual/expected comparison. Confidence basis Model: Expected Flow AI flow-v2 · Pilot-2 learned healthy operating behaviour · 65 daily operating samples
9 Sept 2026	—	222.5 LPM MEASURED Field flow telemetry aggregated by authoritative Pilot 2 intelligence	233.4 LPM MODELLED Pilot-2 learned healthy operating behaviour	Advisory review; no intervention executed NOT APPLIED · ADVISORY	10.9 LPM MODELLED Measured actual compared with modelled expected Modelled shortfall / potential water-delivery opportunity.	OPPORTUNITY EVIDENCE Why? / Evidence Actual flow is MEASURED field telemetry. Expected flow is MODELLED by Expected Flow AI. The difference is an opportunity, not a saving. No executed intervention or before/after outcome is supplied. AI recommendation Advisory review; no intervention executed Actually executed? No What happened afterward? No qualifying before/after outcome evidence Data quality 61 operating samples support the daily actual/expected comparison. Confidence basis Model: Expected Flow AI flow-v2 · Pilot-2 learned healthy operating behaviour · 61 daily operating samples
8 Sept 2026	—	283.1 LPM MEASURED Field flow telemetry aggregated by authoritative Pilot 2 intelligence	279.3 LPM MODELLED Pilot-2 learned healthy operating behaviour	Advisory review; no intervention executed NOT APPLIED · ADVISORY	0 LPM MODELLED Measured actual compared with modelled expected Measured performance agrees with the model; this is not a saving.	OBSERVED Why? / Evidence Actual flow is MEASURED field telemetry. Expected flow is MODELLED by Expected Flow AI. The difference is an opportunity, not a saving. No executed intervention or before/after outcome is supplied. AI recommendation Advisory review; no intervention executed Actually executed? No What happened afterward? No qualifying before/after outcome evidence Data quality 66 operating samples support the daily actual/expected comparison. Confidence basis Model: Expected Flow AI flow-v2 · Pilot-2 learned healthy operating behaviour · 66 daily operating samples
7 Sept 2026	—	186 LPM MEASURED Field flow telemetry aggregated by authoritative Pilot 2 intelligence	197.8 LPM MODELLED Pilot-2 learned healthy operating behaviour	Advisory review; no intervention executed NOT APPLIED · ADVISORY	11.8 LPM MODELLED Measured actual compared with modelled expected Modelled shortfall / potential water-delivery opportunity.	OPPORTUNITY EVIDENCE Why? / Evidence Actual flow is MEASURED field telemetry. Expected flow is MODELLED by Expected Flow AI. The difference is an opportunity, not a saving. No executed intervention or before/after outcome is supplied. AI recommendation Advisory review; no intervention executed Actually executed? No What happened afterward? No qualifying before/after outcome evidence Data quality 65 operating samples support the daily actual/expected comparison. Confidence basis Model: Expected Flow AI flow-v2 · Pilot-2 learned healthy operating behaviour · 65 daily operating samples
6 Sept 2026	—	231.9 LPM MEASURED Field flow telemetry aggregated by authoritative Pilot 2 intelligence	232.3 LPM MODELLED Pilot-2 learned healthy operating behaviour	Advisory review; no intervention executed NOT APPLIED · ADVISORY	0.4 LPM MODELLED Measured actual compared with modelled expected Modelled shortfall / potential water-delivery opportunity.	OPPORTUNITY EVIDENCE Why? / Evidence Actual flow is MEASURED field telemetry. Expected flow is MODELLED by Expected Flow AI. The difference is an opportunity, not a saving. No executed intervention or before/after outcome is supplied. AI recommendation Advisory review; no intervention executed Actually executed? No What happened afterward? No qualifying before/after outcome evidence Data quality 65 operating samples support the daily actual/expected comparison. Confidence basis Model: Expected Flow AI flow-v2 · Pilot-2 learned healthy operating behaviour · 65 daily operating samples
5 Sept 2026	—	256.2 LPM MEASURED Field flow telemetry aggregated by authoritative Pilot 2 intelligence	266.9 LPM MODELLED Pilot-2 learned healthy operating behaviour	Advisory review; no intervention executed NOT APPLIED · ADVISORY	10.7 LPM MODELLED Measured actual compared with modelled expected Modelled shortfall / potential water-delivery opportunity.	OPPORTUNITY EVIDENCE Why? / Evidence Actual flow is MEASURED field telemetry. Expected flow is MODELLED by Expected Flow AI. The difference is an opportunity, not a saving. No executed intervention or before/after outcome is supplied. AI recommendation Advisory review; no intervention executed Actually executed? No What happened afterward? No qualifying before/after outcome evidence Data quality 67 operating samples support the daily actual/expected comparison. Confidence basis Model: Expected Flow AI flow-v2 · Pilot-2 learned healthy operating behaviour · 67 daily operating samples

Evidence classifications

MEASURED: Direct field sensor, VFD or RMS telemetry.
MODELLED: Calculated from an existing trained or authoritative model.
ESTIMATED: Indirectly calculated because direct instrumentation is absent.
VERIFIED: Outcome supported by an executed intervention and before/after evidence.
DATA REVIEW: Anomaly, inconsistency or insufficient data makes the result unsuitable for a benefit claim.
UNAVAILABLE: Required evidence does not exist.

Opportunity is never counted as verified benefit unless an intervention was executed and supported by before/after outcome evidence.