Photo Evidence Auto-Linking
“A photo in WhatsApp becomes evidence in the AI's analysis — automatically.”
📖 The Scene
Siti saw a cross-module insight linking rain to a boiler drop, but it was only 72% confident. She tapped 'Share to WhatsApp' to ask the field operators. Budi replied confirming the flooded storage, and Rina sent a photo of a dark flame. NayaCore's Bayesian engine immediately correlated their replies, boosting the insight to 91% confidence.
Rina
Security Guard (T1)
📱 See It In Action
Real WhatsApp conversations between operators, supervisors, and NayaCore — in Bahasa Indonesia, exactly as they appear on the phone.
NayaCore mendeteksi kemungkinan fuel basah menyebabkan boiler turun (efisiensi 81%).
Hujan 68mm tadi malam di Block C.
Budi, bisa cek kondisi gudang fuel? Apakah ada air masuk?
Balas dengan observasi Anda. 📷 Foto juga boleh.07:35
Wet Fuel → Boiler Efficiency Drop
Confidence: 72% → 91%
Budi & Rina mengkonfirmasi fuel basah di lapangan.
💡 Rekomendasi: Pindahkan fuel ke storage kering sebelum batch berikutnya.07:49
Discussion → WhatsApp Evidence Loop
Insights are just hypotheses until validated. NayaCore bridges the gap by allowing dispatchers to request field confirmation via WhatsApp, automatically feeding replies back into its Bayesian confidence model.
Share Insight to Field
Siti reviews Insight INS-0847 on the PWA (Wet Fuel → Boiler Drop). The confidence is only 72% based purely on sensor data. She taps 'Share to WhatsApp', and NayaCore suggests Budi and Rina based on their proximity to the fuel storage and boiler.
Field Reply & Photo Evidence
Budi explicitly confirms water ingress. Rina uploads a photo of a dark boiler flame. NayaCore's NLP parses Budi's text as a high-relevance confirmation, while the Vision Pipeline tags Rina's photo as an abnormal flame color.
Auto-Correlation & Bayesian Update
NayaCore ingests the two new Evidence Cards and runs a Bayesian update on the insight. The confidence jumps from 72% to 91%, graduating the status to HUMAN_VALIDATED. NayaCore notifies Siti of the successful loop closure.
📊 Before vs. After
Traditional Operations vs. NayaCore
| Metric | Traditional | NayaCore |
|---|---|---|
| Insight to field validation time | 30–60 min (walk to floor, ask) | 3 minutes (WhatsApp share → reply) |
| Field observations captured | 0% (verbal, forgotten) | >90% (NLP-extracted evidence cards) |
| Alternative hypothesis generation | Never (gut feeling assumed right) | Automatic (if field contradicts AI) |