WhatsApp Instant SOP Lookup
“Ask the AI for any procedure — get the answer in seconds, on your phone.”
📖 The Scene
Pak Hendra was at dinner when he asked NayaCore for the week's OEE. Four seconds later, he received a formatted table. On the factory floor, Ahmad asked for a low-pressure SOP. Five seconds later, he received a 3-step summary and the full PDF. NayaCore seamlessly routes unstructured questions to SQL databases or semantic vector stores.
Ahmad
Shift Supervisor (T2)
📱 See It In Action
Real WhatsApp conversations between operators, supervisors, and NayaCore — in Bahasa Indonesia, exactly as they appear on the phone.
Dari: SOP-STERILIZER-001 §4.2
1️⃣ Periksa gauge tekanan di panel
2️⃣ Periksa sumber steam (Boiler efisiensi, valve, dll)
3️⃣ Jika tidak naik dalam 30 mnt, HENTIKAN batch baru.
⚠️ JANGAN paksakan proses dengan tekanan di bawah 2.5 bar — risiko FFB tidak matang sempurna.
Boiler #1: 86% ✅ (batas: 85%)
Boiler #2: 91% ✅
⚠️ Boiler #1 mendekati batas minimum.14:23
Minggu ini Bulan lalu Delta
OEE: 87.3% 84.1% ↑ +3.2
├ Avail: 92.1% 89.4% ↑ +2.7
├ Perf: 95.8% 95.2% ↑ +0.6
└ Quality: 99.0% 98.9% → +0.1
📈 Peningkatan utama: Availability naik karena downtime berkurang setelah bearing Press #3 diganti minggu lalu (INC-1294 resolved).
Target: 88.0%
Saat ini: 87.3% (minggu 1)
Gap: -0.7pp
Untuk capai target:
• Availability harus ≥ 92.5%
• Perhatikan sterilizer #2 (efisiensi 89%, mendekati batas)20:46
Quality minggu lalu: 93.8% (normal: 96.2%)
📊 AKAR MASALAH (Multi-Module):
🌧️ Hujan deras (Cakrawala)
🌿 Panen terlambat & overripe (Plantation)
🏭 Sterilizer drop 91%→86% (MillOS)
📉 FFA naik → CPO quality turun
💡 REKOMENDASI:
Terapkan aturan harvest-delay untuk Block C saat hujan > 50mm.09:00
Semantic Search & Ad-Hoc Queries
NayaCore dynamically routes WhatsApp queries into SQL searches or RAG semantic searches, maintaining context across multiple questions.
Knowledge Query (RAG)
Ahmad is on the factory floor and needs the low-pressure SOP. Instead of walking to the control room, he asks WhatsApp. NayaCore runs a pgvector RAG search and delivers a precise, 3-step summary with a PDF attachment of the full SOP.
Data Query (SQL)
Pak Hendra is at dinner and wants to check OEE. NayaCore classifies this as a data query, builds a TimescaleDB SQL query, and returns a formatted table with trend analysis in 4 seconds.
Cross-Module Reasoning
A complex 'Why' question requires NayaCore to pull weather data (Cakrawala), harvest delays (Plantation), and SCADA metrics (MillOS). It chains them together into a root-cause explanation.
📊 Before vs. After
Traditional Operations vs. NayaCore
| Metric | Traditional | NayaCore |
|---|---|---|
| Time to answer an OEE question | 15–30 min (ask staff, pull report) | 4 seconds (WhatsApp reply) |
| SOP lookup time | 5–10 min (find binder) | 5 seconds (RAG retrieval) |
| Dashboard usage after 1 month | ~15% still check daily | 80%+ use WhatsApp daily |