Selisih is a machine-learning audit platform that learns from thousands of historical audit decisions. Every morning, it ranks the riskiest claims, explains each one in language auditors can cite, and even drafts the audit letter — so your team spends the day deciding, not searching.
Synthetic example — no BPJS PII.
Trillions of rupiah lost every year — not because your team isn't capable.
BPJS Kesehatan receives thousands of claims every day. Manual checks and random sampling leave the more sophisticated patterns invisible. Bad actors learn the sampling rules. Your team burns through cycles on claims that often aren't the most suspicious.
The result: avoidable losses, findings that are hard to defend in front of a contracted hospital, and a workload that simply isn't human-scale. Selisih was built to fix all three — in one flow that feels familiar to auditors, not to data scientists.
“I know there's fraud in the claims I don't see. But I can only review what my supervisor picks at random.”
A machine-learning model picks the 50 most urgent claims every morning — trained on BPJS historical audit patterns. Your team starts from the highest-value work.
Behind every claim: three concrete reasons from AI analysis — cost patterns, length of stay, hospital patterns. Ready to paste into an audit letter.
BPJS rules live as a rule-based layer running alongside the AI model. Your SOP team can add or adjust rules without a programmer.
Instead of starting anywhere, your team opens today's 50 most suspicious claims — already ranked, already labelled.
Three concrete reasons per claim — in sentences you can cite, not in code.
| Pola yang diperiksa | Temuan |
|---|---|
| Biaya vs tarif INA-CBG | 180% (peer 102%) |
| Lama rawat vs peer | 98 persentil |
| Pola rumah sakit | 3× rata-rata daerah |
| Jarak readmisi | — |
| Diagnosis sekunder | Tidak ada |
Confirm, reject, or escalate — with a short note. Every decision strengthens the model for next week.
A full audit letter with evidence and cited rules — the auditor just reviews and signs.
Keputusan auditor ditulis ke tabel anotasi · model retrain mingguan menggunakan label ini · versi model tersimpan untuk audit trail 5 tahun.
A workflow that used to take the whole day now fits into a few hours — with a complete audit trail behind every decision.
At the core of Selisih: classical machine learning with decades of research behind it — deterministic, auditable, runs without internet. On top: five optional AI-copilot features (generative AI for narratives, RAG for regulations, NLP for queries) that help your team move faster. The copilot can be switched off at any time; the core keeps running. The final decision always belongs to the auditor.
'Claims from Hospital X with length-of-stay anomalies in Q1' — the queue filters itself. No query language to learn.
Three technical reasons condensed into one paragraph — ready to send to a hospital or supervisor.
Compares the physician's note with the coded diagnosis — helps auditors catch gaps that routinely go unnoticed.
'What's the readmission-within-7-days rule?' — a brief answer with citations from Peraturan BPJS and PERMENKES.
The system prepares the letter; the auditor approves. All the evidence and rule citations are already attached.
↳ AI only helps explain and compose. Scoring and decisions stay with your team.