Pharmacy first

Pharmacy first. Pharmacy deepest. Pharmacy only.

Every generalist FWA platform has a pharmacy module. We have a pharmacy product.

Closed-loop reconciliation

The receipt is not the source of truth. The authorization is.

Every claim ties back to an authorization registered before dispensation. Forgery becomes structurally impossible.

Real-time AI detection

Real-time AI fraud detection. Before the claim is paid.

25 rules plus an AI anomaly layer at claim submission. Not batch audits after the check cleared.

LATAM intelligence

Built for Latin America, not translated into it.

Latin American pharmacy, nothing else. Production-proven in Bolivia today.

Proven results

43.4% of pharmacy spend was detectable anomaly. Observed, not modeled.

Observed on a real 50,000-subscriber Latin American book. $5.1 million of detectable loss in one year on that book.

Three-week POC

3-week proof of concept. No integration required.

We ingest a claims sample. You get your own WAFL breakdown on real data. International Innovation Award 2025.

The same small patterns repeat thousands of times

Each transaction looks like less than $50 USD. Frequency turns pennies into millions.

$4.53M/yr

$16/dispense — Brand dispensed when generic available

71% of branded products have a generic available and it’s clinically safe 99.8% of the time — but less than 5% is dispensed as generic.

$244K/yr

$13/dispense — Same molecule re-authorized within 72 hours

1 in 5 subscribers affected. No clinical scenario requires the same drug twice in 3 days. Each transaction appears valid on its own.

$176K/yr

$41/dispense — Early refills on chronic medications

1 in 16 subscribers. 30-tablet prescriptions refilled with less than 33% consumed. Consistent with stockpiling or diversion.

$126K/yr

$22/dispense — Medication with no diagnostic justification

1 in 6 subscribers received a drug with zero diagnostic justification. 80% were for the wrong body system entirely.

$13K/yr

$14/dispense — Cumulative dose exceeding safe limits

433 subscribers with multiple dispenses of the same molecule in 7 days. If the patient took everything as dispensed, it would be dangerous.

$8K/yr

$26/dispense — Cloned prescriptions reused months later

Same exact medication list, same prescriber, resubmitted months later. Invisible to systems that evaluate transactions individually.

Total projected cost: $5.1M/year per 50,000 subscribers.

Annual costs projected per 50,000 covered subscribers. Based on real-world analysis of Latin American health insurance claims.

How Inspector Works

1

Policy as data

We translate your insurer's coverage rules, formularies, and clinical policies into executable data rules.

2

Clinical routing + real-time rules

Every dispensing request goes through clinical routing and rule validation before authorization.

3

Enforce at dispense

Authorization decisions are enforced directly at the point of dispense, not after the fact.

4

Fingerprint everything

Every decision, exception, and outcome is logged with a full audit fingerprint for compliance.

Real-World Results

43.4% of pharmaceutical spend shows detectable anomalies

Full analysis with 25 detection rules across hundreds of thousands of dispense events from a Latin American pharmacy network.

20.3%

Waste / Utilization

Same-day duplicates, early refills, same-molecule re-authorization within 3 days, same drug family within 7 days, and excessive concurrent medication accumulation

10.8%

Generic Substitution Opportunity

Brand dispensed when a generic alternative is available in Bolivia — 71% of branded products have a generic, but less than 5% is dispensed as generic

7.3%

Clinical Mismatch

Medications with no diagnostic justification, nonspecific diagnoses for targeted therapies, cumulative doses exceeding safe limits, and justifications based only on symptom codes

4.6%

Behavioral Fraud Risk

Incoherent polypharmacy (after 7 clinical exemptions), doctor shopping, cloned prescriptions reused months later, and pharmacy branch dispersion

0.4%

Financial Anomaly

Quantities dispensed exceeding clinical guidelines for the medication

25 detection rules including clinical intelligence, dose validation, prescription clone detection, and generic substitution analysis.

Detection rules validated through structured manual review

Hundreds of thousands of dispense records analyzed across thousands of subscribers from a Latin American pharmacy network

Real-world analysis — Latin American health insurance claims

Recognition

🏆

International Innovation Award 2025

Fundación Iberoamericana Alianza del SeguroMonterrey, Mexico

Inspector AI was recognized as one of the leading innovation projects in the insurance industry, for using advanced analytics, clinical validation, and machine-learning models to help insurers address waste, abuse, fraud, and leakage in pharmaceutical reimbursement processes.

Read more

Frequently Asked Questions

What is WAFL?
WAFL stands for Waste, Abuse, Fraud & Lack of Knowledge. It is the framework Inspector AI uses to categorize pharmaceutical losses: 20.3% waste/utilization, 10.8% generic substitution opportunity, 7.3% clinical mismatch (lack of knowledge), 4.6% behavioral fraud risk, and 0.4% financial anomaly. The Lack of Knowledge category covers clinically inappropriate or rules-violating approvals made by doctors and reviewers who did not know the guideline — not malice, just uninformed decisions.
How much do insurers lose to pharmacy leakage?
In our analysis of Latin American health insurance claims, 43.4% of pharmaceutical spend showed detectable anomalies. This translates to a projected cost of $5.1M per year per 50,000 covered subscribers.
How does Inspector AI detect pharmacy fraud?
Inspector applies 25 independent detection rules covering generic substitution, same-molecule re-authorization, early refills, clinical mismatches, dangerous cumulative doses, and cloned prescriptions, among other patterns.
Do I need to replace my current systems?
No. Inspector integrates with your existing infrastructure. It is FHIR PAS compliant for standard pharmacy integration, connects with call center software like Freshdesk, and offers an OpenAI-compatible API to connect with any system.
What results can I expect?
Results depend on your network, but in real-world analyses we have found: $16/dispense median for generic substitution ($4.53M/yr per 50K), $13/dispense for same-molecule re-auth ($244K/yr), $41/dispense for early refills ($176K/yr), and $22/dispense for clinical mismatches ($126K/yr).
How long does implementation take?
We can have a proof of concept running in 3 weeks. No system integration is needed for the initial analysis — we work with a sample of your dispensing data.
Is Inspector AI available in my country?
Inspector AI is designed for insurers and benefits administrators in Latin America. Our system is built for the regulatory and operational context of the region.
What is the Innovation Award?
Inspector AI received the International Innovation Award 2025 from the Fundación Iberoamericana Alianza del Seguro in Monterrey, Mexico, for using advanced analytics, clinical validation, and machine-learning models to address pharmaceutical waste, abuse, fraud, and leakage.

Let us prove it in your data — Free

We’ll load a sample extract into Inspector AI and show what’s slipping through. If there’s nothing there, we walk away.

Request free analysis