Getz Pharma

Finding

Getz manufactures 300+ branded generic formulations, each requiring one or more Active Pharmaceutical Ingredients sourced predominantly from India and China. SAP ERP records every purchase order: molecule, supplier, quantity, price, date. Indian API prices fluctuate with feedstock costs, capacity utilisation, and global demand. The price Getz locked in six months ago for a given molecule may differ by 15-30% from what is available today. On a CY2024e raw material spend of approximately PKR 39bn (40% of ~PKR 97bn estimated revenue), even a modest systematic recovery rate produces material value. Without a benchmarking process running continuously against current market rates, price movements are captured opportunistically rather than systematically.

The price was right at negotiation. The market has moved since, molecule by molecule, without anyone tracking it.

Our Solution

An AI procurement intelligence agent ingests Getz's SAP purchase order history by molecule and supplier, normalises for pack size and purity specifications, and cross-references against live API pricing from Indian and Chinese manufacturer databases. Three outputs: molecules where Getz is paying above the current market benchmark (renegotiation candidates); molecules sourced from a single supplier where dual-sourcing options exist at lower cost; and consolidated buying opportunities where multiple Getz products share an API and volumes could be bundled. Procurement keeps the relationships. The AI agent keeps the market in view. The top 20-30 molecules by spend account for 70-80% of the total savings potential. That is where it starts.

Finding

Getz distributes 300+ products domestically across Pakistan and exports to 45 countries from its Karachi facility. CY2023 export revenue was PKR 23.9bn, PACRA confirmed, representing 31% of group revenue. Through 9MCY24, export share held steady at approximately 30% of PKR 74bn, implying ~PKR 22.2bn in nine-month export revenue. SAP EWM manages the warehouse side. Across this combined domestic and export footprint, freight is purchased mostly on a transaction-by-transaction or annual-contract basis without the kind of AI-assisted benchmarking that systematically identifies where consolidation, modal shift, or distributor replenishment optimisation would reduce cost. In domestic distribution, the consistent inefficiency for Pakistani pharmaceutical manufacturers is the disconnect between what Getz ships to distributors and what distributors actually sell to pharmacies. Without granular secondary sales data feeding production planning, decisions run against distributor order patterns rather than end-market demand.

45 markets. Freight priced one shipment at a time, distributor inventory managed one order at a time.

Our Solution

A distributor performance and secondary sales AI agent reads SAP outbound data against distributor sell-through reports (or proxy demand signals) and flags distributors running high inventory relative to sales velocity, markets losing shelf presence due to supply inconsistency, and production planning signals that reduce over-ordering. For export freight, an AI freight audit tool reviews all international shipments against carrier market rates, identifies consolidation opportunities on lanes where shipments go to the same region within close time windows, and flags where airfreight is handling shipments that cold-chain sea freight could serve at lower cost. Both tools are read-only and advisory. Supply chain keeps the decisions.

Finding

Getz holds product registrations across 45 export markets for a portfolio of 300+ products. Each registration requires an initial regulatory dossier, periodic renewal submissions on 5-year cycles, and variation submissions whenever a formulation, packaging, or API supplier changes. Across 45 countries and 300 products, the theoretical registration universe is 13,500 product-country combinations. Assume 30% are active at any time and the regulatory team is managing approximately 4,000 live registrations with continuous renewal and variation activity. Regulatory dossier preparation is overwhelmingly a document assembly task: pulling Certificates of Analysis, stability data, and manufacturing site files from the quality management system and SAP, and formatting them into country-specific templates that already exist. This is where the time goes, not in the regulatory judgment itself.

4,000 active registrations. The data to populate each dossier already exists in the QMS. It just is not queryable.

Our Solution

A regulatory document preparation AI agent pulls the required data elements from Getz's QMS and SAP for each submission, populates the relevant country-format dossier template (CTD, ACTD, WHO, local format), and outputs a near-complete draft for the regulatory affairs officer to review. Preparation time per dossier falls from an estimated 8-15 hours to 2-4. A registration status tracking system maintains the renewal calendar across all 45 markets, alerting the team six months before each deadline. The AI agent handles the assembly. Regulatory judgment stays entirely with the qualified regulatory affairs professional. Every output is reviewed and approved before submission.