IOCL – AI-Powered Material Codification & Data Governance (Extended Shareholder Report)

At Xapny Solutions, our engagement with Indian Oil Corporation Limited (IOCL) is not just another IT project; it represents a transformational leap for India’s largest refiner in how it governs and manages material master data. IOCL’s vast supply chain spans nine refineries, 125+ depots, pipelines, and thousands of vendors. At the heart of this complex ecosystem lies the material master database — a system with over a million records that directly impacts procurement, maintenance, and capital projects. Yet, this backbone suffers from duplication, poor taxonomy, and non-standardized entries, which have historically led to inefficiencies, excess working capital, and compliance risks.

Xapny was entrusted with designing and implementing an AI-powered Material Codification and Governance System. This system is not a generic tool but a PSU-calibrated, enterprise-grade solution engineered for IOCL’s operational realities: on-premises hosting, SAP MDG/MM integration, CVC and CAG audit requirements, and scalability to millions of records. The vision is simple yet profound: to turn material master data into a trusted asset, preventing duplication at source, enabling semantic search, standardizing taxonomy, and embedding governance across the lifecycle of every code.

1. Context and Problem Statement

Material codification in large PSUs has historically relied on manual review committees, often creating bottlenecks. In IOCL, delays in code creation extend procurement cycles by weeks, leading to stockouts or ad-hoc purchases. Multiple descriptions for identical materials inflate inventory and tie up capital. Audit reports (such as CAG’s observations on PSU procurement inefficiencies) have repeatedly flagged these issues. The absence of AI-driven checks also means that human reviewers struggle to detect near-duplicate requests across such large datasets.

By addressing these pain points, Xapny’s solution aims to reduce procurement cycle time, unlock working capital, improve inventory accuracy, and enhance compliance. This directly supports IOCL’s broader goals of operational efficiency and transparency under PSU governance frameworks.

2. Technical Architecture and Solution Components

Our solution is deliberately structured in layers to ensure resilience, scalability, and compliance:

  • Data Ingestion & OCR: Handles intake of material requests, drawings, and spec sheets. Optical Character Recognition parses legacy PDFs and images, transforming them into structured attributes.
  • Duplicate Prevention Engine: An AI model built on Sentence-BERT embeddings compares new requests against the existing master to flag potential duplicates in real time. Configurable thresholds allow flexibility across material categories.
  • AI Description Generator: Produces both short (40-word) operational descriptions and long (500-word) extended texts. This ensures uniformity across material categories while retaining engineering depth.
  • Classification & Taxonomy Standardization: AI-driven classification aligns codes with IOCL’s taxonomy and UNSPSC standards, ensuring consistency.
  • Workflow & Governance Console: Role-based dashboards with SLA timers, escalation workflows, and immutable decision capture for reviewers and approvers.
  • Integration with SAP MDG: Approved codes are automatically handed over into SAP staging with zero data loss. Our connectors are designed for future S/4HANA migrations.
  • Security & Auditability: Every transaction is logged. Role-based access control, encryption-at-rest, and alignment with ISO/IEC 27001:2022 ensure data security and PSU audit readiness.

This architecture reflects a balance between AI innovation and PSU governance requirements — preventing the very pitfalls that have caused other AI pilots to fail.

3. Proof of Concept (PoC) Framework

Following IOCL’s technical evaluation, Xapny has been cleared to execute a structured 90-day Proof of Concept. Unlike typical pilots, this PoC is outcome-driven and tied to measurable KPIs:

  • Scope: ~500 material requests covering engineering spares, consumables, and project items.
  • KPIs: Duplicate creation rate reduced below 4%; classification accuracy at 95–99%; turnaround time reduced by 80–90%.
  • Governance: Weekly operational reviews with IOCL Materials team; bi-weekly Steering Committee with senior management.
  • Deliverables: KPI dashboards, audit logs, comparison reports, and SAP integration tests.

The PoC is structured as a fixed-fee, milestone-based engagement, aligning risk with deliverables. This approach minimizes IOCL’s exposure while ensuring accountability from Xapny. Successful PoC completion will pave the way for a wave-based rollout across IOCL’s refineries, marketing, and pipeline divisions.

4. Commercial and Strategic Value

From a commercial standpoint, the project has a two-stage revenue model:

  • Short-term: PoC revenue with limited scope but high visibility.
  • Long-term: Multi-year enterprise rollout with AMC and support contracts, representing recurring revenue.

From a strategic perspective, the project provides Xapny with unparalleled market validation:

  • Credibility: Clearing IOCL’s technical evaluation establishes us as the first Indian startup trusted to deploy AI at the material master level in a Fortune 500 PSU.
  • Replicability: The solution architecture is modular, enabling deployment in other PSUs (BPCL, HPCL, NALCO, NTPC) with minimal adaptation.
  • Policy Alignment: Embedding governance by design aligns with Government of India’s emphasis on transparency, efficiency, and MSME participation in PSU digitization.

5. Strategic Implications for Shareholders

For shareholders, IOCL represents both short-term value capture and long-term positioning:

  • In the short term, the PoC ensures immediate revenue, operational visibility, and credibility.
  • In the medium term, enterprise rollout secures multi-year recurring revenues and references for cross-sector expansion.
  • In the long term, the project cements Xapny’s reputation as a trusted AI partner for national infrastructure — a position few startups achieve.

By deliberately addressing the gaps where AI initiatives often fail — lack of integration, poor governance, weak auditability — Xapny has created a blueprint that is cost-effective, scalable, and PSU-ready. This project exemplifies our mission: to deliver AI that is not just intelligent, but reliable, auditable, and aligned with India’s governance priorities.

Conclusion

The IOCL codification project is not a technology pilot; it is a strategic transformation of material governance at the largest energy PSU in India. It combines cutting-edge AI with PSU-grade governance, creating measurable outcomes in procurement efficiency, compliance, and capital optimization. For Xapny, this is both a proving ground and a launchpad — showcasing our ability to bridge AI innovation with national-scale infrastructure needs. For shareholders, it represents a cornerstone engagement that underwrites credibility, growth, and long-term market leadership.