The Government of India’s procurement system, accounting for nearly 20-22% of GDP, is undergoing transformation in 2025. The integration of Artificial Intelligence (AI) into tender evaluation is addressing long-standing challenges of transparency, speed, and accuracy. This shift is critical for sectors ranging from defence to healthcare and infrastructure.
Scale and Complexity of Government Procurement
Government procurement in India covers Union, State, and Public Sector Undertakings. It involves trillions of rupees annually. Platforms like Government e-Marketplace (GeM) have digitised tender submissions. However, evaluating bids remains a manual, document-heavy task. Officers review thousands of pages, verifying eligibility, certificates, and financial data under tight deadlines.
Challenges in Tender Evaluation
Each tender attracts multiple bids, often with extensive documentation. For example, GeM processed over ₹3 lakh crore worth of goods and services in FY24 with more than one crore product listings. Bids may include thousands of pages of statutory, technical, and financial documents. This volume causes cognitive overload, delays, and errors in evaluation.
Role of AI and Machine Learning
AI, combined with Optical Character Recognition (OCR) and semantic parsing, can automate data extraction and comparison. AI tools can ingest bids directly from platforms like GeM, Central Public Procurement (CPP), or Indian Railways E-Procurement System (IREPS). They verify eligibility criteria and generate compliance matrices, denoting issues for human review. This reduces time and errors while enhancing transparency.
Proposed AI-Enabled Evaluation Framework
Several innovations are proposed: – Machine-Readable Bid Annexures (MRBA) – Bidders submit structured data sheets alongside PDFs, enabling rule-based checks. – Trusted Data Cross-Checks – AI links with government databases such as MCA-21, GSTN, EPFO, BIS, and UDYAM for real-time verification. – Clause-to-Evidence Mapping (CEM) – AI maps tender clauses to exact bid documents, creating audit-ready trails. – Risk and Exceptions Register (RER) – AI flags anomalies like missing documents or low bids, requiring human decisions. – AI-Led Rate Reasonability – AI benchmarks prices using historical procurement data adjusted for region and scope.
National Procurement Database and Integration
A centralised procurement database, akin to PM Gati Shakti’s multi-layered data integration, is envisioned. This would improve AI’s benchmarking accuracy by consolidating Last Accepted Rates (LAR) across ministries and states. Such a database would strengthen procurement integrity and efficiency nationwide.
Human Oversight and Accountability
AI supports but does not replace human judgment. Evaluation committees retain final decision-making authority. AI outputs act as preliminary screening tools denoting non-compliance. Clear government guidelines will ensure responsibility and maintain audit trails. AI can also assist in drafting Requests for Proposals (RFPs) and standardising evaluation formats, especially in departments with frequent staff changes.
Implementation Strategy
A phased rollout with pilot projects and training is recommended. Integration of AI modules into existing platforms like GeM, CPP, and IREPS will complement policy reforms. This approach aims to enhance procurement efficiency while safeguarding transparency and accountability.
Questions for UPSC:
- Point out the challenges faced in public procurement systems and estimate how technology can address these issues.
- Critically analyse the role of Artificial Intelligence in enhancing transparency and accountability in government processes with suitable examples.
- Underlining the importance of data integration, what are the benefits and risks of creating a centralised national procurement database?
- With suitable examples, estimate the impact of automation on human decision-making in public administration and governance.
