Beyond E-Procurement: How Paharpur Unlocked $12.5M in Strategic Savings with AI
Oaktech Team
Industry Insights

A mature e-procurement system can digitize purchasing processes, standardize approvals, and improve operational efficiency. But what happens when procurement decisions require analyzing hundreds of pages of technical specifications, comparing complex engineering requirements, and evaluating suppliers beyond their quoted prices?
For Paharpur Cooling Towers Ltd., a global industrial manufacturing leader, this was the challenge. The company had already established digital procurement workflows, but complex engineering purchases continued to demand significant manual effort, technical expertise, and time.
By partnering with Oaktech, Paharpur introduced Cognitive AI into its existing procurement infrastructure, unlocking $12.5 million in strategic savings, reducing technical evaluation time by 75%, and achieving 100% audit compliance across complex engineering purchases.
This case demonstrates how AI can move procurement beyond transactional automation and support more informed, data-driven commercial decisions. The Procurement Challenge: Beyond Traditional E-Procurement
Founded in 1948 and headquartered in Kolkata, India, Paharpur Cooling Towers Ltd. is a global leader in the engineering and manufacturing of industrial cooling towers and heat exchangers.
The company operates six integrated manufacturing plants worldwide, generates annual revenue exceeding $850 million USD, and maintains a global workforce of more than 2,697 direct personnel, scaling up to 8,200+ employees across consolidated global subsidiaries.
Its procurement requirements reflect the complexity of its industrial operations.
Engineering accounts for 33.8% of its headcount, while Finance & Operations represents 43.1%. Procurement teams manage high-volume and technically demanding acquisitions, ranging from custom metallic alloys and specialized structural components to engineered-to-order (ETO) equipment essential for heavy industrial infrastructure.
Paharpur had already invested in digital procurement maturity. Purchase requisitions, multi-tier approval matrices, and purchase order issuance were fully digitalized through its existing e-procurement platform and standardized ERP workflows.
However, digitalizing transactions was not enough to address the complexity of strategic procurement.
1. The limitations of static catalogs
Many industrial purchases cannot be standardized into traditional product catalogs.
Technical requisitions frequently included datasheets exceeding 50 pages, complex metallurgical specifications, international ASTM/ASME standards, and custom CAD drawings.
Evaluating these requirements demanded detailed technical interpretation and comparison, something conventional catalog-based procurement workflows were not designed to handle.
2. Time-consuming technical evaluations
Senior buyers spent up to three weeks manually evaluating vendor proposals.
The process involved cross-referencing line-item pricing, international freight terms, delivery lead times, and metallurgical compliance requirements.
Much of this work happened in external spreadsheets, outside the ERP environment, creating additional complexity in the purchasing process.
3. Unmanaged tail spend and spot sourcing
Non-cataloged spot purchases represented a substantial portion of annual spend.
Under tight project deadlines, procurement teams lacked the analytical capacity to conduct exhaustive, multi-round technical evaluations for every non-standard purchase.
The challenge was clear: Paharpur did not need another transactional procurement platform. It needed a way to improve the intelligence behind procurement decisions. Cognitive AI in Procurement: The Oaktech Approach
Rather than replacing Paharpur's existing e-procurement infrastructure, Oaktech introduced an Intelligent AI Procurement Agent that operates as a cognitive decision layer integrated directly into established enterprise workflows.
The objective was to augment procurement teams with AI capabilities that could interpret complex technical documents, evaluate supplier proposals, and support commercial decision-making.
The solution was structured around three specialized capabilities.
1. Document AI and Technical Specification Extraction
Complex technical documentation was one of the biggest bottlenecks in Paharpur's procurement process.
The AI agent was designed to parse technical datasheets, vendor PDFs, and engineering specifications in seconds.
It automatically extracts relevant information, including:
Metallurgical parameters.
Technical tolerances.
Engineering specifications.
Compliance requirements.
The system also identifies specification gaps and potential non-compliance against internal engineering standards before commercial evaluations begin.
This changes the way procurement teams approach technical documentation. Instead of manually reviewing extensive documents and extracting information into spreadsheets, buyers can work with structured technical information to support their evaluations.
2. Total Cost of Ownership (TCO) AI Decision Engine
A competitive purchase price does not necessarily represent the best procurement decision.
For complex industrial acquisitions, evaluating suppliers requires considering several commercial and operational factors beyond the initial quotation.
Oaktech implemented a Total Cost of Ownership (TCO) AI Decision Engine to build multidimensional supplier comparison matrices.
The engine analyzes:
Base pricing.
International freight tariffs.
Import duties.
Historical vendor delivery performance.
Warranty terms.
By consolidating these variables, the solution recommends the optimal supplier based on the lowest true Total Cost of Ownership, rather than simply selecting the lowest initial bid.
This capability allows procurement teams to evaluate supplier proposals with a broader financial perspective and make more informed purchasing decisions.
3. Automated Tail Spend and Spot Sourcing Engine
Non-cataloged purchases often require additional clarification and supplier coordination.
To address this challenge, the AI agent introduced automated technical clarification rounds with pre-approved vendors.
The solution aggregates bid comparisons, verifies vendor certifications, and prepares a fully analyzed recommendation matrix for final buyer approval.
This capability helps procurement teams handle non-standard requisitions with greater analytical coverage, without requiring buyers to manually coordinate every technical clarification and comparison.
The result is a more structured approach to tail spend and spot sourcing, while keeping final purchasing decisions under human approval. Business Results: Measuring the Impact of Cognitive AI in Procurement
Integrating Cognitive AI into Paharpur's procurement workflows generated measurable improvements across operational speed, financial performance, and corporate governance.
Three outcomes stand out.
75% Reduction in Technical Evaluation Time
Automated document parsing and intelligent specification matching reduced the turnaround time for evaluating complex vendor proposals from 18 days to just 4 days.
This improvement accelerated procurement evaluations and supported project delivery schedules across Paharpur's global manufacturing sites.
$12.5M USD in Strategic Savings
$12.5 million USD in auditable financial savings, representing approximately 1.5% of annual spend.
By identifying technical equivalencies in metal alloys and optimizing Total Cost of Ownership across high-value procurement categories, the AI decision engine identified savings opportunities against historical purchase baselines.
This result demonstrates the financial potential of applying AI to complex purchasing decisions, particularly when supplier evaluation involves technical and commercial variables that go beyond standard price comparisons.
100% Audit Compliance on Complex Engineering Purchases
The AI agent established complete transparency and automated decision tracking across all non-cataloged purchases.
The result was 100% adherence to internal governance policies and international manufacturing standards across complex engineering purchases.
This is particularly relevant in industrial procurement, where technical specifications, supplier qualifications, and purchasing decisions require clear documentation and traceability. Why Cognitive AI Is Different from Traditional Procurement Automation
Paharpur's transformation highlights an important distinction in enterprise technology: transactional automation and cognitive intelligence solve different problems.
Traditional e-procurement systems digitize transactions and enforce process compliance. They standardize requisitions, approvals, and purchase orders, creating consistency across purchasing operations.
Cognitive AI extends these capabilities by addressing the analytical work behind complex decisions.
It can interpret unstructured engineering documentation, extract technical requirements, compare supplier proposals, and evaluate multiple variables to support strategic purchasing decisions.
In Paharpur's case, the opportunity was not to replace an established procurement platform. It was to extend its capabilities with an intelligence layer that could handle complex technical and commercial analysis.
This distinction matters because organizations with mature digital infrastructure may still face significant operational bottlenecks that traditional automation cannot resolve.
For engineering-intensive businesses, procurement decisions frequently depend on technical information that does not fit neatly into predefined catalogs or structured ERP fields.
Cognitive AI makes it possible to bring that information into a more structured decision-making process. From Administrative Processing to Strategic Value Creation
The impact of the Paharpur case goes beyond faster supplier evaluations.
By applying AI to complex data analysis, the company enabled engineering and procurement teams to shift their attention from administrative processing toward strategic value creation.
The three capabilities work together to address different stages of the procurement decision:
Document AI structures technical information and identifies specification gaps.
TCO AI evaluates commercial variables to identify the optimal supplier.
Automated Spot Sourcing supports supplier clarification, bid comparison, and recommendation preparation.
Together, these capabilities address technical complexity, commercial optimization, and procurement governance within existing enterprise workflows.
The case also reinforces an important principle for enterprise AI adoption: meaningful transformation does not always require replacing existing systems.
Sometimes, the greatest opportunity comes from adding intelligence to the infrastructure organizations already use. The Future of E-Procurement Is Intelligent Decision-Making
Paharpur's results illustrate what becomes possible when Cognitive AI is applied to a specific business problem rather than introduced as a standalone technology initiative.
With $12.5 million in strategic savings, a 75% reduction in technical evaluation time, and 100% audit compliance, the company demonstrated the measurable business impact of bringing AI into complex procurement workflows.
For organizations managing high-value acquisitions, engineered-to-order components, and non-cataloged spend, the opportunity extends beyond digitizing purchasing transactions.
It is about improving how technical information is interpreted, how suppliers are evaluated, and how purchasing decisions are made.
At Oaktech, we help enterprises turn AI capabilities into measurable business outcomes by connecting intelligence with existing products, processes, and enterprise systems.
Is your procurement operation ready to move beyond transactional automation?
