By Akash Agarwal, Head – Central Planning Cell, Power Transmission & Distribution Independent Company, L&T Construction
The scale of operations in the power sector is significant, particularly in project execution and asset creation. Every year, L&T Construction undertakes asset creation worth nearly Rs 140 billion in India and around Rs 300 billion globally in the power sector through customer contracts and project execution activities. Given this scale, the volume of transactions and operational requirements is extensive, making it increasingly difficult to rely solely on conventional human systems.
With the emergence of artificial intelligence (AI), new tools and techniques have become available, which can enhance efficiency and enable faster processing of information. However, AI adoption cannot happen in isolation. It must build upon a strong digital foundation.
Existing digital systems and tools, which have been in use for nearly 15-20 years, have matured over time and continue to serve as the backbone of operations. These deterministic systems will remain central to future digital transformation efforts, with AI functioning as an additional layer rather than replacing them.
Moreover, digitalisation and AI will coexist over the next one to two years. During this period, digitalisation systems will continue to mature while AI adoption progresses in parallel. Once digitalisation reaches a higher level of maturity, the transition into full-fledged AI-driven operations will become more feasible.
AI across the project life cycle
From a contractor’s perspective, project execution involves an entire life cycle, beginning from the bidding stage and extending to commercial operations. AI has the potential to support several activities throughout this life cycle, particularly those involving repetitive tasks, document-intensive processes and analytical assessments.
AI in tender and contract management
The bidding stage involves extensive document processing, including reading contract documents, extracting bill of quantities, carrying out design estimations and cost assessments, and analysing macroeconomic conditions. These activities present strong opportunities for AI implementation.
L&T has already begun testing certain AI-based tools during the pre-bid stage, although full-scale implementation is still under way. One of the key applications involves contract clause analysis. Based on past experience, disputes often arise because of the wording and framing of contract clauses. In many situations, even when employers are willing to support contractors during genuine cases, contractual language may restrict such flexibility.
AI tools can assist by analysing and reframing clauses to improve clarity and reduce future disputes. In addition, AI can support risk detection, clash analysis and employer profiling. Once a bid is floated into the market, AI tools can analyse customer capabilities, including aspects such as funding arrangements and project viability, enabling faster and more informed assessments.
Document processing
Project tender documents often range between 2,000 and 3,000 pages, making it difficult for individuals to process them within limited time frames. AI-based large language systems can significantly improve document processing by extracting relevant insights from extensive data sets. Rather than manually reading lengthy tender documents, feeding them into AI systems equipped with large language context can enable faster analysis and decision-making.
AI in design engineering
Design engineering is another area where AI can improve efficiency by reducing repetitive efforts. Every year, substantial engineering activities are undertaken, including the preparation of single-line diagrams (SLDs), cable schedules, general arrangement layouts and transmission line tower designs. A significant portion of this work involves repetitive processes. AI can help minimise manual effort by automating repetitive design activities and simulating multiple design scenarios to identify optimal solutions. Since there is often more than one way to design a system, AI can assist in evaluating alternatives that optimise cost and improve efficiency.
This capability can also help employers and customers reduce project costs through more effective design solutions. In addition, when organisations enter new geographical or business territories, AI can support macroeconomic analysis and feasibility assessments. Activities that previously required expert consultants and extensive human effort can now be undertaken more efficiently by teams equipped with proper research approaches and AI tools. AI can also help identify regulatory and permitting requirements that may not be explicitly defined in contracts. By leveraging prior experience and available information, these tools can assist project teams in identifying compliance requirements more effectively.
A live example demonstrated by L&T involved feeding SLDs and general arrangement layouts into an AI-enabled system. While the basic engineering was carried out by human experts, the derived engineering tasks, including cable schedule preparation, cable sizing and cable length estimation, were performed automatically through AI.
AI in project scheduling and monitoring
Project scheduling, monitoring and control remain among the most challenging aspects of project management. Existing systems such as Excel continue to be widely used, while advanced project management platforms such as Primavera involve learning curves that often limit widespread adoption.
AI has played a significant role in enhancing team capabilities within L&T’s Central Planning Cell. Young engineers with limited experience have been able to rapidly develop competencies in areas such as Python programming and Primavera APIs. Teams have also been able to automate schedule creation by converting Excel templates into Primavera schedules.
The organisation is now working towards implementing mini large language models (LLMs) to automate rule-based project scheduling systems. Creating and monitoring realistic project schedules is a complex task, and AI can support the development of systems capable of automatically generating and managing schedules.
Delay analysis is another area where AI applications are being explored. Although AI-generated outputs may not yet be entirely accurate, they already provide valuable insights and enable graduate engineers to undertake analytical tasks that previously required much higher levels of expertise and experience.
AI in procurement and vendor discovery
Procurement forms a major component of engineering, procurement and construction projects, particularly in power distribution, accounting for nearly 70 per cent of project value and effort. AI can support procurement activities by improving vendor discovery and market analysis, particularly when organisations expand into new territories. L&T is currently testing tools capable of identifying vendors by consolidating information from multiple portals and historical performance records.
For example, if a project requires the procurement of 10,000 metric tonnes of steel in a foreign market such as Saudi Arabia, AI tools can analyse available data, consolidate vendor information and recommend suppliers based on performance ratings and historical records.
Site-level AI applications
AI applications are also being tested at project sites through vision analytics and equipment monitoring systems. One such initiative, referred to as Site AI, is currently in the testing phase. These applications are being used for monitoring equipment performance, particularly expensive assets, as well as overseeing workforce activities and administrative functions. Vision analytics can also support project execution monitoring by using bots to inspect sites, ensure compliance with safety requirements and generate quality punch points.
Progress monitoring through image processing is another important area where AI can complement existing digital systems. Since conventional digitalisation tools alone may not be sufficient for image-intensive tasks, AI-enabled visual analytics can help address these challenges.
Building digital maturity alongside AI
The transition towards AI adoption must begin with stronger digital maturity. Core systems such as Primavera, enterprise resource planning systems and computer-aided design tools must first be fully adopted and utilised effectively across organisations. Project management professionals must also expand their skill sets beyond conventional project management functions. There is a growing need for professionals to develop software-related capabilities alongside domain expertise.
According to L&T’s experience, the pace of progress achieved over the past one and a half years has exceeded expectations.
Going forward, organisations may increasingly adopt orchestrated mini-LLMs and on-premise LLMs, particularly where significant investments can be justified by measurable operational outcomes. For organisations capable of making such investments, AI presents an opportunity to accelerate the journey towards higher levels of digital maturity and operational efficiency.
