NHAI deploys in-house AI system to flag issues in highway projects

SEeking to curb bit high-priced errors during main road twist, the subject Highways Authority of India (NHAI) is relying on an in-house artificial intelligence (AI) tool to detect discrepancies in detailed project reports (DPRs), verify conditions on the ground and flag potential issues before they escalate into delays or cost overruns.DPRs serve as the blueprint for highway projects, determining what will be built, how it will be executed and the estimated cost.“Detailed Project Reports are our biggest problem. In a ₹2,000-crore project, variations of ₹200 crore to ₹300 crore can arise simply because something was missed during planning. This tool helps identify those gaps at the DPR stage itself,” NHAI chairman Santosh Kumar Yadav said.The software, called the Technical Schedule Analyzer, compares information contained in project reports with technical standards, policy provisions and a growing repository of field data. At present, it analyses Schedule B and Schedule C sections of DPRs, checking project quantities and specifications against Indian Roads Congress (IRC) codes, Ministry of Road Transport and Highways norms and NHAI standards, while flagging inconsistencies and omissions before work begins.Also Read:NHAI identifies land needed for two flyovers at Rajiv Chowk“Soon, the system will also draw data from drone surveys, Network Survey Vehicle records, geospatial information, monthly progress reports and site photographs to verify whether what has been planned on paper matches conditions on the ground,” a member of NHAI’s AI Cell said.Officials said such capabilities could help identify mismatches between project designs and site realities much earlier in the project cycle, reducing the need for costly changes after construction has commenced.The Technical Schedule Analyzer is among a growing suite of AI applications being developed by NHAI’s in-house AI Cell, which was established in August 2025.The initiative emerged from an internal assessment which found that project directors and regional officers were spending a substantial amount of time handling routine administrative matters, contractual disputes and document reviews rather than focusing on site inspections and project monitoring.To address this, NHAI assembled a multidisciplinary team comprising civil engineers, AI engineers and design specialists, including young recruits from leading institutes and the private sector. Over several months, the team digitised and organised thousands of circulars, technical manuals, policy documents and engineering standards into a structured knowledge repository.The effort has since evolved into a broader AI ecosystem centred around Margsarthi, NHAI’s in-house AI assistant trained exclusively on NHAI circulars, Acts, IRC codes and other approved technical documents. Unlike public AI platforms that draw information from the internet, Margsarthi operates within a closed environment and relies only on vetted institutional records, project databases and technical standards.Launched in April, the chatbot has already handled more than 50,000 queries from around 1,100 users across the organisation, according to NHAI. Officials said every response generated by the system is linked to source documents, allowing users to verify the basis of recommendations.The system is currently accessible to NHAI officers across the country and is expected to be extended to authority and independent engineers associated with highway projects.One of its most frequently used functions is resolving technical disagreements at project sites. Instead of manually searching through hundreds of pages of codes and circulars, engineers can query the system and receive an answer backed by the relevant provision.“Many disputes arise because the contractor and the authority engineer interpret a provision differently. Earlier, both sides would spend hours going through manuals and circulars. Now they can put the issue to Margsarthi, which points them to the relevant IRC code or policy provision. Once everyone is looking at the same rule, disagreements are resolved much faster,” Yadav said.The platform is also being used to analyse lengthy legal and arbitration records. Tasks that previously required officers to review hundreds of pages can now be completed in minutes, helping officials identify key issues and prepare responses more quickly.Another application provides engineering guidance from field photographs. Recounting a recent example,Yadav said he uploaded an image of an improperly cut slope while travelling, and the system immediately identified the deficiency and suggested corrective measures based on applicable standards.NHAI is also developing AI tools capable of continuously scanning project data and automatically alerting officers to delays, quality deficiencies and emerging risks. Future applications under development include automated road defect analysis, maintenance planning, AI-assisted drawing reviews, signage planning and support for environmental and forest clearance applications.However, officials stressed that the tools are intended to assist decision-making rather than replace it. While engineers and project directors are encouraged to use AI-generated insights, responsibility and accountability for project decisions will continue to rest with the officers concerned.