NHAI turns to AI and open source to preserve institutional knowledge lost to officer transfers
The National Highways Authority of India is using AI and open-source tools to preserve and share knowledge among its officers.
Reported by News · context written by Hushread
In brief
- NHAI has developed an internal AI assistant called 'Marg Sarthi' to help officials access policies, contracts, and project data.
- Frequent officer transfers at NHAI often lead to loss of institutional knowledge and require new officers to rebuild information bases.
- The AI team is converting recurring administrative processes into reusable workflows to make expertise accessible across offices.
- The assistant uses open-source software, allowing faster experimentation, adaptation, and feedback from NHAI employees.
- 'Marg Sarthi' has handled over 200,000 queries since April 2026, with high engagement from more than 1,000 employees.
What happened
The National Highways Authority of India (NHAI) introduced an internal AI assistant named 'Marg Sarthi' to help officials quickly access and understand policies, contracts, and project data. This system was developed using open-source software, enabling the team to experiment and adapt tools to the organisation's needs more efficiently.
Why it matters
NHAI manages large-scale infrastructure projects, and frequent officer transfers can result in important knowledge being lost or siloed. By making expertise and workflows reusable, the AI tools aim to improve continuity, reduce errors, and save time for officials. This could lead to more efficient project management and better use of public resources.
How we got here
Officers at NHAI previously relied on personal folders and had to consult multiple sources for routine and complex tasks. The lack of a central, reusable knowledge base made it difficult for new officers to get up to speed after transfers. The internal AI team began using open-source software to build tools that address these challenges and encourage collaboration between technology staff and engineers.
What happens next
The team is expected to continue refining the AI assistant and developing new tools based on officer feedback. Increased use and further integration of reusable workflows may help standardise procedures and reduce repeated mistakes across the organisation.