Road Safety Audits (RSA) are a critical component of highway project delivery, ensuring that safety risks are systematically identified and mitigated before roads are opened to traffic. Among all audit stages, Stage 3—pre-opening—is one of the most crucial, as it represents the final opportunity to detect design, construction, and operational safety issues.
IRC SP 88 defines a structured framework for conducting Road Safety Audits in India, outlining methodologies, team composition, audit processes, and reporting standards. However, as road networks expand and project timelines tighten, traditional audit methods face challenges in scalability, consistency, and speed.
This raises an important question: can modern technologies such as AI in road safety audits replicate or augment the judgement of certified RSA teams while remaining compliant with IRC SP 88?
Stage 3 audits are conducted just before a road is opened to traffic. At this stage, construction is substantially complete, and the road environment closely reflects actual operating conditions.
The objective is to identify any safety deficiencies that may not have been evident during earlier design or construction stages. These include issues related to geometry, signage, markings, roadside hazards, visibility, and user behaviour.
Unlike earlier stages, Stage 3 audits require a strong emphasis on real-world conditions, including traffic interaction, lighting, and roadside environment.
.png)
IRC SP 88 establishes several key requirements that define how Stage 3 audits should be conducted.
The audit must be carried out by an independent team with certified expertise in road safety engineering. The team should not be involved in the design or construction of the project to ensure unbiased evaluation.
Auditors are required to conduct thorough site inspections under different conditions, including daytime and nighttime. This ensures that safety issues related to visibility, lighting, and driver perception are properly identified.
The audit process follows a structured checklist covering multiple aspects:
This structured approach ensures consistency and completeness.
Auditors must identify potential hazards and assess their severity based on likelihood and potential impact. This requires professional judgement and experience.
The final audit report must clearly document:
The report should be actionable, precise, and aligned with engineering standards.
While IRC SP 88 provides a robust framework, implementing it at scale presents several challenges:
Manual inspections are time-intensive and often cover limited stretches, especially in large highway networks.
Different auditors may interpret the same situation differently, leading to inconsistencies in risk assessment.
Audits are typically conducted at a single point in time, missing evolving risks that emerge after project completion.
Manual reporting can lack geo-referenced evidence, making it difficult to validate findings or track compliance.
Advancements in artificial intelligence are transforming how infrastructure is monitored and evaluated. Technologies such as computer vision for road inspection enable automated detection and analysis of road conditions at scale.
AI systems use cameras mounted on vehicles to capture high-resolution video of road environments. This allows rapid and consistent data collection across large networks.
Using AI-based defect detection, systems can identify safety issues such as missing signage, faded markings, potholes, and roadside hazards.
Through automated road safety analysis, AI systems apply consistent logic to evaluate safety conditions, reducing subjectivity.
AI platforms generate precise, location-based data using GIS-based road inspection, enabling accurate mapping and verification of issues.
The ability of AI to meet IRC SP 88 requirements depends on how well it aligns with each core component of the audit framework.
AI significantly enhances coverage by enabling large-scale surveys in a fraction of the time required for manual audits. This aligns well with the need for comprehensive inspection.
Unlike human auditors, AI applies standardized rules across all observations, improving consistency in identifying and classifying risks.
While AI can detect and quantify issues, replicating nuanced engineering judgement remains a challenge. However, predictive analytics for road safety can support risk prioritization based on historical patterns and severity scoring.
AI excels in generating detailed, geo-tagged, and time-stamped records, improving transparency and auditability.
The most effective approach is a hybrid model where AI augments human auditors rather than replacing them.
AI can handle large-scale data collection and preliminary analysis, providing auditors with structured insights.
Certified RSA professionals can interpret AI outputs, validate findings, and make final decisions.
With smart road monitoring system, roads can be assessed regularly even after opening, ensuring ongoing compliance and safety improvements.
As road infrastructure becomes more complex, the integration of AI will become essential. Technologies such as digital road asset management platforms will enable centralized monitoring, reporting, and decision-making.
AI will not replace auditors but will redefine their role—shifting from manual inspection to strategic analysis and decision-making.
IRC SP 88 provides a comprehensive framework for Stage 3 Road Safety Audits, emphasizing systematic inspection, professional judgement, and structured reporting. While AI cannot fully replicate human expertise, it can significantly enhance the audit process by improving coverage, consistency, and data quality.
The future lies in combining AI capabilities with human judgement to create a more efficient, scalable, and reliable road safety audit ecosystem.
RoadVision AI is building the world’s first Autonomous Road Engineers, combining vision intelligence and language intelligence to transform how road safety audits, condition monitoring, and infrastructure decisions are handled.
Its platform enables end-to-end automation powered by AI road inspection software, delivering geo-tagged, video-backed insights aligned with engineering standards. From detecting safety risks to generating audit-ready reports, RoadVision AI helps RSA consultants and highway departments move from manual processes to intelligent, data-driven workflows.
The Autonomous Road Engineer
AI that is transforming road infrastructure, from reactive to predictive.
.png)
