Intelligent Data Capture & Schedule-Linking Layer for Infrastructure Project Management: Real-Time Actual Progress Tracking (Planning-to-Execution Bridge)
“Large infrastructure projects like oil refineries have detailed plans, but site workers report progress using messy, inconsistent notes that don't match the official schedule. This system acts as a digital translator that reads these messy updates and automatically updates the master project plan in real-time.”
Deconstructed Problem Blueprint
WHAT TO BUILD?
An intelligent middleware platform that acts as a bridge between unstructured field data (voice, text, spreadsheets) and structured project management software (Primavera/MS Project).
WHO IS IT FOR?
Primary: Site supervisors and contractors. Secondary: Project managers and planners at Oil India Limited.
WHY DOES IT EXIST?
Current reporting is manual, fragmented, and disconnected from the master schedule, leading to 'data silos' where project managers don't know the true status of work until it's too late to fix delays.
INPUT → PROCESS → OUTPUT ARCHITECTURE PIPELINE
Daily progress reports (PDF/Word), discipline-specific spreadsheets (Excel/CSV), scanned site diaries (OCR), and voice-to-text inputs from supervisors.
LLM-based entity extraction, fuzzy string matching against WBS (Work Breakdown Structure) codes, confidence-scoring, and automated API-based schedule synchronization.
Real-time dashboard updates, automated progress alerts, discrepancy reports for planners, and a structured historical database of project execution patterns.
CONSTRAINTS
- Must handle offline-first data entry for remote sites.
- Must support fuzzy matching for non-standardized industry terminology.
- Must maintain a strict audit trail for every automated schedule update.
- Must work with synthetic data since live project data is restricted.
JURY METRICS
- Fuzzy matching accuracy (Precision/Recall).
- Time-to-update (latency from report submission to schedule reflection).
- User friction score (number of clicks/seconds to log an activity).
- Percentage of unmatched activities flagged for manual review.
CORE BOTTLENECK
The 'Semantic Gap': Mapping colloquial site language (e.g., 'finished welding joint 4') to formal, rigid WBS codes (e.g., 'ACT-PIP-004-WLD') with high confidence.
Why You Should Choose This PS (Pros)
- High industrial relevance for Oil & Gas sector.
- Clear, measurable problem with a direct ROI for the organization.
- Strong potential for AI/LLM integration to show technical depth.
Challenges & Risks to Watch (Cons)
- Requires deep domain understanding of project management (Primavera/WBS).
- Risk of 'hallucination' in LLM-based mapping if not properly constrained.
- Integration with legacy PMIS software can be complex.
Recommended Tech Stack & Grand Finale Features
Killer Features to Impress the SIH Jury
View Official SIH 2026 Portal Raw Text▼
? Actual progress data is fragmented, delayed, and inconsistently structured across disciplines and contractors.
? Manual reconciliation with the baseline schedule is slow, error-prone, and often lags the schedule update cycle by days or weeks.
? Downstream performance analytics, delay/ risk analysis, and forecasting inherit this poor-quality, late data- undermining the AI performance-monitoring stack that depends on it.
? Once a project closes, the hard-won knowledge of what actually happened
? Offer an LLM-based conversational or voice interface ('time agent') for site supervisors across disciplines to log activity start/end with minimal friction, replacing rigid manual forms while still producing structured output.
? Fuzzy-match and link extracted discipline-specific activity descriptions to the correct L5/L6 plan node,handling terminology differences and granularity mismatches, and flag unmatched/new activities for planner review rather than silently dropping them.
? Auto-update actual start/end dates in the schedule/PMIS in near real time, with a confidence score and audit trail per entry.
? Produce a clean, structured, discipline-tagged actual-progress dataset that serves two purposes: (a) live input for performance analytics, delay/risk pattern discovery, and forecasting, and (b) a foundation for institutional memory building
A working prototype demonstrating ingestion of 2–3 varied input formats (e.g., a free-text daily report and a discipline spreadsheet), extraction, and schedule-linking logic would be ideal; full production-grade OCR/ASR is not required.