Havish M Consulting
Customer Case Study · Project Control Room
Highway Infrastructure (EPC) × Havish M Consulting

From site report to boardroom, every kilometre in view

A leading Indian EPC contractor builds highways, expressways and rail works across India. HMC built a project control room for leadership and site teams: GenAI reads daily progress reports, voice notes and site photos; progress is tracked by chainage against plan; fleet and fuel are watched machine by machine; tenders are read in hours, not days; and site safety is checked from photos.

SectorHighway Infrastructure (EPC) FunctionProjects & Operations ThemeProject Control Room · GenAI
Schedule slippage flagged earlier
0weeks
▲ from the daily report, not the monthly review
Fleet utilisation
0%
▲ from about 52%
Fuel variance against norms
0%
▼ from about 14%
First read of a tender document
0hours
▼ from about 3 days

Figures indicative; validated metrics available on request.

The Challenge

Many sites, one view that came too late

Leadership saw progress, machines, bids and safety in separate reports, weeks after the fact.

Progress in WhatsApp

Daily reports arrived as texts, voice notes and photos; the monthly review was the first consolidated view.

Idle machines, missing diesel

Telematics existed but nobody compared hours, idle time and fuel issued across sites.

Three days per tender

Each RFP ran to hundreds of pages; risky clauses surfaced late in the bid cycle.

Safety after the incident

Observations were logged on paper; patterns across sites were invisible.

HMC Approach

Built with site, plant, tendering and HSE teams

One control room, four lenses, starting from data the teams already produced.

Select a step to see what happened.

Discover · What the CMD needs every Monday

Interviews from site engineers to the CMD's office.

  • How daily progress reports are written and sent.
  • Telematics, fuel issue registers and hire costs.
  • The tender review checklist and recent bid post-mortems.

Deliverable: Decision map and data inventory.

Design · Chainage as the common key

Every record tied to a project, a chainage and a date.

  • Activity sequence and weights per pavement layer.
  • Fuel norms by machine type; utilisation on available hours.
  • Clause risk rules agreed with the tendering head.

Deliverable: Data model and rule book.

Build · Four modules, one control room

GenAI where reading is the bottleneck, rules where the numbers decide.

  • Report, voice and photo reading into structured progress.
  • Fleet, fuel and redeployment analytics on a map.
  • Tender reader, estimate and bid score; photo safety checks.

Deliverable: Control room prototype.

Adopt · Three pilot projects

Site teams kept sending reports the way they always had.

  • Weekly CMD summary generated from the control room.
  • Plant head reviews fuel flags every morning.
  • Bid committee starts from the tender reader output.

Deliverable: Pilot and playbook.

Scale · Every project, every machine

Rolled out across road, rail and new verticals.

  • All projects on the same chainage model.
  • Telematics feeds for the full fleet.
  • Safety scorecard in the monthly board pack.

Deliverable: Roll-out plan.

The Solution

Open the control room

Four working modules on synthetic projects, machines, tenders and sites. Try the presets or bring your own report, telematics file, RFP or site photo.

✦ Connect GenAI

Optional. Without a key every module runs on presets and built-in rules, and you can type or tag things yourself. Add your OpenAI key to read site photos and voice notes, extract any RFP, check your own safety photos and write the summaries.

Rules
Your key goes only from this browser to OpenAI.
01 · Highway progress tower

Planned versus actual, kilometre by kilometre

A time-chainage chart shows where each crew is: every line is one pavement layer moving along the road. Move the status date, read a daily progress report, and see the forecast finish and the weekly summary for the CMD.

Rules

Project

Daily progress report

What the report says

ActivityFromToLengthAgainst planNote
Pick a report and select Read report.

The summary combines status, forecast and the latest report.

Synthetic projects, chainages and schedules. Activity weights and rates are illustrative.

02 · Equipment fleet & fuel

Every machine: working, idle or leaking diesel

Last seven days of telematics for 48 machines across seven locations. Fuel issued is compared with the norm for the hours worked; under-used machines are matched to sites that are short of them.

Fuel by day

Redeploy before you hire

Machines ·

MachineSiteUtil.IdleFuel var.Flags

Draft the morning note on fuel, idling, breakdowns and moves.

Synthetic fleet. Fuel norms in litres per working hour by machine type; idle burn at 35% of the norm; diesel at ₹92 a litre.

03 · NHAI tender reader

Clauses, risks, quantities and a bid score

Pick a sample RFP or upload one. Key clauses are pulled out and rated; the bill of quantities is priced from the rate library, and a bid / no-bid score combines risk, margin, fit and capacity. Click a clause to find it in the document.

Rules

Key clauses ·

ClauseFoundRiskWhy

Estimate from the bill of quantities · rates editable

ItemQtyRate ₹₹ crore

The note summarises clauses, estimate and score.

Sample RFPs are synthetic extracts, not real NHAI documents. Rate library, net worth and order book are illustrative.

04 · Site safety from photos

Hazards marked on the photo, incidents sorted, sites scored

Pick a site photo to see the hazards found, or upload your own. Without a key you can drag a box on the photo and tag the hazard yourself. Incident notes are sorted by type and potential severity, and every site gets a safety score.

Rules
Drag on the photo to tag.

Findings

Incident and near-miss notes

By type · last 90 days

Site safety scorecard

SiteOpen obs.High-pot.TrainedScore

Preset photos are drawn illustrations with pre-marked findings to show the output. Incident notes are synthetic.

Impact Dashboard

Closer to plan, less diesel lost

With slippage visible from the daily report and fuel compared every morning, projects tracked closer to plan and fuel variance fell at every site.

Eight projects, planned and actual progress

S-curve small multiples: cumulative % complete by month

PlannedActual

Fuel variance by site and month

Heat map: % of diesel issued above the norm

Figures indicative; validated metrics available on request.

Before vs After

How the week runs now

DimensionBeforeAfter
Daily progressTexts, voice notes and photos in group chatsRead into activity, chainage and length the same day
Schedule viewMonthly review, % complete per projectTime-chainage chart with forecast finish
FleetHire first, check idle machines laterRedeploy under-used machines before hiring
FuelIssue registers reconciled at month endVariance against norm flagged every morning
TendersThree days to a first readClauses, risks, estimate and score in hours
SafetyPaper observations, counted monthlyPhoto checks, sorted incidents, site scorecard
Tools & Tech Used

The control room stack

OpenAI GPT (vision) reports, photos, RFPs Whisper site voice notes Equipment telematics hours, idle, location Leaflet maps sites and machines D3 time-chainage and S-curves Power BI board pack
Key Takeaways

What any EPC contractor can borrow

01

Let sites report their way

Reading texts, voice notes and photos beats asking engineers to fill another form.

02

Chainage is the common key

Progress, machines, quantities and incidents all line up once tied to a kilometre.

03

Redeploy before you hire

The cheapest machine is the one already idle on another site.