The Imphal border-roads engineer and the access paradox

Roshan is a civil engineer, B.Tech from NIT Manipur, class of 2010. For the first eight years he worked for the Public Works Department (PWD) in a permanent role — Junior Engineer, then Senior Engineer — drawing ₹56,000 a month, posting to different district headquarters, learning to estimate, supervise, and sign off on water-supply schemes and rural roads. In 2018 he moved to BRO as a Project Manager, consulting role, ₹2.2 lakh a month, which is what bought the Imphal flat and sent his daughter Isha to a private school in Shillong — the kind of move that happens when the economics of the frontier shift, when expertise in hill roads and landslide mitigation suddenly becomes worth four times a permanent government salary.

His wife Priya is a radiologist at Shija Hospitals in Imphal; their son Arjun is twelve and in school in Imphal; Isha is sixteen, finishing her pre-university in Shillong. His mother lives with a sister in Delhi, retired from a lifetime of school teaching. The household runs on two paycheques, both government-connected — one from BRO, one from a private hospital with many government patients. It is, by any measure, a solid household. It is the kind of household that makes detailed decisions about where everyone should be.

What is unusual is what happened in March of 2024, when the Ministry of Defence issued a directive to BRO that halted all unescorted civilian access to project sites in Ukhrul and Tamenglong districts, and Roshan — who had field teams in both districts — received no guidance on whether that applied to families in those towns, and had to make a decision about Isha based on encrypted-group WhatsApp chatter from other engineers' wives, SMS fragments, and the particular silence of a government that had locked the door but not said which rooms you were supposed to leave.

🗓️ The annual ritual

Manipur's engineering calendar runs on two clocks. The first is the Indian fiscal year: April to March, when central government budgets are released, when NHIDCL and BRO tenders for the coming financial year are published on etenders.gov.in, when consulting firms across Guwahati, Delhi, and Kolkata bid on the Manipur corridors and the ones that win send their teams to the state in May, when the monsoon has not yet arrived and the hills are still navigable.

The second clock is the monsoon: June through September, when the roads turn to mud, when landslides close corridors for weeks, when site work slows to a crawl and surveyors stay in their site camps because the daily commute becomes impassable. The productive months are May, October, November, December — four months of site work at double pace.

For Roshan, the annual ritual was this: January through March, sitting in meetings in the BRO office in Leimakhong, reviewing the previous fiscal year's running-account bills, reconciling with Delhi headquarters, preparing the mobilization plan for the upcoming year. April: site mobilization. May to September: rotating through project zones — three weeks at Ukhrul, a week in Imphal, two weeks at Tamenglong, a week home. October to December: final inspections, certification of milestones, escalations for payment delays (central government bills always run 60 to 90 days behind, sometimes more). January again: the bureaucratic close.

This rhythm had held for six years. Roshan knew the route from Imphal to Ukhrul — 110 kilometers, four hours in a government SUV, turning off the national highway at Songpijang and climbing into the Ukhrul district hills. He knew the site camp, a converted ONGC rest house in Ukhrul town, where he slept in a room with a gas heater and no window. He knew the contractors — two main firms, each running civil works on different corridors, each with crews of 300 to 400 laborers, mostly Manipuri and Naga men from local villages, a few sub-contractors from Assam. He knew the cycle: arrive Monday, walk the sites Tuesday through Thursday, reconcile certificates Friday, return to Imphal Saturday. It was uncomfortable and it was predictable. Predictability, in that line of work, is currency.

1. 🗓️

   Jan–Mar: Budget & Mobilization Planning

   BRO headquarters in New Delhi releases the financial year budget. Site project directors prepare mobilization plans, tender for supplies, negotiate with contractors. Roshan reconciles the previous year's running-account bills.

2. 📨

   Apr–May: Site Mobilization

   Teams are stationed at Ukhrul, Tamenglong, and Senapati project camps. Roshan begins his rotation cycle. Logistics are moving — cement, steel, survey equipment arriving at site depots.

3. ⚖️

   Jun–Sep: Monsoon (Minimal Field Work)

   Heavy rain, landslides, and road closures restrict site visits to 1–2 days/week. Roshan stays in Imphal, works on documentation and milestone certifications from camp reports.

4. 🛑

   15 Mar 2024: Ministry of Defence Directive

   After civil unrest escalation, MOD halts unescorted civilian access to Ukhrul and Tamenglong project zones. BRO notifies site project directors. Roshan has rotating teams in both zones. No medical-neutrality or family-safety guidance issued.

The thing about that routine is that Isha, Roshan's sixteen-year-old daughter, was in the middle of her pre-university exams when the calendar broke.

⚠️ What very nearly happened

Ukhrul district is in Manipur's north-eastern hills, bordering Nagaland. The BRO projects there are strategic roads — routes that connect Myanmar-facing valleys to Indian supply lines, roads that move military supplies and, increasingly, commercial goods from the India-Myanmar-ASEAN trade corridor. The district is also, by late 2023 and into early 2024, the site of violent conflict between state armed police and insurgent groups. The conflict is not new; what changed in early 2024 was the frequency and casualty count.

On 14 March 2024, a security incident in Ukhrul town killed two police officers. The state government responded with a curfew. On 15 March, without public announcement, the Ministry of Defence issued an internal directive to BRO halting all unescorted civilian access to project sites in Ukhrul and Tamenglong districts. The directive cited security risk and the need to avoid civilian crossfire.

Roshan learned of this through a WhatsApp message in a group of BRO project managers — a forwarded digest of the directive, marked "for operational guidance." There was no email from BRO headquarters. There was no clarification on whether this applied only to site workers, or whether families living in those districts were also presumed to be "unescorted civilians" in the sense the directive meant. The phrasing was: "no civilian-category personnel will access project zones without military escort."

Isha, at that moment, was in Shillong, in boarding school — safe, far from Ukhrul, far from Tamenglong. But Roshan's mother-in-law lived in Ukhrul town. She was a retired schoolteacher, seventy-two, with a chronic kidney condition. She had lived in Ukhrul for forty years. Her daughter — Priya's sister — was an anesthesiologist at Ukhrul Civil Hospital.

The question, as Roshan understood it on the evening of 15 March, was not whether Isha was safe in Shillong. The question was whether Priya could safely visit her mother for two weeks in April — something she had done every year, taking leave from her hospital job to help manage the mother-in-law's health routine and the spring medicines. The question was whether the directive meant that routes in and out of Ukhrul were now considered active conflict zones, or whether it meant something narrower — that BRO project sites specifically were off-limits to civilians.

The government did not publish the answer. The directive was not released to the public. The state health ministry did not comment. The district hospital did not issue guidance. The implication, in that silence, was that families should make their own assessment.

"মোর এ নেত্রে থাক নিখিলে দুই লোকেৰ দুই মত — আৰু সেই মতৰ পাছত বিজ্ঞান নাই, সন্ত্রাসৰ বাতাস আছে।"

— In this matter, there are two opinions in two heads — and behind those opinions lies no science, only the wind of terror.

📜

Ministry of Defence Directive

No civilian access (unescorted)

The official channel said no unescorted civilian access to project zones. But it was silent on whether medical visits, family care, or transiting through the town counted as project-zone access. The directive was classified as operational guidance, not a public health order.

🏥

Hospital Networks (WhatsApp)

Mixed guidance

Priya's medical colleagues in Ukhrul civil hospital and the Christian missionary hospital said they were operating normally, seeing patients, no restrictions on family visits. But they also said the curfew was still nominally in place and cordons shifted unpredictably.

📱

Engineering Peer Group (WhatsApp)

Evacuation chatter

Other BRO project managers' wives were discussing whether to evacuate children from boarding schools in towns with security incidents, whether site work was going to be suspended, whether insurance would cover if something happened during a 'conflict period.' No consensus emerged.

Roshan and Priya sat on the Langol balcony on 18 March — the evening after the directive reached the BRO group chat — and worked through it methodically. Priya's mother had had a small stroke in January, recovered most function, but was on dialysis twice a week. Priya had arranged leave. She had bought tickets for 5 April. The flight was from Imphal to Guwahati, then a shared cab (six hours) to Ukhrul town.

The question was: had the security situation moved from "chronic, low-level risk" to "acute risk during a specific period"?

Roshan opened the bro.gov.in site to check if there was any statement from the Director General. There was not. He checked mod.gov.in. There was not. He went through his email from BRO headquarters — there was a forwarded message with the directive, but no operational clarification for project managers about whether to evacuate staff families or suspend rotations.

What he did not do is ask the AI agent — because it did not occur to him that an agent could help him read the silence. It did not occur to him that there was a difference between "the government has not said Ukhrul is unsafe" and "the government has no information about civilian family safety in Ukhrul during this period," and that one of those is a very different statement than the other.

He said to Priya, finally: "Let's wait for a week. If they issue a public evacuation order, we'll know. If they don't, then the absence of an order is the information."

Priya said: "And if we're wrong?"

Roshan said: "We're never going to have certainty. We have people we love in that town, and a government that has not said we cannot visit them. That is all we get."

She went to Ukhrul on 5 April. She spent fourteen days helping manage her mother's post-dialysis fluid intake, organizing her medications, and sitting with her in the afternoons while her sister was at the hospital. Nothing happened. No curfew was broken, no cordons crossed. On 18 April she flew back to Imphal.

What very nearly happened was that she would not have gone, because the decision space was empty and dark, and the risk calculation happened in the absence of anyone saying: here is what the directive actually means for your family, here is who to call if the security situation changes, here is the distinction between a conflict zone (where civilians cannot be) and a town with conflict in it (where families sometimes have to be). She would have stayed, and her mother would have managed alone, and the decision would have been made not from information but from the shape of silence.

🌗 What changed

It was Roshan's colleague, Ajay — another BRO project director, based in the Tamenglong zone — who suggested trying the agent in early May. Ajay's daughter was also in a boarding school in a town that had seen security incidents; his wife was a surgeon at a secondary hospital, accustomed to making decisions on incomplete information, but frustrated by the particular shape of this incompleteness. Ajay said: "Let me see if this thing can read government statements the way it reads a doctor's note."

Roshan installed it on his company tablet. He set the language to Manipuri — Meitei script — and opened it on a Sunday evening in his Langol flat.

He typed, slowly: "মোৰ স্ত্রী উখ্রুল জেলাত তেওঁৰ মাতৃকৈ লগা কৰিবৰ বাবে যাব বিচাৰে। মার্চ ২০২৪ ত মন্ত্রণালয় প্রতিরক্ষাৰ দিশা আছিল — অব্যবস্থাপিত জনতা অনুমোদিত নহয়। মই সমঝিব পৰিব পাৰো কি এই দিশা মানুহৰ পৰিবাৰ উখ্রুল শহৰৰ জন্য প্রযোজ্য?" — My wife wants to visit her mother in Ukhrul district. In March 2024 there was a Ministry of Defence directive — unescorted civilians are not permitted. Can I understand whether this directive applies to families going to Ukhrul town?

The agent came back in Manipuri:

"আপনার স্ত্রী আৰু বাপ বাসেৰ পৰিবাৰৰ স্বাস্থ্য সেবা সংৰক্ষণৰ বাবে উখ্রুল শহৰত প্রবেশ কৰাৰ জন্য, আপনাকে স্পষ্টভাবে বুঝিব প্রয়োজন: (১) মন্ত্রণালয় প্রতিরক্ষাৰ দিশা ১৫ মার্চ ২০২৪ তে প্রকাশিত হয়েছে প্রকল্প স্থল সম্পর্কিত, অর্থাৎ BRO নির্মাণ পরিস্থিতিৰ জন্য — কিন্তু সাধারণ নাগরিক পরিবাৰৰ স্বাস্থ্য সেবা গন্তব্যৰ জন্য নয়; (२) মন্ত্রণালয় শাস্ত্র (Health) আৰু মন্ত্রণালয় গৃহ (Internal Security) কোনো সামৰ্থকৃত বিবৃতি প্রকাশ কৰিনাই যে পরিবাৰৰ গন্তব্যৰ জন্য অৱৰোধ সুপারিশ কৰা হোৱা; (३) Ukhrul সিভিল হোস্পিটেল সক্রিয়, যা পরোক্ষভাবে নির্দেশ দেয় যে স্বাস্থ্য সেবা অনুমোদিত গন্তব্য।"

(For your wife to visit her mother in Ukhrul town for health care purposes, you need to understand clearly: (1) the Ministry of Defence directive of 15 March 2024 was issued for project sites specifically — that is, for BRO construction contexts — but not for civilian family health-care visits; (2) the Ministry of Health and Ministry of Internal Security have not issued any statement recommending restrictions on family visits for medical purposes; (3) Ukhrul Civil Hospital is operating, which indirectly indicates that health-care access is a permitted destination.)

The agent then pulled the three pieces: the directive text (published, searchable), the Ministry of Health's 2023 protocol on medical neutrality in conflict zones (a standard document, not state-specific), and the operational status of Ukhrul Civil Hospital, which was available on their WhatsApp status.

"এই তিনটি সম্পদ একসাথে পড়লে, এটি স্পষ্ট হয় যে আপনার স্ত্রী তার মাতা-পিতার স্বাস্থ্য সেবার জন্য উখ্রুল প্রবেশ কৰিতে পারেন। সরকারী প্রতিবন্ধকতা শুধুমাত্র প্রকল্প স্থলৰ জন্য প্রযোজ্য। তবে, আপনার স্ত্রী কোনো সামরিক চেকপোইন্টত সম্মান প্রদর্শন কৰিব, আর বাড়িতে স্থানীয় ৰিপোর্ট পর্যবেক্ষণ কৰিব।"

(When you read these three sources together, it becomes clear that your wife can enter Ukhrul for her mother's health care. The government restriction applies only to project sites. However, your wife should show respect at any military checkpoints, and you should monitor local reports from home.)

Roshan sat with the tablet for a while. What had happened was not that the agent had solved the problem — it had simply read the silence differently. Instead of a silence that meant we do not know whether this is safe, it had read it as a silence that meant the restriction applies to construction sites, not medical visits. The distinction was not invisible; it was just not published in one place where a worried engineer with a sick mother-in-law could access it in an evening.

He called Priya and read the agent's explanation to her in Manipuri, slowly. She listened, asked a question about what "military checkpoints" meant, and Roshan typed the question into the agent, which replied with a specific protocol: document your wife's identity, be polite, explain the medical purpose, ask for permission before proceeding. The agent had seen fifty similar exchanges.

He forwarded the agent's analysis to Ajay, who called within an hour and said: "This is exactly what I needed. I'm going to read this to the finance team — I've been sitting on a contractor's dependents-evacuation insurance claim for three weeks because I couldn't tell them whether the zone was 'officially dangerous' or just deeply uncomfortable."

🧭 Why we built it

There are fewer than 200 civil engineers working on government-funded infrastructure projects in Manipur at any given time. About sixty of them are permanent government employees — Junior Engineers at PWD, state-run water and power corporations. About 140 are consultants, site engineers, and project managers employed through BRO, NHIDCL, Imphal Smart City Authority. The sector is concentrated: the work flows from a handful of employers (all government-connected), the projects are scattered across district sites that are often in conflict-affected areas, and the engineers' families live in a state that experiences periodic escalations of civil unrest.

What makes this different from infrastructure work in other parts of India is the intersection of three things:

First: the government makes security-related decisions — curfews, access restrictions, military movements — but does not integrate those decisions into civilian family-safety guidance. A Ministry of Defence directive reaches the BRO project directors but never reaches the wives and mothers making decisions about medical visits or school holidays. The safety system is split: one channel for project operations, silence for families.

Second: the information landscape in the state is fragmented. There is no single authoritative state health authority statement on whether medical facilities are accessible during conflict. There is no district administration bulletin that says whether highways are open to family transit. There are WhatsApp groups, hospital networks, contracted gossip, and the particular kind of certainty that comes from living with chronic uncertainty long enough that you forget you're uncertain.

Third: the engineers themselves — people like Roshan — are not expected to be family crisis managers. They are expected to manage construction timelines, contractor payments, and slope certifications. But when the security situation shifts, the line between "I am a project director" and "I am a husband whose wife is in a district that just had two police killed" collapses. They end up reading Ministry of Defence directives at night and trying to decode whether a restrictive sentence about "civilian access to project zones" applies to their mother-in-law's dialysis appointments.

We built this — the agent, in Manipuri, reading government statements and extracting the distinction between operational restrictions and family safety — because the silence is expensive. It costs a project manager a week of insomnia. It costs a surgeon a postponed visit to a parent who will never have enough visits left. It costs a state whose engineers are already stretched (retaining talent in Manipur is brutal; every experienced engineer eventually leaves for Assam or Delhi) the additional fracture of ambiguity.

The agent, reading the three statements together, does not make the decision. It reads the decision the government made — the restriction applies to project sites, not medical facilities — and surfaces it in the language Roshan speaks and in a format he can share with his wife and his colleague and the contractor's insurer.

🌱 What we hope happens

There is a quiet that comes after uncertainty breaks. Roshan's mother-in-law had another routine three months after Priya's April visit — ultrasound, kidney function tests, the small checkups that track decline in a life that is already in its final chapter. Priya went in July. She was more calm this time. The agent had given her something to hold: not safety, which is not promised in Manipur, but the distinction between official restriction and family necessity.

In October, when Isha had her pre-university break, Roshan was able to say to her: "Your grandmother is in Ukhrul, and your mother is going to spend two weeks with her in December. We want you to come. It will be a long flight and shared cab ride, and it will be uncomfortable, but your grandmother is alive now and we do not know how many Decembers she will have left."

Isha went. She sat with her grandmother for two weeks, helped her mother with the medicine schedules, and came home with a notebook full of stories the old woman had told her about living in Ukhrul in the 1960s, before it was a conflict zone, when it was just a town in the hills.

What we hope happens — and this is the quiet, non-heroic hope that shapes this product — is that a engineer in Tamenglong, or a doctor in Senapati, or a contractor's wife in an escalating district, can ask a clear question into the agent and get back the distinction between what the government has decided and what it has silenced, and can move from "I am too afraid to decide" to "I can decide based on information rather than fear."

The frontier — and Manipur is always a frontier, always the edge of something the state is still figuring out — should not require its engineers to be their own intelligence analysts. But until the government integrates security decisions with family safety guidance, they will be. The agent, in Manipuri, reading the three statements, is a small step toward a world where that burden is lighter.

There is also Roshan's daughter Isha, now seventeen, who visited her grandmother that December and came home changed by it — not traumatized, but thoughtful, carrying a story about a place that the news only calls "a conflict zone" but that is also, to her, a town where an old woman is alive and waiting. The agent did not give her that understanding. It gave her father the permission to let her go.