➜~cpchander@dev:~$
cpchander@dev:~/blog$ cat covid-1200-employees-office-to-home.mdx
➜ ~ cat covid-1200-employees-office-to-home.mdx
# published May 16, 2026

1,200 employees, 5 days: the COVID remote-work transition playbook

In March 2020 we moved 1,200 employees from a Noida office to their homes in five days — VPN, internet, payroll, attendance, vendor payments, and every department running smoothly from day 6. Here's the operational playbook, day by day, system by system, department by department.

Most "business continuity" content is theoretical. This isn't. In March 2020, with a 12-hour notice from the Indian government's first national lockdown announcement, I had to move 1,200 employees from our Noida office to working-from-home — keeping every client engagement live, every salary paid, every attendance recorded, every invoice sent, every vendor paid.

By the end of the first week, every department was operating from home. By the end of the second week, productivity had returned to within 5% of pre-lockdown levels. Most of those 1,200 people never came back to the office.

This is what we actually did. Day by day. Department by department. System by system. It's the playbook I'd hand to any IT operations leader staring at a similar disruption — pandemic, geopolitical shutdown, natural disaster, anything.

The pre-COVID baseline (what we had to preserve)

Virtual Employee Pvt. Ltd. was a remote-staffing company — paradoxically, all 1,500 of our staff worked from our office in Noida. They provided remote work to clients in the US, UK, Australia, Canada. So our IT operations had two layers that had to keep functioning:

Layer 1 — Our office IT: desktops/laptops, our LAN, internet links, the in-house HRMS, ERP (Microsoft Dynamics NAV / Navision), CRM, Asterisk + OpenSIPS telephony, file shares, internal apps, the help desk.

Layer 2 — Client engagements: each agent connected daily to a client's environment — sometimes via VPN to the client's on-prem systems, sometimes via SaaS tools we managed access to, sometimes via specialized hardware (KVMs, dedicated phones, secure laptops). About 40% of our clients had compliance requirements (HIPAA, PCI, US government contractors) that constrained how agents could connect.

If either layer broke, the entire revenue stream paused.

March 22, 2020 — The trigger

The Government of India announced a 14-hour national curfew on March 22 and a 21-day full lockdown on March 24, effective midnight March 25. We had about 30 hours of working time to figure out the entire operational pivot.

Our crisis response started at 6 PM on March 22 with a meeting between our CEO Narinder Singh Mahil, the operations head, the finance head, and me. The decision was made in under an hour: we keep operating, from home, starting Wednesday March 25. No furlough. No client downtime. No salary delay.

The five-day countdown started that night.

Day -1 (March 23, Monday) — Asset triage and the "what does each person need" matrix

Before any technology, the most important step was an inventory. Within 6 hours of the lockdown announcement, the IT team built a spreadsheet that mapped every employee against four columns:

| Employee | Office hardware needed | Internet at home | Client connectivity model |

The team manager for each function (sales, ops, finance, HR, each client delivery pod) filled this in for their reports. By morning of Day -1 we had a complete picture. Highlights from that data:

  • ~750 employees needed only a laptop + internet — they could work from anywhere
  • ~300 employees needed their desktop + dual monitor setup taken home
  • ~100 employees needed dedicated client hardware (KVMs, secure laptops, IP phones)
  • ~50 employees needed nothing taken home but had concerns about home broadband
  • ~200 employees had no home broadband at all

That last number — 200 people with no home internet — became the single hardest operational constraint of the week.

Procurement war room

By 10 AM on Day -1 I had a procurement war room running. We needed, in the next 60 hours:

  • Additional VPN concentrator capacity — our existing appliances could handle ~400 concurrent VPN sessions; we needed to scale to 1,500+
  • Mobile hotspots — 4G dongles + data plans for the 200 staff without broadband, plus emergency capacity for anyone whose broadband failed
  • Laptop bags / power adapters / lock-and-load kits for the desktop-takers
  • A logistics partner to drive desktops to employees' homes if needed — we ended up using a combination of cab services and the office's own transport vans

Vendor calls all day. Fortinet, our firewall vendor, agreed to ship VPN licenses overnight and stretch our terms. Airtel and Jio sales reps fast-tracked 250 mobile hotspot devices with 14-day delivery cut to 36 hours. Our office facilities partner agreed to use his transport vans through the curfew with proper movement permits.

Day 0 (March 24, Tuesday) — The physical movement

The night before the lockdown was the only window to physically move equipment. By 9 AM on March 24 we had:

  • All employees who could take a laptop home, taking one
  • A schedule for the 300 desktop-needing employees — each slot, who carries what, which vehicle, which neighborhood
  • IT team members at the office assembling and tagging boxes by employee name
  • A standing call running on Microsoft Teams with HR and operations to track who had what

By midnight (the lockdown moment), every employee had:

  • Their working hardware at home
  • Pre-installed VPN client software (configured during Day -1 testing)
  • Their HRMS login credentials reconfirmed
  • A printed crib sheet of phone numbers for IT help, HR, their manager
  • The morning Teams meeting calendar invite

This was the high-anxiety moment. We genuinely didn't know if the systems would handle the load.

Day 1 (March 25, Wednesday) — Lockdown morning

9 AM. The office was empty. The lockdown was active. 1,200 people were sitting at their kitchen tables, dining tables, makeshift desks. The VPN was the first test.

The VPN scaling crisis (resolved by 11 AM)

Our Fortinet VPN concentrators were configured for ~400 concurrent sessions. By 9:15 AM we had 950 connections trying to establish. We had three saving graces:

  1. The emergency Fortinet licenses from Day -1 — already loaded but not active. The IT team turned them on at 9:20 AM.
  2. Split-tunneling configuration — by 10 AM we'd pushed an updated VPN profile that routed only client-business traffic through VPN. Web browsing, video calls, personal traffic went direct. This 5x'd our effective capacity.
  3. A second VPN edge — a backup AWS-hosted VPN we'd set up the night before, ready to take overflow. We activated it at 10:30 AM.

By 11 AM, 1,150 VPN sessions were stable. The remaining ~50 had connectivity issues that turned out to be home-broadband-related, not VPN.

The telephony test

At 9:30 AM the inbound calls started — clients calling for status updates, agents calling clients for daily work, internal calls between teams. Our Asterisk + OpenSIPS telephony platform was already cloud-resilient because we'd designed it that way through years of office-power-outages. SIP softphones on employee laptops registered cleanly to our SIP edge. Outbound international calls went through unchanged. Inbound DIDs ringed agents at home as if they were at the office.

Telephony was the part of the transition we worried about most. It turned out to be the smoothest piece. The investment in SIP-based telephony with OpenSIPS routing — instead of older on-prem PBX hardware — paid off in a single day.

Department-by-department on Day 1

Sales (40 people). Working from CRM (Microsoft Dynamics) over VPN. Outbound calls via SIP softphones. Demo calls via Teams. Zero downtime on the sales floor; pipeline activities continued.

Operations / Client Delivery (~900 people). This was the heart of the company — the agents actually doing client work. By client engagement type:

  • Pure-SaaS clients (~500 agents): immediate — they just logged into the client's tool over the internet, no VPN needed
  • VPN-to-client-network clients (~300 agents): up by 11 AM after our own VPN stabilized; their client VPN sat on top of ours
  • Specialized-hardware clients (~100 agents): up by Day 3 once we'd shipped or driven specific KVMs to homes

Finance (15 people). This was tricky because Navision (Microsoft Dynamics NAV) was on-prem. Finance worked through VPN with strict access controls. Day-end reconciliation that night was 30 minutes late but complete.

HR (12 people). The HRMS we'd built in-house was already web-based; HR could log in from anywhere. Recruitment, attendance reviews, employee queries all moved to Teams calls + HRMS forms. They didn't miss a beat.

IT (the team of about 25). Half of us stayed at the office for the first three nights to manage the physical infrastructure — VPN concentrators, switches, the data center, helpdesk volume spike. The other half worked from home running remote support over Teams + phone.

By 6 PM on Day 1, we'd handled approximately 800 support tickets — most around VPN config, soft-phone setup, "I can't open Outlook from home." A team of 20 IT staff worked from 8 AM until midnight to clear them.

Day 2 (March 26, Thursday) — The systems that broke

Every business continuity plan has a "the second day is when reality hits" moment. Ours was on Day 2 morning when we realized:

Home internet was hugely variable. Some employees had 200 Mbps fiber; others had 4 Mbps copper-DSL that couldn't sustain a Teams call. Our 200 mobile-hotspot allocation — bought for the no-internet employees — had to be expanded to another 150 people whose home internet wasn't enterprise-grade enough. We re-ordered 200 more hotspots that morning.

Power cuts. Half of Noida's residential neighborhoods had 2–4 hour power cuts daily during summer. Employees had no UPS at home. We did two things: (1) shipped small home UPS units to 80 critical-role employees within 48 hours, and (2) instructed all employees to keep laptops charged and have a "go bag" for unexpected outages.

Family environments. Children at home (schools were closed too), spouses also working from home, ambient noise on calls, no dedicated workspace. We added a soft policy: agents on client-facing voice work got first priority on the home's quietest room. HR ran 1:1s with the 50 most-impacted employees to find solutions.

The "where's my desktop" problem. A few employees' desktops, despite the Day 0 logistics, hadn't arrived at homes by Day 2 morning. We scrambled cabs to deliver overnight. Five employees worked from their phones for 24 hours.

By Day 2 evening, however, the support ticket volume had dropped from 800 (Day 1) to ~250. The biggest issues were behind us.

Day 3 (March 27, Friday) — Payroll execution

Day 3 was the moment of truth for the payroll function. Salaries for ~1,500 staff were due on the 30th. Normally finance ran payroll from the office, with HR validating attendance from the office, and HR-coordinated bank file transfers from the office. None of that physical infrastructure was available.

Here's what we did:

Attendance. Our in-house HRMS already tracked attendance per agent, integrated with Asterisk CDR for call-handling agents and with the VPN logs for back-office staff. So even though physical badge swipes at the office no longer happened, digital attendance was actually more accurate, not less. Attendance for March was complete in the HRMS by Day 3 lunchtime.

Payroll calculations. The HRMS pushed attendance + leave + overtime to a payroll spreadsheet that finance ran through Navision (Dynamics NAV). The on-prem Navision was VPN-accessible. The finance manager ran the monthly payroll batch from her home over VPN by Day 3 afternoon.

Bank file generation. Navision generated the BACS-equivalent bank file for our payroll bank. The file was usually couriered as a USB to the bank or uploaded from an office PC. We'd never run this from outside the office. We tested an approach with the bank's relationship manager: upload via the bank's corporate portal from finance head's home connection, with the file signed off by two authorized signatories via Microsoft Teams screen-share before upload. The bank accepted it.

Salary disbursement on March 30. All 1,500 salaries hit bank accounts on time. Not delayed by a day. No mistakes — finance reconciled each row before the file went up.

This was the single moment that defined the transition. Employees who'd seen colleagues at other companies miss salary by a week or two trusted the company more after March 30 than they had before. Years later I'd hear people refer to it as "the salary that came on time during lockdown."

Day 4 (March 28, Saturday) — Vendor payments and invoicing

On any normal weekday, finance:

  • Sent out client invoices via email (some via client-portal upload)
  • Processed vendor invoices coming in
  • Paid vendors via NEFT/RTGS through the corporate banking portal
  • Reconciled cash positions

All of this had to keep working remotely.

Client invoicing moved cleanly. Invoices were generated in Navision; PDFs were emailed via Exchange Server (also accessible over VPN). Client-portal uploads worked from home if the agent doing the upload had VPN to the client's network.

Vendor invoice intake moved to a shared mailbox + a vendor-invoice-tracker workbook in SharePoint. Vendors had been emailing invoices for years already; nothing structurally changed.

Vendor payments were the trickier piece. Our corporate banking portal required a hardware token + dual approval. We solved this by:

  • Issuing extra hardware tokens to finance department's home offices
  • Setting up Microsoft Teams as the dual-approval channel — the proposed payment was screen-shared, the second signatory verbally approved, then physically clicked approve on their own portal session

By Day 4 evening, ~80 vendor payments had been processed without delay.

Day 5 (March 29, Sunday) — Catch-up, monitoring, recalibration

Day 5 was the first day where IT support tickets dropped to a normal volume (~50). The transition crisis was largely over. The remaining work was:

  • Monitoring. We'd deployed our employee monitoring system (integrated with HRMS) to track productivity on remote machines. By Day 5 we had real data showing productivity was at ~94% of pre-lockdown baselines. Within two weeks it would return to 99%.
  • Remote firewall posture for client-confidential work. ~50 agents working on US-government-contractor or HIPAA workloads had stricter requirements. We pushed standardized home-firewall configs (small Fortinet 30E units couriered to those homes) for the strictest client environments.
  • Communications cadence. We instituted a 9 AM daily all-IT Teams call, a 6 PM department-heads call, and a weekly Friday town hall with the CEO addressing the entire company. The communication overhead was real but kept the company aligned.

What I'd do differently

In hindsight, three things would have been better:

1. Practice before you need it. We had business continuity plans on paper. We'd never run a drill. A two-day BC drill once a year would have surfaced the VPN capacity problem before it was a crisis, not during.

2. Buy more buffer capacity upfront. The VPN licenses we paid premium-emergency-shipping rates for could have been on standby for 5% of the cost as cold spare. Same with mobile hotspots, UPS units, headsets.

3. Invest in home-office stipends earlier. Within a few months we instituted a home-office allowance for chairs, desks, lighting. Some of that should have been pre-positioned with critical-role employees as part of standard onboarding.

What this story is actually about

This is the case study I lead with when I talk about IT operations leadership. Not because the technology was novel — it wasn't. VPN scaling, mobile hotspots, payroll-over-VPN, telephony-over-SIP — all standard. The story is about operational discipline.

The transition worked because the foundation had been built over the previous decade:

  • The HRMS was already web-based (built ~2012)
  • Telephony was already cloud-resilient over SIP/OpenSIPS (re-architected ~2014)
  • ERP (Navision) was accessible over VPN (configured ~2016)
  • Email was on hybrid Exchange — accessible from anywhere (deployed ~2017)
  • Our employees were already trained on Microsoft Teams (rolled out 2019)
  • Our IT team was practiced at handling outages (years of operational scar tissue)

If any one of those foundations had been weaker, the 5-day transition wouldn't have worked. The crisis didn't reward heroes. It rewarded preparation that had been done years before, often when it looked like over-investment.

I think about this every time I'm building IT operations for a new venture. The systems you build in calm times are the only systems that work in chaos. And the operations team you trust on the boring days is the team that saves you on the impossible ones.

Cross-references

← back to /blog