Continuity from client to revenue: what 450 desktops taught us about BCP

A plan written for a different disaster

On Friday, March 20, 2020, I was running a 450-person operation across Hyderabad, Bengaluru and Delhi, delivering managed services for capital markets and shareholder communications. By Monday, almost all of it had to run from people's homes.

We had a Business Continuity Plan. It was documented, audited and table-top tested. But like most BCPs of that era, it planned location by location: one site goes down through fire, flood, a local lockdown or a geopolitical event, and work shifts to another. Nobody planned for every location, in every country, closing on the same day.

Our BCP was a fire drill for one building. COVID was every building in every city catching fire at once, with the fire brigade told to stay home too.

There was no written plan for moving desktops, no plan for power or connectivity at home, and no project manager. And we were heading into the peak of corporate actions, the season that carries about 90% of our annual volume.

The SEZ rule most people don't know about

Many outsourced operations in India run inside Special Economic Zones, where client data is meant to stay within defined boundaries. Teams on the most sensitive work are isolated further, and some of these restrictions are written into client contracts. Remote work on that data was, by design, a no-go.

So the moment we declared BCP, we pulled together the numbers, presented our data to our risk and compliance owners and our clients, and sought their approval before anyone could start. If your operation sits in an SEZ, your continuity plan carries a constraint most BCP templates never mention.

The first week: what broke, in order

Nothing broke all at once. Each fix revealed the next problem.

1. The desktops. Our floor ran on desktops, not laptops. That meant 450 complete sets of CPU, monitor, cables, keyboard and mouse, each boxed, labelled, logged as an SEZ asset and mapped to one associate. Nobody had ever written down how to do this.
2. The power. Desktops need steady power, and home power cuts are common. We had battery backups, but only a limited number.
3. The people weren't where we thought. On Monday, many associates were in their hometowns, and travel was now restricted. Three offices became more than 20 locations. Some associates collected their sets; for others, we delivered to their hometowns and collected an acknowledgement that each device was working.
4. Remote access. No associate had home access, and the VPN had never been sized for the whole operation. It maxed out almost immediately. Its security key ran on a smartphone, and over half our associates didn't own one; phones weren't company assets, so they had to source their own. With no guide to follow, I wrote the connection SOP on the fly.
5. The data. Teams relied on local servers inside the office, never designed to be reached from outside. That raised harder questions: why was this data held locally, was it backed up, and who owned it? We brought in a Citrix farm, and even that capacity ran out.
6. The real showstopper: no internet. Almost 90% of associates had no internet where they lived. Many stayed in PGs, where broadband simply wasn't part of life. We sourced data dongles and activated SIMs, one location at a time.
7. The ones we couldn't reach. For about 2% of the team, no delivery route existed. Many held key-person responsibilities, so we converted their personal devices into official assets, applying every security control remotely before they touched client work.

Throughout, no one was running it as a project. I picked up one problem at a time while peak-season work waited for people who could not yet log in. Whatever could break, broke.

The next 60 days

Getting people connected was only the start. We work directly with the Street, and our clients still expected their work delivered at the busiest time of the year. The last-mile associates owned every client obligation, so we built SME chat and email channels to support them in real time, and stayed available 24/7 as we moved from nothing, to recovery, to business as usual.

Then came problems no plan had imagined. Hardware faults at people's homes meant engaging vendors on the fly; software faults often meant getting the device back to the office. Every security patch needed the device online and restarted, across 20+ locations, and we had to prove each one was compliant. Once stable, we re-mapped every asset down to the last cable.

Working through a full lockdown also created an always-on culture overnight. Wellbeing had to be designed deliberately, through small cohorts and one-on-ones. COVID reached our teams and their families and touched all of us, so we built skill backups quickly and made sure the company showed its human side while the business kept running.

From the moment we declared BCP, the plan became a live document. We recorded every action, named every unknown and gave it an owner, and appended each gap as we found and fixed it. Once everyone was working, we made it sustainable: a table and chair for every associate, desktops converted to laptops, and stable internet in place of dongles.

The result: no SLA breaches and 100% quality, measured in hard data, in work where a miss has direct market impact. And as some vendors in the chain simply stopped supporting their part, we owned the last leg ourselves, so end customers never felt the gap. The credit for that belongs to our 450 associates.

What the crisis exposed

The logistics were the visible problem. The lasting lessons were gaps that COVID revealed across the industry, in operations built for a world where everyone came to the same floor every day.

The thinnest operations cracked first. The biggest cracks were in vendors, intermediaries and managed operations running on thin margins, thin teams and thin infrastructure. Some went out of business, and their failures rippled straight into their clients. Your continuity is only as strong as the weakest operation you depend on.

Automations and data lived on individual desks. BOTs ran locally with no central server, data sat on local servers, and critical access was often shared. On one floor, that works. Once people scatter, nobody can see what is running where, whether it is backed up, or who owns it. Spotting this made us ask what else was hiding in plain sight, and the answer usually sat in the layers most plans ignore: the associate's unwritten knowledge, the supervision above them, and improvements kept local instead of built into the process.

Mission-critical didn't mean redundant. Critical activities often had no backup in people or systems, and knowledge concentrated in a few key people. A crisis, and especially illness, turns that into a single point of failure.

The view from the top missed the floor. The full picture of how processes connect usually sits with the people doing the work, not with P&L owners, senior leaders or continuity planners. In global organisations, each site owned its piece, but often nobody owned the whole process end to end. A plan built from that view can keep every site running and still let the process fail in the gaps between them.

The lesson: continuity from client to revenue

Most BCPs ask two questions: can our people work, and can the process run? COVID showed that continuity is a chain, and it breaks at its weakest link. A real plan maps every link, from the client all the way to the revenue.

1. 1The client and the contractWhat you have committed to deliver, where, and under what SEZ and regulatory conditions.
2. 2The product and workflowThe work itself, its volumes, peaks and cut-offs.
3. 3The peopleWhere they actually are, what they know that isn't written down, and who backs them up.
4. 4The layer above the peopleSupervision, decisions, escalations and quality checks.
5. 5Automations and dataEvery BOT, server and dataset: where it runs, whether it is backed up, and whether it can be reached remotely.
6. 6Access and securityVPN, Citrix, security keys, credentials, and who is allowed into what.
7. 7Physical infrastructureDevices, power, internet, smartphones, servicing and patching, and a place to sit.
8. 8Locations and ownershipEvery site, and one owner for each process end to end.
9. 9Vendors and intermediariesEvery third party in the chain, and who picks up their work if they fail.
10. 10The end customer and the revenueThe SLAs, the customer's outcomes, and the P&L that depends on everything above.

If any link is missing from your plan, that is where your next crisis will start.

The heroes nobody recognised

The unsung heroes were our IT team: 8 to 9 people, working in shifts to cover 24/7 for 60 days straight. They packed and tagged devices, stretched remote access every time it hit its limits, secured personal devices, patched machines across 20+ locations, and stayed on call for every associate connecting from a hometown hundreds of kilometres away. Continuity stories usually credit leadership. In reality, a small group of IT staff refused to go home until everyone else could work from theirs.

And then there were the 450 associates, who went from no access at all to business as usual at the peak of the year, from homes and PGs never meant to be offices. Leaders, myself included, worked almost the full 60 days without a break. The plan didn't carry us. People did.

Eleven questions to ask of your BCP this quarter

We had done table-top testing. That weekend was the real test: real devices, real people in real places, real client work at peak volume. Most BCPs I review today would pass an audit and still fail it.

1. Does your plan cover every location going down at once? Local and geopolitical scenarios are not enough.
2. Is there a written plan to move, map and track devices out of the building? Who packs, labels, logs and approves.
3. If you operate in an SEZ, do you know which approvals you need, and how to get them fast? Find out now, not on a Friday night.
4. Could your remote access carry your entire critical workforce at once? Check real VPN and Citrix capacity, and whether everyone can use the security key.
5. Is a connect-from-home SOP ready to send? Writing it mid-crisis costs hours you don't have.
6. Do your people have internet and a smartphone where they live? Personal phones are not company assets.
7. Can you service, patch and track devices you can't physically reach? Including personal devices you may need to bring under control.
8. Does every mission-critical activity have redundancy in people and systems? If one person, one illness or one machine can stop it, it isn't protected.
9. Where are your hidden dependencies? Look for desk-level automations, local servers, shared credentials and unwritten knowledge.
10. Who owns each process end to end, and who runs the response as a project? Name them in advance, and plan for the BCP to become a live document.
11. How thin are the vendors you depend on? Know who picks up their work if one of them fails.

If you can't answer three or more with confidence, your plan is probably certified rather than ready.