Seven weeks ago the received wisdom in business process outsourcing was that the work could not be done from home. Security required a controlled floor. Productivity required supervision. Quality required a team leader who could see the screens. Clients would never permit it, regulators would not accept it, and the infrastructure did not exist. Then it happened anyway, in about ten days, across every delivery geography at once. Accounts payable, payroll processing, collections, claims, reconciliations and a large share of voice work are currently being performed from bedrooms and kitchen tables in Manila, Bengaluru, Cairo, Kraków and Dubai. The interesting question is not how the industry managed it. It is what the speed of the change reveals about what we believed.
Three categories, three very different outcomes
Seven weeks in, the pattern is clear enough to describe, and it has almost nothing to do with technology. Transaction work with system-recorded queues moved in days. Invoice processing, payment runs, data entry, cash application, ticket handling. The work arrives in a system, is allocated by a system, and is recorded by a system. Move the person and the work follows. In several operations I have looked at, throughput on this category is currently above its pre-crisis level. Judgement work with informal escalation moved in weeks, painfully. Exception handling, disputes, complex reconciliations, anything where the standard operating procedure covers eighty per cent and the remainder is resolved by asking the person at the next desk. These processes did not fail, but their cycle times lengthened and their error rates rose, because the invisible part of the process, the turning around and asking, had no remote equivalent for the first fortnight. Work anchored to a physical artefact or a floor never moved at all. Anything involving original documents, cheques, stamps, sealed envelopes, secure printing or a device bolted to a desk. These processes stopped, were batched, or are being performed by a skeleton team under permit. No amount of remote enablement changes them, and pretending otherwise wastes the effort that should be spent redesigning them. The lesson buyers should take from this is uncomfortable: the separability of work was always higher than the industry claimed, and the parts that genuinely could not separate were never the parts we were told about.
| Work type | Constraint described | Intervention |
|---|---|---|
| System-queued transactions | Supervisory visibility can disappear. | Assign each item, expose queue age and name an owner. |
| Judgement and exceptions | Informal desk-side escalation does not travel. | Roster an escalation path and document exception decisions. |
| Physically anchored work | Original documents or fixed devices remain on site. | Provide controlled local coverage or redesign the physical dependency. |
Qualitative summary of this article's source text, not a measured outcome or performance estimate.
What broke was management, not delivery
The failures of the past seven weeks have been managerial almost without exception. Supervision by line of sight disappeared and was not replaced. Team leaders who managed by walking the floor found themselves with no instrument at all, and many responded by scheduling meetings, which consumed the productivity the move had gained. Quality sampling, which in most operations happens by a supervisor glancing at screens, stopped for weeks in some teams. Queue balancing, done by eye at 10am, became a spreadsheet nobody owned. And the training pipeline collapsed: new joiners learn this work by sitting next to someone, and that apprenticeship has no remote substitute that anybody had built. Three substitutes work, and they are cheap. Allocate work explicitly rather than letting people take it. A pile that people pull from is invisible; a queue with assignment, ageing and a name against each item is manageable from anywhere. Publish a daily operating rhythm: a short structured start-of-shift call covering volumes, exceptions and blockers, a visible board showing queue status, and a defined end-of-day position. Not a meeting culture, a heartbeat. And rebuild escalation as a named path with a maximum wait. The single most effective intervention I have seen in this period is a rule that no item sits with an unresolved question for more than thirty minutes without going to a named person who is rostered to answer. It replaces the desk turn, and it exposes how much of the work was never documented. One caution on measurement. The instinct to install activity monitoring, keystrokes, screen captures, idle time, is strong and mostly counterproductive. Measure the queue, the cycle time, the error rate and the rework, all of which you already have, and be careful about lagging indicators: quality problems created in April will surface in the reconciliation in June.
The economics have quietly changed
If delivery is no longer tied to a seat, the seat stops being the unit of cost, and everything priced on it comes under question. Facility cost, campus shuttle transport, the premium paid for buildings with the right certifications: all of it is currently being avoided, and buyers know it. Expect that to become a pricing conversation this year rather than next. Providers will argue, correctly, that they have absorbed device costs, connectivity subsidies, additional security tooling and a great deal of management overhead. Both things are true, and the negotiation will be won by whichever side has the better numbers. The more durable change is access to labour. A delivery operation that no longer requires people to live within commuting distance of a campus can hire from smaller cities and from people, particularly women with caring responsibilities, who were previously excluded by the commute and the shift pattern. That is a structural improvement in the talent pool, and it is worth more over five years than the rent saving.
Practical Guidance for Remote BPO Optimization
- Classify every process into the three categories and stop treating them alike. System-queued, judgement-with-escalation, physically anchored. Each needs a different intervention and only the third needs a building.
- Replace pull with explicit allocation. Assigned items, visible ageing, named owner. This single change recovers most of the supervisory visibility that was lost.
- Roster escalation and cap the wait. A named person available for questions at all times, with a maximum response time, published to the team.
- Restart quality sampling immediately, and sample earlier in the process. Errors introduced now appear in downstream reconciliations in two months; sampling at the point of entry shortens that loop.
- Rebuild onboarding deliberately for remote joiners. Recorded walkthroughs, a named buddy, a first-fortnight checklist and supervised work on live items. Apprenticeship does not happen by accident on a video call.
- Measure outcomes, not activity. Cycle time, first-time-right rate, rework, queue age, escalation volume. Activity monitoring buys you conflict and very little information.
- Settle the cost-sharing question in writing with your provider. Devices, connectivity, security tooling, facility savings. An unspoken position now becomes a dispute at renewal.
- Document the processes that were only ever in people's heads. The past seven weeks have exposed exactly which ones those are. That list will not be this visible again.
The Regional Angle
Four consequences matter for Gulf organisations, whether you buy delivery offshore or run your own operations here. The first is infrastructure in the delivery geographies you depend on. Home working in parts of South and Southeast Asia means power interruptions, shared and capped broadband, and connectivity that varies by neighbourhood, which is why providers there have been shipping mobile data devices, small generators and uninterruptible power supplies to agents' homes. Ask your provider what proportion of its home-based agents have a secondary connection, because for a Gulf buyer the visible symptom is not a headline about infrastructure; it is a Tuesday morning when a third of the team is unreachable. The second is a problem sitting quietly in the payroll of almost every regional group: staff who are now working from another country. People travelled in February and could not return, or repatriated when flights allowed, and are continuing to do their jobs remotely on a Gulf employment contract. That arrangement raises immigration status, wage protection compliance, social contribution and personal tax questions in the country they are sitting in, and at scale it raises a corporate permanent establishment question as well. Nobody decided this; it accumulated. Build the list, by person and by country, and get advice on the ones who have been outside for more than a few weeks. The third is that the economics of keeping back-office work in the region have improved, and almost nobody has noticed. The historical argument against a Gulf shared service centre was cost per seat: office rent, visas, accommodation allowances, transport. Remove the seat and much of that gap narrows, while the arguments for proximity remain: Arabic language work, statutory filings, bank and authority interaction, and the same time zone as the business. For groups that were about to offshore statutory accounting for cost reasons, the calculation is worth redoing this quarter. The fourth is hiring reach. Regional employers with national workforce targets have always been constrained by who could physically get to an office in a major city on a fixed shift. Distributed operations make it possible to employ nationals in secondary cities and governorates, and to employ people whose circumstances rule out a daily commute. Remote-work employment frameworks have existed in parts of the Gulf for several years and have been lightly used; this is the moment they become practical, and the organisations that build the supervision model properly will find a labour pool their competitors are not looking at.
The objection worth taking seriously
Calling this a miracle flatters everybody involved and obscures who paid for it. The transition worked in significant part because employees absorbed its costs: electricity, internet, a chair, a corner of a bedroom, childcare while schools are closed, and a working day that now has no boundary. Agents on modest salaries in Manila and Bengaluru are subsidising the facility savings their employers and clients are currently celebrating. Any honest account of the past seven weeks has to record that transfer, and any durable model has to reverse it. The second objection is analytical. Productivity looks excellent partly because volumes in many queues fell with economic activity, so the same people are processing less work, and partly because quality is a lagging indicator that has not yet reported. Operations that look efficient in May may look different when the June reconciliations run and the error rate from April becomes visible. Declaring victory on seven weeks of throughput data, while demand is depressed and quality is unmeasured, is exactly the mistake a delivery director under pressure will make. The response is not scepticism about remote delivery, which plainly works better than the industry believed. It is to fund it properly, pay the connectivity and equipment costs rather than externalising them, and wait for two clean quarters of quality data before rewriting the operating model. The evidence will probably support the change. It has not arrived yet.
Common Questions
Will delivery centres reopen?
Some will, at lower density, for the physically anchored processes, for training, and for clients who require it contractually. The full-occupancy campus model is unlikely to return in its previous form.
Can we renegotiate price on the basis of facility savings?
You can raise it, and you should expect a counter-claim for devices, connectivity and security. Do it with numbers from both sides rather than as a percentage demand, and consider taking value in capacity or transition credits instead.
What about regulated or card-handling processes?
They remain the hardest. Compensating controls, virtual desktops, restricted environments and explicit documented approval are the realistic path, and a small clean floor for the residue is a legitimate answer.
What should we expect over the next twelve months?
Expect hybrid delivery to settle as the default: a smaller floor for controlled and physical work, most transaction processing distributed. Expect clients to write remote capability into contracts as a requirement rather than an exception. Expect the first serious quality and control incidents from this period to surface in the second half of the year and to shape the rules that follow. And expect location strategy to be reopened everywhere, because an industry built on where people sit has just proved that it knows less about that question than it thought.
Remote BPO Optimization — we classify what actually moved, rebuild the supervision and escalation that disappeared with the floor, and settle the cost and quality questions before they become a renewal dispute.