Buying laptops as they fail is the most expensive way to run a fleet. Windows 11 has just made the alternative unavoidable.
Reactive replacement produces a fleet of a dozen models nobody can support consistently, a budget that arrives as surprises, and devices that are always slightly too old. A planned refresh produces a small number of standard models bought on a schedule, provisioned without an engineer through Autopilot, and disposed of with a certificate. The Windows 11 hardware bar has forced the decision on every organisation still carrying pre-2018 devices, and the readiness assessment tells you exactly how many that is.
- Few modelsStandardise, then support
- PhasedBy role and risk, not by failure
- Registered at purchaseSo Autopilot works from the box
- Certified disposalA DPDP matter, not housekeeping
Eight decisions that turn a replace list into a programme.
Standardising on a small number of models
Every model in the fleet is a driver set to maintain, a firmware procedure to test, an Autopilot profile to verify and a spare parts line to hold. Two or three standard models, one per role type, is manageable. A dozen accumulated through reactive purchasing is not. Standardisation is the decision that makes every subsequent piece of support cheaper, and it only happens if somebody decides it before the order.
Matching models to roles rather than to people
A finance manager, a field engineer and a receptionist need different devices, and giving everyone the same laptop either overspends on the receptionist or underserves the engineer. Role-based specification, a knowledge worker standard, a mobile standard, a fixed-desk standard, is how a fleet gets the right device to the right person without a per-user procurement conversation.
Phasing by role, risk and business calendar
The replace list is sequenced, not dumped. Devices handling sensitive data or facing the internet go first. Departments with a month-end or a busy season are scheduled around it. Sites with disposal logistics are grouped. The phasing is what makes a refresh deliverable by an IT team that also has a day job, and it is what gives the ESU exit dates something real behind them.
Supplier registration and delivery
Devices registered for Autopilot at the point of purchase arrive already known to your tenant and can ship directly to a user anywhere in India. That is a clause in the order, agreed before it is placed. Without it, every device passes through IT for registration first, which turns zero-touch into one-touch and adds a shipping leg to every delivery.
Warranty, support and what happens when one breaks
The warranty term and the support model decide the total cost of the fleet far more than the purchase price does. On-site next business day for a distributed workforce, or a swap pool held centrally, or an AMC arrangement that covers the fleet after the warranty lapses. Deciding this at purchase is cheap. Discovering the answer at the first failure is not.
Peripherals, docks and the things nobody budgets
A refresh of laptops is also a refresh of docks, monitors, headsets and the cables between them, and the compatibility between a new device and an old dock is not guaranteed. This is the line that surprises the budget, and it is worth listing per role before the order rather than discovering it desk by desk during rollout.
Disposal of the outgoing fleet
Every retired device needs a data wipe to an evidenced standard, a closed asset record, and disposal through an authorised channel that provides a certificate. Under the DPDP Rules a laptop leaving your control with residual personal data is a potential breach with a reporting obligation, and Indian e-waste rules require authorised disposal. Building this into each wave prevents the cupboard full of old laptops that almost every organisation has.
The cycle after this one
A refresh is not an event, it is the start of a cycle. Deciding now that a third of the fleet is replaced each year, or that devices are replaced at a defined age, turns the next refresh from a crisis into a budget line. The Windows 11 refresh is the moment to establish the cycle, because the whole fleet is being looked at anyway.
Eight refresh failures, and the decision that prevents each.
A dozen models accumulated over three years
Bought reactively as devices failed, each on whatever the supplier had in stock. Every one is a separate driver set, firmware procedure and spare parts problem.
- Standardise on two or three models before the first order
- Buy in tranches of the standard, not one-offs
- Refuse the substitution the supplier offers when stock is short
Devices arrive unregistered
The supplier did not register the hardware hashes, or registered them to the wrong tenant, so every device has to come through IT before it can go to a user.
- Put registration in the order as a condition of acceptance
- Verify in the tenant before accepting delivery
- Keep a fallback for capturing hashes, and hope not to need it
The whole fleet lands in the same month
Procurement was efficient and the rollout was not. Fifty devices in boxes, an IT team of two, and users asking when they get their laptop.
- Phase deliveries to match provisioning and support capacity
- Sequence by role and risk, not by whatever arrived first
- Autopilot removes the provisioning bottleneck, not the support one
The docks do not work with the new laptops
The refresh budget covered laptops. The existing docks, monitors and cables were assumed compatible and were not.
- List peripherals per role before the order
- Test the new model against the existing dock before buying fifty
- Budget the peripherals line explicitly
The first failure has no support path
A device breaks in a remote city, the warranty is return-to-base, and the user is without a laptop for a fortnight.
- Decide the support model per role at purchase
- On-site or swap pool for anyone who cannot be without a device
- Know when the warranty ends and what covers the fleet after
Old devices accumulate in a cupboard
Retired laptops with data still on them, no asset record closed, no disposal arranged. A DPDP exposure sitting in a storeroom.
- Wipe and dispose wave by wave, not at the end
- Use an authorised channel that provides a certificate
- Close the asset record when the certificate arrives
Everyone gets the same device
A single specification for the whole fleet overspends on the roles that need little and underserves the roles that need more, and both groups notice.
- Two or three role-based standards
- Specify from what the role does, not from what the loudest user wants
- Review the standards at each cycle, not each order
The next refresh is another crisis
The fleet was replaced in one go, so it ages in one go, and in four years the same emergency happens again with the same budget shock.
- Establish a rolling cycle from this refresh onward
- A third of the fleet per year turns a crisis into a line item
- Track device age so the next order is planned, not forced
We run fleets after the refresh, so we plan the one we would want to inherit.
We start from the readiness assessment, not from a catalogue
The replace list from a device-side scan is what a refresh plan should consume. Starting from what the fleet actually needs, after the firmware recount has shrunk it, produces a smaller order and an honest budget. Starting from a supplier catalogue produces neither.
Microsoft Partner, so Autopilot registration is part of the order
We know what to write into the procurement to get devices registered at purchase with the right group tags, and how to verify it in the tenant before accepting delivery. It is a small clause that turns a refresh from one-touch into zero-touch.
We standardise honestly, including for executives
The premium executive tier is the most common standardisation failure, and it creates a support exception for the people least tolerant of support problems. We will make that case, and we will lose it sometimes, and the fleet is better when we win it.
We manage estates afterwards, so the plan has to work
A refresh partner who leaves after delivery has no reason to care about model count, support models or disposal. We support fleets for clients on an ongoing basis, so a fleet of a dozen models with a cupboard of old laptops is our problem next year as well as yours.
Why organisations plan one now.
The Windows 11 hardware bar
The dominant trigger in 2026. Devices below the processor bar have a defined end date, and the readiness assessment tells you how many there are. Most refresh programmes now start here.
An ESU exit date approaching
Devices bridged on Extended Security Updates with exit dates that need an order behind them. The refresh plan is what makes those dates real rather than aspirational.
A distributed or remote workforce
Staff across cities who cannot be handed a laptop, where direct-to-user shipping through Autopilot is the only sensible provisioning model and the refresh is the moment to adopt it.
Regulated organisations
Unsupported devices are an audit finding, and disposal without a certificate is another. The refresh plan is as much a compliance document as a procurement one.
Education with bulk purchasing cycles
Large tranches bought years apart, so the fleet ages in blocks. The refresh is the chance to move to a rolling cycle so the next block does not arrive as an emergency.
Growth past reactive purchasing
Organisations that bought devices one at a time as they hired and now have a fleet of a dozen models nobody can support consistently. The refresh is the reset.
What each role actually needs, and what it does not.
| Role type | What matters | What does not | |
|---|---|---|---|
| Knowledge worker, office and hybrid | Reliable, light enough to carry, a good screen and keyboard, a dock at the desk, a long warranty. Standard memory and storage. | High-end graphics, the largest screen, the thinnest chassis. These cost money and change nothing about the work. | |
| Mobile and field | Battery life, durability, cellular connectivity where coverage allows, and a support model that does not require returning the device to an office. | Docking and desk peripherals. A field device that spends its life in a bag has different priorities. | |
| Fixed desk and shared | Cost efficiency, ease of provisioning with a self-deploying profile, and a locked-down configuration. Often a small form factor desktop rather than a laptop. | Portability and battery. A device that never moves does not need to be a laptop. | |
| Engineering, design and data | Memory, processor and graphics matched to the actual workload, which is worth measuring rather than assuming. A larger screen or an external one. | The lightest chassis. Performance and portability trade against each other, and this role picks performance. | |
| Executive | The knowledge worker standard, honestly. Reliability and a long warranty matter more here than anywhere. | A separate premium tier. It creates a support exception for the people least tolerant of support problems. | |
| Kiosk, shop floor, reception | Whatever the fixed-purpose application needs, provisioned self-deploying, with the software vendor consulted. Sometimes a thin client or a Cloud PC. | General-purpose specification. These devices do one thing and should be sized and locked down for it. |
Reactive replacement against a planned refresh cycle.
| Feature | Dimension | Reactive | Planned |
|---|---|---|---|
When devices are bought | When one fails, or when a hire starts | On a schedule, in tranches | |
Models in the fleet | Whatever was in stock each time | Two or three standards, one per role | |
Provisioning | An engineer per device | Autopilot, registered at purchase, shipped direct | |
Support | Discovered at the first failure | Decided per role at purchase | |
Peripherals | Found not to fit, desk by desk | Tested and budgeted before the order | |
Outgoing devices | A growing cupboard | Wiped and certified per wave | |
The budget | Surprises through the year | A line item, known in advance | |
The next refresh | Another crisis in four years | A rolling cycle, already underway |
The old laptops in the cupboard are a DPDP exposure, not a storage problem.
Almost every organisation we work with has a store of retired devices nobody has dealt with, and almost all of them still hold data. Under the DPDP Rules, a device leaving your control with residual personal data is a potential breach with a reporting obligation attached, and under Indian e-waste rules disposal has to go through an authorised channel. A refresh is the moment this either gets built into the process, wave by wave, with a wipe standard and a certificate per device, or gets deferred into a cupboard that grows with every wave. The second outcome is the default. It should not be.
- Wipe to a standard you can evidence, and record which standard
- Dispose through an authorised e-waste channel and keep the certificate
- Close the asset record when the certificate arrives, not before
- Do it per wave, because a cupboard of a hundred devices is a project nobody starts
Five stages, and the standards come before the catalogue.
- 1
Take the replace list and define the role standards
The readiness assessment output, after the firmware recount, sorted by role. Two or three role-based specifications written from what each role actually does, agreed with department heads, and a decision on executives that we will argue for and sometimes lose.
- 2
Select models and test them
Standard models chosen against the specifications, one device of each tested against existing docks and peripherals before any volume order, and the peripheral line budgeted per role. Warranty term and support model decided per role at the same time.
- 3
Write the order with Autopilot registration in it
Registration at purchase with group tags as a condition of acceptance, direct-to-user shipping for remote staff, and delivery phased to match provisioning and support capacity rather than arriving in one month. Verified in the tenant before delivery is accepted.
- 4
Roll out in waves with disposal built in
Waves sequenced by role, risk and business calendar, aligned to any ESU exit dates. Each wave includes the wipe and certified disposal of the outgoing devices, so nothing accumulates. Asset records updated at delivery and closed at disposal.
- 5
Establish the cycle
Device age tracked in the asset register, a rolling replacement cadence decided, and the next tranche planned as a budget line rather than a future emergency. Then handover to your team, or ongoing fleet management if that is what you want.
Twelve things to settle before procurement places the order.
What to buy
- Two or three standard models chosen, one per role typeNot whatever is in stock
- Role-based specifications written downFrom what the role does, not from requests
- New model tested against existing docks and peripheralsBefore buying fifty
- Peripherals budgeted per roleThe line that surprises finance
How it arrives
- Autopilot registration written into the order as a conditionWith group tags for profile assignment
- Delivery phased to match provisioning and support capacityNot all in one month
- Direct-to-user shipping agreed for remote staffThe whole point of Autopilot
- Warranty term and support model decided per roleOn-site, swap pool, or AMC after warranty
What happens after
- Wave sequence agreed by role, risk and business calendarESU exit dates aligned to it
- Wipe standard and disposal channel arrangedCertificate per device, DPDP and e-waste rules
- Asset register updated at delivery and closed at disposalOne record per device, start to finish
- The rolling cycle for the next refresh decidedSo this is the last crisis
Device refresh, answered plainly.
What sits either side of this.
Windows 11 readiness assessment
The scan that produces the replace list this plan consumes, after the firmware recount has shrunk it.
Learn moreWindows Autopilot deployment
The provisioning that makes a refresh of any size deliverable: registration at purchase, profiles, and shipping direct to users.
Learn moreWindows 11 vs Windows 10 for business
The device-by-device decision behind the replace list: what Windows 11 changes, what ESU really costs, and where each device lands.
Learn moreDecide the standards before the catalogue arrives.
The order is the point of no cheap return, and everything that makes a refresh work, the model count, the registration clause, the support model, the disposal route, is decided before it. Tell us where your replace list stands and we will plan from there. Remote-first from Hyderabad, serving all of India.
Related Services
Explore more solutions that work great with this service
Windows 11 Migration
Sort the fleet, bridge the exceptions, provision without hands
Learn moreWindows 11 Readiness
Device-side scan, every machine sorted, report as the plan
Learn moreWindows Autopilot
Devices that configure themselves from the box
Learn moreComputer AMC Hyderabad
Hardware AMC for device fleets
Learn more