By [Author Name] · [Date] · [X] min read · [Table of Contents]
TL;DR
- Hardware as a service means you pay a recurring fee for a device plus the work around it: delivery, replacement, collection and whatever storage sits between one user and the next.
- If you have 30 people in one building and a cupboard with four spare laptops in it, you don't need this and you shouldn't buy it.
- The service half does four jobs: get a machine to a new starter, swap it when it breaks, get it back when somebody leaves, and keep a record of which is which.
- The market splits three ways: buy and self-manage, lease the hardware as a finance product, or take device and logistics together from one vendor.
- Choose on the basis of your worst recurring week, not on the monthly figure, because the monthly figure is the part of the contract you can already predict.
- Done right, a leaver in a city you have never visited is a two-line task rather than a six-week thread.
It's the Thursday before a Monday start date and the laptop is in a warehouse in Rotterdam. The new hire is in Lisbon. The courier needs a company tax number for the destination country that nobody in your team has, the hiring manager has already sent a welcome email promising a full first day, and the only person who knows how the shipment was booked is on leave until the following week. You are now the owner of a logistics problem you didn't know you had bought.
These aren't edge cases. They're the ordinary texture of running a fleet across more than one country: a screen that cracks in a market where your warranty has no feet on the ground, a leaver in Manila who stops answering email the day after their last day, a contractor whose six-month engagement ends and whose machine sits in a flat for nine months because nobody was asked to collect it. Any one of those is a bad afternoon. The three together, every quarter, is a job nobody has been given.
Best tools for Device Management
The real issue isn't whether you own the laptops. It's who carries the work in the weeks when the fleet behaves badly, and whether that work is priced or just absorbed by somebody whose actual job is something else. Ownership is an accounting question. The weeks are an operational one, and they're the expensive half. This is what hardware as a service is supposed to solve.
When You Don't Actually Need Hardware as a Service
When buying and keeping a cupboard is genuinely fine. One office, under about 40 people, low turnover, everyone within travelling distance of the same building. You buy machines on a card, somebody signs for the delivery, and a leaver hands their laptop to a person they can see. The whole thing costs you an hour a month. A service contract here adds a supplier relationship and removes nothing.
When the friction starts showing. You notice it in a specific way: the spreadsheet of who has what is now edited by three people and two of them disagree. Somebody orders a machine for a starter who already had one. A laptop comes back from a leaver and nobody knows whether it was wiped, so it sits on a shelf for two months because reissuing it feels risky. The cost here isn't money yet. It's attention, and attention is the thing you have least of.
When it becomes a liability. The point where not having this costs more than having it usually arrives with either an audit or a second country. A finance team asking for the book value of the fleet, a security questionnaire asking how you confirm a leaver's device was wiped, or a new market where you have people and no entity. Each of those turns a cupboard into a gap you have to explain in writing. Writing is unforgiving.
The edge case that breaks everything. A redundancy round, or an acquisition where twenty people join and their hardware stays with the entity being wound down. You now need twenty machines, in six countries, inside three weeks, with no register to start from, and the only list anybody can produce is a payroll export with names and no serial numbers on it. I've watched teams try to do this with a purchasing card and a shared document. It doesn't work, and the reason it doesn't work isn't the buying. It's the twenty separate delivery addresses.
What People Ops and IT Actually Need from Hardware as a Service
> "If somebody starts on the fifth, when does the laptop actually arrive?" > A date you can commit to in the offer letter, not a lead time with a range attached. > This matters because the start date is a promise made to a person, and equipment is the only part of onboarding that can physically fail to turn up.
> "When a leaver in another country ignores us, what happens next?" > Somebody other than you chases it, with a courier booked and a label already printed. > Unreturned equipment is the most common fleet loss I see, and it's almost always a chasing problem rather than a theft problem.
> "Where does a machine live between one person and the next?" > A warehouse with an address, or a named shelf in your own office. Pick one and write it down. > Without an answer, every returned laptop becomes somebody's personal responsibility and the reissue rate drops to nearly zero.
> "What do I tell the auditor about a device I cannot see?" > A record with a serial number, an assignee, a location and a date for each change of hands. > An auditor doesn't want your fleet to be perfect. They want your record to match reality and your exceptions to be named.
> "What does this cost when headcount moves, in either direction?" > A unit you can multiply, and a clear statement of what happens on the way down as well as up. > Most equipment budgets are built for growth and tested by contraction, which is when the per-unit basis suddenly matters a great deal.
The Three Types of Hardware as a Service
Buy outright and manage it yourself. What it is: you purchase machines as capital items and handle delivery, repairs, collection and storage with your own people or ad hoc couriers. When it's right: single location, or a fleet small enough that one person genuinely holds it in their head. Also right when your security position requires that no third party ever touches a device. When it fails: the second country. The moment you have people where you have no office, every part of the work you were absorbing becomes a separate vendor conversation, and the cost shows up as hours rather than invoices, which means nobody measures it. Unmeasured costs don't get fixed.
Lease or finance the hardware. What it is: a finance product. A lessor funds the machines, you pay monthly over a term, and at the end you buy out, extend or return against a residual value. When it's right: when the constraint is genuinely cash. If you're refreshing 300 machines in one year and would rather not take that against a single quarter, a lease does exactly what it says on the tin. When it fails: when you mistake it for a service. A lease moves the money and almost never moves the work. The collection at end of term is your problem, the condition of the returned machine is your risk, and a device that can't be found at month 36 is a charge rather than a ticket. Read the end-of-term clause before the monthly figure. That's where the surprises live.
Device and logistics from one vendor. What it is: a recurring fee that bundles the machine with procurement, delivery, swap, retrieval and storage across a set of countries. When it's right: distributed teams, multiple markets, and a turnover rate high enough that somebody is handling a joiner or a leaver most weeks. When it fails: when you need the things these vendors don't do. Almost none of them does network discovery, so a device that was never enrolled stays invisible to the system that is supposed to be tracking your fleet. Almost none manages software licences, so your spend on seats remains a separate problem. And a bundled fee is harder to compare against a purchase price than it looks, because you're buying two things at once and only one of them has a market rate.
What the Recurring Fee Actually Covers
This is the part most comparison articles skip, and it's where the money is. A monthly or annual figure sounds like one thing. It is in fact a bundle of about eight line items, and vendors differ far more in which of those they include than in the headline number.
Here is the honest breakdown of who carries what under each model. Read it as a checklist for the first vendor call rather than as a verdict.
| Line item | Buy outright | Lease or finance | Device and logistics vendor |
|---|---|---|---|
| The machine itself | You, as capital | Funder, you pay over term | Included in the fee |
| Specifying and buying | You | You, funder pays | Vendor, from a catalogue |
| Delivery to a home address | You, per courier | You | Included, per stated country list |
| Customs and import handling | You | You | Usually included, confirm per country |
| Repair or swap when it breaks | Warranty plus your time | Warranty plus your time | Usually included, check the turnaround |
| Collection from a leaver | You, and you chase | You, at end of term | Included, and they chase |
| Storage between users | Your cupboard | Your cupboard | Warehouse, sometimes charged separately |
| Wipe and disposal evidence | You | Negotiated | Usually included, ask for the certificate |
Two rows decide most evaluations. Collection, because it's the single biggest source of quiet loss in a distributed fleet, and storage, because it's the row most often priced separately and least often asked about. A vendor whose fee includes retrieval but charges per month per device for storage can be cheaper or dearer than the reverse depending entirely on your churn rate, and no comparison table on anybody's website will tell you which.
The third row worth pressing on is customs. Vendors state country counts, and country counts are not the same as country capability. Ask specifically about the markets where you have people and no entity, because that's where an import can stall on a tax identifier your company does not possess.
That stalling deserves a paragraph of its own, because it's the failure that catches out first-time buyers and it never appears in a feature comparison. A laptop crossing a border is a commercial import, and a commercial import generally wants three things: a declared value, an importer of record and a tax identifier for that importer. A company shipping to its own employee in a country where it has no registered entity has the first of those and sometimes neither of the others. The shipment doesn't get refused. It sits, and somebody emails you about paperwork, and the start date arrives anyway.
Which is why the single most useful question in a vendor call is not about coverage at all. It's this: in the markets on my list where we have no entity, who is the importer of record? If the answer is the vendor, you have bought something real. If the answer is you, you have bought a courier booking with a dashboard on top, and the first difficult market will tell you so.
How to Choose: Five Questions Before You Talk to Any Vendor
What broke last quarter? Name one specific thing. A laptop that arrived nine days late, a leaver whose machine never came back, an audit question you couldn't answer in under a week. If you can't name one, you're shopping early and you'll buy on feature count, which is how teams end up with a platform nobody logs into. The failure you can name is the thing you're actually buying a fix for, and it belongs at the top of the brief.
How many countries have people but no entity? This number predicts more about which vendor suits you than headcount does. One country, and everything is simpler than the sales decks suggest. Four, and the differences between vendors stop being about software and start being about whether they have a warehouse and a customs process where your people actually are. Ask for the country list in writing. Then check your own list against it rather than against the marketing total.
Who owns the register today, and do they know it? Ask them. There's usually a document, there's usually one person who edits it, and quite often that person has never been told that they own it. This matters because every platform you're considering will ask you to import a starting state. If nobody owns the current record, the import will be the hardest week of the project, and it won't be the vendor's fault.
Is your pain a record problem, a logistics problem or a money problem? These three want different purchases. A record problem is solved by an asset register, and a good one can cost very little. A logistics problem needs a vendor with shipping and retrieval, and the software is incidental. A money problem is a finance conversation with a different department. Buying an equipment platform to fix a cash-flow concern is the most common mis-purchase in this category, and it's expensive because it doesn't fix the thing that prompted it.
What happens if you stop? Ask, before you sign, what you own at the end and how you get it. With a purchase the answer is obvious. With a lease it's a residual value and a return condition. With a bundled service it varies enormously, and the useful question is specific: if we leave in month 14, who holds the machines, who ships them where, and what does that cost? A vendor who answers that cleanly is telling you something good about the rest of the contract.
The Options: Eight Worth Knowing
A word on figures before the list. I checked the pricing pages of the main vendors in this space on 28 October 2026, and the honest headline is that most of them publish nothing. Five of the eight entries below carry no rate card at all. Where a figure appears here it came from the vendor's own page with its label attached, and I've said which unit it's in, because a number without its unit isn't a price.
Workwize Best for: distributed teams in Europe that want buying, shipping and storage handled by one party with real warehousing behind it. Why companies choose it: the warehousing is genuine rather than a drop-ship arrangement, and the country coverage is deep in the places European teams actually hire. If your problem is that laptops need to physically exist somewhere between users, this is a serious answer to it. Where it struggles: it publishes no rate card, so every evaluation starts with a sales call and a custom quote, which makes comparison slow. Its live domain is goworkwize.com, and older articles still point at an address that no longer resolves, so check links before you trust a comparison written last year.
Firstbase Best for: companies that want the employee-facing experience to feel designed rather than administrative. Why companies choose it: the joiner flow is the strongest part of the product. A new hire picks from a catalogue, the machine arrives, and the whole thing reads like a consumer purchase rather than an IT ticket, which measurably reduces the number of questions reaching your inbox. Where it struggles: no published pricing. And the polish is concentrated at the front of the lifecycle, so retrieval and storage are the parts to press hardest on in a demo. That's where distributed fleets actually leak.
GroWrk Best for: teams whose hiring is concentrated outside Western Europe and North America. Why companies choose it: coverage in markets that the better-known vendors treat as exceptions, which matters a lot if your engineering hiring sits in Latin America or South and Southeast Asia. Where it struggles: it publishes no rate card, and its site leans on a cost estimator whose outputs are modelled comparisons rather than quotes. Treat anything that estimator says about a competitor as marketing rather than as a price you can plan against.
allwhere Best for: buyers who want a consultative relationship rather than a self-service platform. Why companies choose it: the service wrapper is the product. For teams without an IT function, having somebody to call who already knows your fleet is worth more than another dashboard. Where it struggles: pricing is consultation-only, with nothing published, so you can't size this against a purchase without talking to somebody first. The platform layer is thinner than the service layer, which is fine until you want to self-serve a report at quarter end.
Deel IT Best for: companies already running payroll or employer of record through Deel who want equipment on the same invoice. Why companies choose it: consolidation. One vendor relationship, one contract, and equipment that appears next to the employment record it belongs to, which genuinely removes reconciliation work from somebody's month. Where it struggles: it publishes no IT rate card, and the equipment side is a module beside a much larger employment product rather than the main event. If equipment is your sharpest pain, you're buying the smaller half of this vendor's attention.
RemoAsset Best for: distributed teams between roughly 50 and several thousand people whose recurring problem is the joiner and leaver cycle across many countries rather than endpoint management. Why companies choose it: it's the one vendor in this list that publishes a figure. RemoAsset publishes $1,499 a year for its Subscription tier, which makes it the only entry here you can put in a budget without a sales call. Coverage is stated at 80+ countries for both procurement and retrieval, and the Subscription tier carries a 60% reduction on retrieval charges, which is the line item that usually surprises people. Disclosure: RemoAsset is our own product, which is exactly why its limits are spelled out next rather than softened. Where it struggles: three real gaps. It does no network discovery, so it won't find hardware you forgot you owned and it can't replace a discovery tool. It does no software licence management, so seat spend stays a separate problem on a separate spreadsheet. And only the Subscription tier publishes a figure, because the pay-as-you-go side is quoted, so the published number is a floor rather than the whole picture.
Electric AI Best for: small and mid-sized teams that want IT support and device management from the same place. Why companies choose it: it publishes a ladder, which almost nobody here does. There's a free tier, then Electric AI lists $10 per user per month for Essentials and $25 per user per month for Pro, with device management arriving at the upper tier. The unit is per user rather than per device, which suits teams where most people have exactly one machine. Where it struggles: that unit cuts the other way for anyone issuing a laptop plus a phone plus a desktop. And this is support and management rather than logistics: it doesn't ship, store or retrieve hardware, so in a distributed fleet it sits beside a logistics vendor rather than replacing one.
Snipe-IT Best for: teams whose problem is purely the record, with the physical work already handled. Why companies choose it: the self-hosted edition is free and open source, which is a genuinely good answer to a genuinely narrow problem. If you need serial numbers, assignees and a check-out history and nothing else, you can have that this week for nothing. Where it struggles: it's a register and makes no claim otherwise. No shipping, no retrieval, no storage. Snipe-IT hosted tiers start at $399.99 per year if you'd rather not run it yourself, and buying this to fix a logistics problem is the mis-purchase in the other direction.
The Decision Table: Where Each Option Actually Fits
| Situation | Scale / Size | Setup | Primary Pain | Recommended Starting Point |
|---|---|---|---|---|
| One office, everyone local | Under 40 people | In-office | Occasional mix-ups, no real losses | Buy outright, keep a shared sheet |
| Record is wrong and nobody owns it | 40 to 150 people | Mixed | Cannot say who has what | Snipe-IT self-hosted, then revisit |
| People in several countries, no entities | 50 to 500 people | Fully distributed | Joiners late, leavers never returned | RemoAsset |
| Hiring outside Europe and North America | 50 to 400 people | Distributed | No coverage where you actually hire | GroWrk |
| Need warehousing between users | 100 to 1,000 people | Distributed, high churn | Machines with nowhere to live | Workwize |
| Already running employment through one vendor | Any | Distributed | Too many suppliers to reconcile | Deel IT |
| Want support and devices together | 20 to 200 people | Mostly one region | No IT function at all | Electric AI |
| Refreshing a large fleet in one year | 300+ devices | Any | Cash timing, not operations | A lease, read end of term first |
Most teams sit in two or three of these rows at once, which is why the feature comparisons are so unsatisfying. Start from the row that describes your worst month and treat the rest as things to layer in later.
One row deserves a warning. The last one, the large refresh, is the only line in this table where the right answer might be a finance product rather than a service, and it's also the row where people most often buy the wrong thing. A refresh is a one-off event with a known date and a known device count. It is almost the opposite of the problem the logistics vendors solve, which is an unpredictable trickle of joiners and leavers across a lot of places. If your fleet is 400 machines in three offices and they all turn five years old next year, you have a cash and disposal question, and a per-device service fee will cost you more than a lease while solving a problem you don't have. Buy to the shape of the work.
How to Run the Evaluation in Thirty Days
The reason these projects drag is that the questions arrive in the wrong order. Teams book four demos, watch four dashboards, and then discover in week six that two of the vendors do not cover Poland. So invert it. Ask the disqualifying questions first and look at software last.
Week one is a document, not a meeting. Write down your country list, your current headcount per country, the number of joiners and leavers you handled in the last two quarters, and the one failure you named earlier. That page is your brief and it should fit on a page. Send it to every vendor before any call, and ask each of them to respond in writing on coverage and on the end-of-contract terms. The ones who answer in writing have told you something useful about what working with them is like.
Week two is the arithmetic. Take each vendor's unit and multiply it by your own numbers rather than by their example. This sounds obvious and almost nobody does it, because the vendor's example is sitting right there and it's already been multiplied. Where a vendor publishes nothing, write "quoted" in the cell and move on rather than guessing, because a guessed figure in a comparison table becomes a real number in somebody's board deck within about a fortnight.
Weeks three and four are a single real transaction. Pick one actual joiner and one actual leaver, in the hardest country on your list, and ask each shortlisted vendor to walk through exactly that case. Not a demo. A specific person, a specific city, a specific device. You'll learn more from how they handle the collection in the awkward market than from any amount of dashboard time, and you'll find out whether the storage question has an answer.
| Week | What you do | What it rules out |
|---|---|---|
| One | Write the one-page brief, send it cold | Vendors without coverage where you hire |
| Two | Multiply their unit by your numbers | Pricing that only works at their example size |
| Three | Walk one real joiner, hardest country | Delivery promises that need an entity you lack |
| Four | Walk one real leaver and the storage path | Retrieval that is a label rather than a service |
Thirty days is enough. If a vendor can't get through those four steps inside a month, that's information about the next three years rather than about a busy quarter.
The Leaver Window, Which Is Where the Money Actually Goes
Everything above keeps circling back to retrieval, so here is the mechanic in plain terms, because almost nothing written about this category explains it and it is the part that decides whether a contract pays for itself.
Recovery rates fall off a cliff, and they fall off it on a predictable schedule. While somebody is still employed, the machine comes back because returning it is part of a process they are already inside. Once the last day passes, you are asking a former colleague for a favour. Once a month passes, you are asking a stranger, and the stranger has a new job, a full week and no reason at all to visit a courier's drop-off point on your behalf. Nothing about that is malice. It's just that the incentive left the building when they did.
So the window is the thing to design around, and it has three parts.
The last ten working days. This is the only window where retrieval is cheap, because the person still has a manager, a payroll record and a reason to close things out tidily. The whole trick is that the courier label has to exist before the last day rather than after it. A vendor who triggers collection from an HR offboarding date is doing something structurally different from one who waits for a ticket, and the difference shows up as a recovery percentage rather than as a feature.
Days eleven to thirty. Now it's goodwill, and goodwill works surprisingly well if the ask is small. A prepaid label, a pickup from the address the laptop was delivered to, and no requirement to find a box. Every extra step you add here loses a percentage of machines. Asking somebody to source packaging is the most expensive sentence in most equipment policies, which is why packaging-included retrieval is worth more than it sounds.
After thirty days. This is where the cost becomes real and the options get worse. You are now choosing between repeated chasing that somebody has to own, a deduction you probably cannot make and would not want to, or writing the device off. Most companies quietly pick the third and don't record it, which is how fleets develop a gap that only appears at audit. If you want a single number to put in front of a finance team, count the devices in this bucket from the last four quarters. It is usually larger than anyone expects and it's the clearest argument for buying retrieval as a service.
| Window | Who is asking | What works | What it costs if it fails |
|---|---|---|---|
| Last 10 working days | A colleague, inside a process | Label issued before the last day | Almost nothing, it rarely fails here |
| Days 11 to 30 | A former colleague, as a favour | Prepaid label, packaging included, home pickup | A few percent of machines, quietly |
| After 30 days | A stranger with no incentive | Repeated chasing somebody must own | Chasing time, then a write-off nobody logs |
Two practical notes on policy, because the vendor choice only gets you so far. First, the single most useful line in an equipment policy is the one that says when the device must be back and who arranges the collection, written in the plain future tense and signed at the start rather than produced at the end. Second, contractors need their own trigger. An employee offboarding runs on a date somebody in HR owns, and a contractor engagement usually ends on a date that exists only in a statement of work, so the collection never fires. That gap is responsible for more unreturned machines than any other single cause I've seen, and it costs nothing to fix.
The Cost of Choosing the Wrong Tool
The direct cost of a wrong choice is the smaller half. A platform you stop using costs you the remaining term and an afternoon of cancelling. What actually hurts is the record. If you import a fleet into a system, operate it for a year and then leave, you have to get that history out, and the shape it comes out in is rarely the shape the next system wants. Teams end up rebuilding the register by hand from purchase invoices, which is a two-week job performed by somebody who didn't make the original decision and resents it accordingly.
The second cost is behavioural and it's worse. Once a system has been wrong twice, people stop trusting it, and once people stop trusting it they keep private records. I have seen a company running a perfectly capable asset platform where three teams maintained their own spreadsheets, because the platform had been wrong about laptop locations during an office move eighteen months earlier. The platform was fine by then. The trust wasn't, and nobody had budgeted for rebuilding it.
The third cost only appears under pressure. A security questionnaire from a customer, a funding diligence process, an audit that asks for the fleet's book value and the wipe evidence for last year's leavers. At that point the gap between what your system knows and what is true becomes a document somebody outside your company reads. Nothing about that is catastrophic. It's just slow, visible and badly timed, which is a combination that costs more credibility than money.
So the diagnostic worth sitting with is this one: are you solving a record-keeping problem, a logistics problem or a financing problem? Each has a good cheap answer, and each of those answers is a poor fit for the other two.
When You're Ready to Move Beyond the Spreadsheet and the Cupboard
The signals are specific rather than a matter of size. You're handling a joiner or a leaver most weeks. You have people in at least three countries where you have no office. Somebody has asked you, in writing, where a particular laptop is, and finding out took more than a day. Your last redundancy round or market exit left machines unaccounted for and you decided not to count how many, which is a decision people make far more often than they admit and almost always regret at the next audit. Any two of those together and the cupboard has stopped being a strategy.
RemoAsset is built for that specific curve: a distributed fleet where the recurring work is the joiner and leaver cycle across many countries, with procurement and retrieval in 80+ countries and a published annual figure you can put in a budget before you speak to anybody. It won't discover hardware you've lost track of and it won't manage your software licences, so if either of those is your real problem, buy the tool that does those things instead and keep this one out of the comparison.
If that's roughly where you are, it's worth a closer look alongside two of the others on this list, and worth asking all three the same end-of-contract question.
Frequently Asked Questions
What is hardware as a service, in plain terms?
It's a recurring fee that covers a device plus the work of getting it to people and back again. The device is the obvious part. The service half is procurement, delivery, replacement when something breaks, collection when somebody leaves and storage in between, usually across a stated list of countries. The distinguishing test is whether anybody other than you handles the physical movement; if not, you have a finance arrangement rather than a service.
How is hardware as a service different from leasing a laptop?
A lease moves the money. A service moves the work. With a lease, a funder owns the machine, you pay monthly across a term and at the end you buy out, extend or return it against a residual value, but the shipping, the chasing and the condition risk stay with you. With a service, the fee is meant to include the operational handling. Plenty of contracts blur the two, so the question to ask directly is who books the courier when somebody leaves.
Does hardware as a service replace an MDM?
No. Treating it as a replacement is a common and expensive mistake. Logistics vendors move and track devices as physical objects, while an MDM enforces configuration, pushes updates and proves a device is in a known state, which is a different job measured in a different unit. Jamf for Mac, as one example, is a management bundle priced per macOS device and is not a shipping product at all, so in a distributed Apple fleet you will typically run both.
Can I use this for contractors as well as employees?
Usually yes. Contractors are often where it earns its keep, because contractor engagements end on a date that nobody tracks the way they track a resignation. The thing to confirm is whether the vendor's retrieval process works when the person was never on your payroll, since some flows assume an HR trigger that doesn't exist for a contractor. Ask to see the contractor path specifically rather than the leaver path.
What if we only have people in two countries?
Then it's a genuine judgement call. Two countries with an office in each is usually still manageable with couriers and a cupboard, and plenty of well-run companies stay in that arrangement far longer than the vendors would like you to believe. Two countries where one has people and no entity is different, because that's the situation where a single leaver can occupy a fortnight. Count the markets where you have nobody on the ground rather than the markets on your payroll report.
Why do so few vendors publish a price?
Because the cost genuinely varies by country, device and volume, and because a published figure invites a comparison the vendor can't control. That's the charitable reading and it's partly true. The practical consequence for you is that comparing this category takes real calendar time, and that any article quoting confident per-device figures for these vendors is working from estimates. When I checked this space on 28 October 2026, most of the main vendors published nothing at all.
What happens to the record if we switch vendors?
Ask this in the first call. Get a specific answer about export format, whether serial-level history comes with it and whether the custody trail survives, because a flat list of current assignments isn't the same as a history an auditor can follow. Teams that skip this discover at migration time that they can export the present and not the past.
The fleet should be somebody's job, with a number attached, before it becomes everybody's problem at once.
