We value your privacy

We use cookies to analyze site traffic and improve your experience. You can accept all cookies or reject non-essential ones. See our Privacy Policy for details.

GR IT SERVICES
  • Contact
Get a quote
  1. Microsoft security
  2. Windows LAPS
Windows LAPS for US organizations

One local administrator password across the estate means one compromised machine is all of them.

Give every machine its own local administrator password, rotate it on a schedule, and keep a copy somewhere your help desk can actually reach it. That is the whole feature, it works on machines joined to Entra ID or to Active Directory, and the benefit Microsoft names first is exactly the failure a shared password guarantees: pass-the-hash and lateral traversal.

Book a Windows LAPS deploymentSee what it covers
Windows LAPS deployment for US organizations
  • FreeThe feature itself, on all supported Windows
  • Per deviceA distinct password, rotated automatically
  • Entra or ADBackup destination, and only one of them
  • DeprecatedLegacy Microsoft LAPS, from Windows 11 23H2
What Windows LAPS does

Eight things to establish before you deploy it.

Very few security improvements return this much for this little, and it costs nothing. Deployments still fail, and they fail for three reasons every time: somebody misread the join state, an existing legacy install was left running alongside it, or nobody checked that the people who need a rotated password can actually get hold of one.

A managed password per device

The password on a local administrator account gets managed and backed up automatically, on anything joined to Entra ID or to your domain. Every endpoint ends up holding a different one, rotated to a schedule, which means whatever an attacker pulls off the first machine they land on will not open the second.

The attack it directly addresses

Top of the published benefits list sits protection against pass-the-hash and lateral traversal. Consider why. One password shared across every workstation is about the shortest route there is between a single phished laptop and your entire environment, and this removes the route rather than waiting to notice somebody using it.

Join state decides where passwords go

A cloud-only machine has exactly one destination available to it, and so does a domain-only machine. Only hybrid joined hardware gets a genuine choice, and even then it is a choice of one, because backing up to both directories is not something the product will do. Workplace joined clients get nothing at all here.

How stored passwords are protected

Two different models depending where the password lands. On the domain side, granular access control lists plus encryption if you want it. On the cloud side, the role-based access control you already use for everything else. What both give you is the same thing: pulling a password becomes a permissioned action that leaves a trace, instead of a value sitting in a spreadsheet somebody shared in 2021.

Password history, which legacy LAPS never had

History is stored, and the reason to care is recovery rather than audit. Restore a machine from an image captured before the most recent rotation and that machine is still expecting the previous password. No history means no way in through the very account you deployed this to guarantee access with, which is a uniquely frustrating way to lose a device.

The legacy product is deprecated

The old add-on reached deprecation at Windows 11 23H2. Its installer will not run on newer builds, and no further code changes are being considered for it. Whatever support remains applies only to older Windows versions and expires when those versions do, so it is a position with a published end date rather than an indefinite one.

Migration from legacy is supported deliberately

Nothing about the modern feature depends on the old one being present, so a fresh deployment simply ignores it. Where a legacy install already exists, there is an emulation mode built specifically for that situation, which lets you move across in stages rather than switching everything one evening and hoping.

Domain controllers get covered too

The Directory Services Restore Mode account on your domain controllers can be managed and backed up the same way. Think about when that password was last set. Somebody typed it during promotion, wrote it down or did not, and it has sat unchanged ever since, which makes it comfortably one of the oldest live secrets anywhere in the environment.

The design decision to get right first

A device backs up to Entra ID or to Active Directory. Never both.

How a machine is joined dictates what options it has. For most estates that means the decision makes itself, and it is only the hybrid population where somebody genuinely has to choose.

  • Cloud-only machines have one destination. Domain-only machines have the other. Hybrid joined hardware can use whichever you prefer, singular, because the documentation is unambiguous that backing up to both directories is not supported.
  • Workplace joined clients are excluded outright. Wherever personal or lightly managed hardware sits in that state, it falls entirely outside this control, and the useful thing to do is write those machines down as uncovered rather than letting an existing tenant policy create the impression they are handled.
  • For the hybrid group, decide it on who does the retrieving. A help desk living in the cloud portal all day is better served by cloud backup and its role-based model. A team already comfortable in the directory, working with access control lists and possibly encryption, is better served staying there.
  • Cost almost never enters into it. The capability is free across every supported Windows platform, storing passwords in the directory carries no extra licensing at all, and the cloud destination asks for nothing beyond the free tier.
Ask us to design the backup model
How we approach it

Four things that make a LAPS deployment stick.

Pushing the policy out is close to trivial. Where these projects come apart is three months later, when the help desk finds they cannot get a password at the moment they need one and quietly stands up a second local account with a password everybody knows, because the tickets still have to be closed.

We design retrieval before we design rotation

Rotation without accessible retrieval is simply an outage you have scheduled in advance, and the workaround your team will reach for undoes the entire exercise. Settling the permission model first, whether that is cloud roles or directory access control lists, is what stops the workaround ever becoming necessary.

We settle the destination question per device group

Two thirds of your estate has no choice to make, and the hybrid third has one choice that excludes the other. In a genuinely mixed environment that amounts to a real design decision, and taking it deliberately is what prevents the half-finished rollout that nobody spots for a year.

We keep password history for the restore case

Any machine brought back from an image older than the last rotation is going to want the previous password. Storing history is something the modern feature does and the old add-on never could, and having it configured is the difference between recovering that device in ten minutes and rebuilding it from scratch.

We migrate legacy rather than running both

The old product is deprecated, will not install on current builds, and receives no further development. An emulation mode was provided precisely so the move can be made in an orderly way, and taking it is considerably better than the alternative, which is two separate products with opinions about the same account.

How a deployment runs

Four phases across roughly four to six weeks.

A short engagement measured against what it buys you. The bulk of the time goes on retrieval rather than on configuration, because retrieval is the part that decides whether your help desk can still function on the Monday after go-live.
  1. 01
    Week 1

    Establish join state, versions, and current practice

    We sort the estate by join state, confirm the update level is high enough for the feature to exist at all, and then establish how local administrator passwords are being handled today. That third question is the one that produces silence in the room, and it is usually the most useful answer of the week.

    • Join state inventory across the estate
    • Windows version and update level confirmed
    • Workplace joined devices identified as out of scope
    • Current local administrator password practice documented
  2. 02
    Week 2

    Design the backup destination and retrieval model

    A destination gets chosen for each group of machines, subject to the one-or-the-other constraint. Then the harder half: which people may pull a password, under which permission model, and what gets recorded when they do. This is the week where the help desk workflow is actually designed instead of being left to work itself out.

    • Backup destination decided per device group
    • Retrieval permissions designed using RBAC or ACLs
    • Password encryption decision recorded for Active Directory
    • Password history retention agreed
  3. 03
    Weeks 3 to 4

    Deploy by policy and validate retrieval

    The policy goes out through the configuration service provider or through Group Policy, and then comes the step almost everybody skips. Somebody retrieves a real password and signs in with it, using the account of a person who will one day be doing this at three in the morning with a director on the phone.

    • Policy deployed to a pilot group
    • Password backup confirmed in the chosen directory
    • Retrieval tested by help desk staff, not by the project team
    • Event log channel monitored for failures
  4. 04
    Weeks 5 to 6

    Broaden, retire legacy, and cover domain controllers

    The policy widens across the estate, any existing legacy deployment is moved across using emulation mode and then decommissioned properly, and the Directory Services Restore Mode password on your domain controllers is brought into the same management.

    • Full estate coverage achieved and measured
    • Legacy Microsoft LAPS migrated and retired
    • DSRM password management enabled on domain controllers
    • Coverage reporting handed to the operational team
Where this matters

Six situations where local administrator passwords are the exposure.

They all look the same in the end. A company that has done serious work on identity for its human beings, and one single credential sitting quietly on every endpoint it owns.

A business built from a standard image

Deploy a thousand machines from one image and all thousand share a local administrator password, typically the one chosen on the afternoon the image was built. This is the single most frequent finding in an endpoint review, and it is also among the quickest things on the report to actually close.

A firm asked about lateral movement controls

Pass-the-hash and lateral traversal protection heads the published benefit list, which matters because insurers, SOC 2 auditors and anyone running a NIST CSF assessment now put the question in writing. Answering that every endpoint holds its own automatically rotated credential, retrievable only by named people, is direct and checkable. Answering anything else invites follow-up questions you would rather not have.

An operator with machines that are rarely online

Terminals on a plant floor and equipment at unstaffed sites can go weeks without checking in, which raises the stakes on both password history and a retrieval process that works first time. In these environments the recovery case is not the edge case, it is the one your team performs most often.

A provider recovering an inaccessible device

Being able to sign in to and recover a machine that is otherwise unreachable is named as a benefit on its own terms. On clinical hardware, where rebuilding means revalidation and a device out of service for days, holding a retrievable local credential functions as an availability control every bit as much as a security one.

An organization still running legacy Microsoft LAPS

Deprecated, blocked from installing on current builds, and receiving no further development. There is no hard deadline forcing your hand this quarter, which is exactly why it keeps slipping down the list, but the path it is on has a known end point. Emulation mode is there to make the transition unremarkable.

A company that has never changed its DSRM password

Somebody set that password while promoting a domain controller, possibly before the current IT manager was hired, and in a great many environments nothing has touched it since. Bringing it under automatic management and backup converts the oldest surviving secret in the building into an ordinary managed credential.

Three positions

How US organizations handle local administrator passwords.

Far more organizations sit in the right hand column than will say so out loud. In an estate that is otherwise run well, it is reliably the most exploitable thing left standing.
Unique password per device
Windows LAPS deployedYes
Legacy Microsoft LAPSYes
Shared passwordNo
Automatic rotation
Windows LAPS deployedYes
Legacy Microsoft LAPSYes
Shared passwordNo
Backup to Entra ID
Windows LAPS deployedYes
Legacy Microsoft LAPSNo
Shared passwordNot applicable
Password encryption in Active Directory
Windows LAPS deployedOptional
Legacy Microsoft LAPSNo
Shared passwordNot applicable
Password history stored
Windows LAPS deployedYes
Legacy Microsoft LAPSNo
Shared passwordNot applicable
Retrieval permissioned and recorded
Windows LAPS deployedYes
Legacy Microsoft LAPSYes
Shared passwordNo
DSRM password managed
Windows LAPS deployedYes
Legacy Microsoft LAPSNo
Shared passwordNo
Product actively developed
Windows LAPS deployedYes
Legacy Microsoft LAPSDeprecated
Shared passwordNot applicable
Lateral movement from one endpoint
Windows LAPS deployedBlocked at this vector
Legacy Microsoft LAPSBlocked at this vector
Shared passwordTrivial
Additional license cost
Windows LAPS deployedNone
Legacy Microsoft LAPSNone
Shared passwordNone
Feature
Windows LAPS deployed
Legacy Microsoft LAPS
Shared password
Unique password per device
YesYesNo
Automatic rotation
YesYesNo
Backup to Entra ID
YesNoNot applicable
Password encryption in Active Directory
OptionalNoNot applicable
Password history stored
YesNoNot applicable
Retrieval permissioned and recorded
YesYesNo
DSRM password managed
YesNoNo
Product actively developed
YesDeprecatedNot applicable
Lateral movement from one endpoint
Blocked at this vectorBlocked at this vectorTrivial
Additional license cost
NoneNoneNone
Platform support

Where Windows LAPS is available, and where it is not.

Coverage reaches further back than most people assume, and the reason is that this arrived in a monthly Windows update rather than requiring a new operating system release.

Platform or state

Windows 11 23H2 and later

Support position
Supported

Platform or state

Windows 11 22H2 and 21H2

Support position
Supported with the April 11 2023 update or later

Platform or state

Windows 10

Support position
Supported with the April 11 2023 update or later

Platform or state

Windows Server 2025 and later

Support position
Supported

Platform or state

Windows Server 2022 and 2019

Support position
Supported with the April 11 2023 update or later

Platform or state

Automatic Account Management CSP settings

Support position
Require Windows 11 24H2 or later

Platform or state

Entra joined devices

Support position
Back up to Entra ID only

Platform or state

Active Directory joined devices

Support position
Back up to Active Directory only

Platform or state

Hybrid joined devices

Support position
Either destination, never both

Platform or state

Entra workplace joined clients

Support position
Not supported
Platform or stateSupport position
Windows 11 23H2 and laterSupported
Windows 11 22H2 and 21H2Supported with the April 11 2023 update or later
Windows 10Supported with the April 11 2023 update or later
Windows Server 2025 and laterSupported
Windows Server 2022 and 2019Supported with the April 11 2023 update or later
Automatic Account Management CSP settingsRequire Windows 11 24H2 or later
Entra joined devicesBack up to Entra ID only
Active Directory joined devicesBack up to Active Directory only
Hybrid joined devicesEither destination, never both
Entra workplace joined clientsNot supported
How an engagement runs

Five steps, and the retrieval test is the one that matters.

The deployment itself amounts to pushing a policy. Whether you are actually safer at the end of it comes down to one thing, which is whether your help desk can still do their work the following morning without inventing a workaround.
  1. 1

    Establish join state and Windows version coverage

    Four populations get counted individually, because three of them determine where passwords will be stored and the fourth cannot participate at all. Alongside that we confirm build and update level on every machine, given that the capability arrived through a servicing update rather than an OS upgrade.

  2. 2

    Decide the backup destination per device group

    One destination or the other, and for the hybrid population the deciding factor is almost always operational rather than architectural: who pulls passwords day to day and which console they already live in. The permission models differ between the two, so matching the destination to your team saves friction later.

  3. 3

    Design retrieval permissions and history

    Three questions get settled: who is allowed to pull a password, under what permission model, and what evidence exists afterwards that they did. History gets configured as a deliberate choice rather than left at whatever the default happens to be, because a restored machine wanting a superseded password is the scenario history exists for.

  4. 4

    Deploy and have the help desk test it

    A pilot group receives the policy by whichever mechanism suits, we confirm passwords are genuinely landing in the chosen directory, and then a real help desk agent retrieves one and uses it. Not the engineer who built the configuration, who already knows where everything is.

  5. 5

    Broaden, migrate legacy, and cover domain controllers

    The rollout widens with coverage reported as a measured figure instead of an assumption, any legacy deployment moves across through emulation mode and is then properly retired, and the restore mode password on your domain controllers comes under management as the final step.

Straight answers

What organizations ask about Windows LAPS.

Nothing, in licensing terms. The capability ships free on every supported Windows platform, keeping passwords in your directory adds no requirement at all, and the cloud destination asks only for the free identity tier. Weighed against what it removes from your risk register, there is very little else that returns this much for nothing.

You do not. Anything from Windows 11 23H2 onward has it outright, and Windows 11 22H2, 21H2 and Windows 10 all qualify provided they have taken the April 11 2023 update or anything later. Servers follow the same shape, with 2025 supported outright and 2022 and 2019 supported once that update is applied.

You cannot, and the documentation says so without hedging. Cloud-joined hardware writes to the cloud directory, domain-joined hardware writes to the domain, and hybrid machines get to pick between the two but only ever end up with one. Design for that constraint rather than discovering it partway through a rollout.

They are outside the scope of the feature entirely. Those machines need their own answer, and the genuinely important part is writing them down as uncovered somewhere visible. The failure mode we see is a tenant with a policy in it, which creates a comfortable impression of coverage that does not survive contact with an auditor.

For anything new, no. Deprecation landed at Windows 11 23H2, the package will not install on newer builds, and code changes are no longer being considered. What support remains covers older Windows versions and expires when those versions reach their own end of support, so there is no long-term position available here.

There is an emulation mode built for exactly this, so an existing deployment can be moved across in stages rather than cut over. Worth knowing too that the modern feature never needed the old one present in the first place, which means anybody starting fresh can disregard this question completely.

Three named additions: passwords can go to the cloud directory, passwords in the domain can be encrypted, and history is retained. Underneath those, the more consequential difference is that this is a distinct implementation living inside Windows rather than a product you install on top, which is why it arrives and stays current through ordinary servicing.

Because of restores. Bring a machine back from an image captured before the most recent rotation and it will be waiting for the password it had at the time. With no history retained, that endpoint may be unreachable through precisely the account this whole project existed to guarantee you access with. It is the neatest available example of a security improvement becoming an availability incident.

Exactly the people you decide on, no more. Cloud-stored passwords are governed by the role-based access control model you already run; directory-stored ones by granular access control lists, with encryption available on top. Getting this right is more important than any other decision in the deployment, and it is the one most often left until last.

Pass-the-hash and lateral traversal, named first among the benefits. Trace the mechanics: one password common to every workstation means whatever an attacker lifts from their initial foothold authenticates everywhere else in the building. Distinct rotating credentials do not make that harder to spot, they delete the path. Which is precisely why the question now turns up on insurance applications and inside NIST CSF and CIS control mappings.

Yes, for the Directory Services Restore Mode account, which gets managed and backed up like any other and can then be retrieved and used by an authorized administrator. Given that this password is typically set once during promotion and never revisited, bringing it into scope is one of the higher-value additions available for almost no extra effort.

Cloud managed machines take it through the configuration service provider, domain managed ones through Group Policy. One caveat that catches people out: the Automatic Account Management settings on the CSP side need Windows 11 24H2 or later, so that particular subset applies to a noticeably narrower slice of the estate than the feature as a whole.

Three routes on the domain side: a dedicated event log channel, a purpose-built PowerShell module, and the properties dialog in Active Directory Users and Computers. Where passwords are going to the cloud directory instead, monitoring and reporting come from there. The event channel is the one worth wiring into whatever you already alert on.

Four to six weeks is normal, and very little of that is engineering. Counting join states, settling destinations, working out permissions and getting a real agent to test a real recovery all take longer than the policy push itself, which is roughly an afternoon. That imbalance is not waste, it is what makes the thing survive its first year.

Quoted per engagement against estate size and how tangled the join states are, since a heavily hybrid environment needs a decision taken group by group. Before contacting anybody, try the free version of the assessment. Pick two machines at random and compare their local administrator passwords. If they match, you have already answered the question of whether this is worth doing.

On older Windows versions where it was already supported, yes, and that support runs until those versions themselves go out of support. Since the installer is blocked from Windows 11 23H2 onward, though, your legacy footprint can only shrink from here. Treat it as a wind-down rather than a strategy.
Before you deploy

Fifteen questions worth answering first.

Answers in the first group tell you how easy this will be. Answers in the third tell you whether your team can still rescue a machine once it is done, and that carries exactly as much weight.

Estate

  • What is the join state of each device group?
    It decides the destination.
  • Are all devices on the April 2023 update or later?
    Or Windows 11 23H2 plus.
  • Do we have workplace joined clients?
    They are not supported.
  • Are domain controllers in scope?
    DSRM can be managed too.
  • Do we still run legacy Microsoft LAPS?
    It is deprecated.

Design

  • Entra ID or Active Directory backup?
    Never both.
  • Will we encrypt passwords in Active Directory?
    It is optional.
  • How much password history do we keep?
    It helps with restores.
  • Policy by CSP or Group Policy?
    Both are supported.
  • Which local account are we managing?
    Decide deliberately.

Operations

  • Who is permitted to retrieve a password?
    By RBAC or ACL.
  • Has a real help desk agent tested retrieval?
    Not the project team.
  • Is the event log channel monitored?
    Failures surface there.
  • What happens when a device is offline for months?
    History matters.
  • How do we measure coverage?
    A number, not an assumption.
Related reading

The pages around this one.

Privileged Identity Management

The equivalent discipline for directory roles.

Learn more

Active Directory security audit

The wider review that usually surfaces this finding.

Learn more

Intune security baselines

The hardening standard local administrator management belongs inside.

Learn more
Next step

Check whether two randomly chosen machines share a local administrator password.

If they do, one compromised endpoint is every endpoint. Windows LAPS is free on every supported Windows platform, and this is the fix.

Book a Windows LAPS deploymentSee what it covers

Related Services

Explore more solutions that work great with this service

Microsoft Entra Privileged Identity Management

Privileged Identity Management deployment for US organizations:

Learn more

Active Directory Security Audit

On-premises Active Directory security audit for US organizations:

Learn more

Privileged Access Audit

Privileged access audits for US organizations: enumeration of every

Learn more

Intune Security Baselines

Security baseline design and management for US organizations: stating

Learn more

Endpoint Security

Endpoint security for US businesses using Microsoft Defender for

Learn more

Microsoft Entra

Identity and access management solutions

Learn more
GR IT SERVICES

IT services for US businesses,
delivering enterprise-grade solutions
remotely, coast to coast.

Microsoft CSP PartnerApple Jamf PartnerCISGuard

Microsoft 365

  • Microsoft 365 Administration
  • M365 Reporting & Auditing
  • Microsoft 365 Licensing
  • Microsoft Copilot
  • Microsoft 365 Apps
  • Windows 365 Cloud PC
  • Microsoft SharePoint
  • Outlook & Exchange

Security

  • Microsoft Defender
  • Microsoft Purview
  • Microsoft Intune
  • Microsoft Entra
  • Compliance Manager
  • Cybersecurity Audits
  • Copilot for Security
  • Microsoft Sentinel
  • Microsoft Priva

Infrastructure

  • Google Workspace
  • Cloud Migration Services
  • Data Analytics & BI
  • Active Directory
  • Server Management
  • Apple Business
  • Apple Jamf Pro
  • IP Telephone
  • Data Backup
  • Website Development

IT Services

  • Managed IT Services
  • IT Support USA
  • IT AMC USA
  • New Office IT Setup
  • IT Relocation
  • Remote IT Support
  • On-Call IT Support
  • Startup IT Business Kit
  • Disaster Recovery & BC

Company

  • About Us
  • Careers
  • Contact
  • Blog

Contact

  • hello@gritservices.io
  • gritservices.io

© 2026 GR IT Services. All rights reserved.

Privacy PolicyTerms of UseCookie PolicyCCPA/CPRA