Technology

Windows 11 Is Logging Which Coding Agents You Use — What It Records and How to Stop It

By Asad Ali ·

27 min readLast updated

Cover image for the article on Windows 11 logging which coding agents you use, showing MissionControlAggregator telemetry and how to turn off optional diagnostic data.

Yes — Windows 11 records which AI coding agents and developer tools you run, and queues that list for upload to Microsoft unless you have turned off optional diagnostic data. What it records is narrow: a set of yes-or-no flags saying tools such as Claude, Codex, Copilot or Cursor were used. It does not record your prompts, your code, your file names, or how long you worked.

Security researcher Xusheng Li found the behaviour by reverse-engineering a Windows system component, and published the full write-up on 16 November 2026. The finding reached the Hacker News front page the same day. This post explains what the component does, what it does not do, whether it is running on your machine, and how to switch it off in two minutes.

READ THIS FIRST — what is true and what is hype

True: a Windows 11 component called MissionControlAggregator.dll keeps a daily tally of which coding agents and developer tools have run on your PC, and queues that tally for Microsoft as optional diagnostic data.

Also true: the tally is a bitmap — one bit per tool. The researcher found no prompt text, no source code, no file names and no session durations in the report.

Hype to ignore: claims that "Windows reads your code" or "Microsoft watches everything you type" are wrong. Nothing in the report contains your work. And the feature is not proven to be active on every Windows 11 PC — the researcher's own clean install from the public ISO had it switched off by default.

Still unknown: why Microsoft collects this. As of the morning of 17 November 2026 (UTC), Microsoft had not publicly commented on the finding.

The short answer: yes, it logs your tools — as bits, not your code

Windows 11 has a built-in diagnostics component that watches for a fixed list of executable names. When it sees one run — claude.exe, codex.exe, copilot.exe, cursor.exe and a handful of others — it sets a bit in a small block of state saved on your disk. Roughly once a day, it turns those bits into a short report and hands it to Windows Diagnostics. If your PC is set to send optional diagnostic data, that report is queued for upload to Microsoft.

Think of it as a census, not surveillance footage. A set bit means "this tool was observed during the reporting window". It does not say what you typed into the tool, what files you opened, what the tool produced, or how long you used it. The report's timestamps describe the collection window, not your working hours.

That distinction is why this story is both real and routinely overstated. The collection is genuine, documented in the binary and demonstrated end to end. The idea that Microsoft is reading developers' code through it is not supported by anything in the report.

Who found it, and how

The researcher is Xusheng Li, a security researcher and a developer on the Binary Ninja reverse-engineering platform. On 16 November 2026 he published a detailed write-up on his site, xusheng.dev, titled "Windows 11 Is Collecting Data on Which Coding Agents and Developer Tools You Use". He had shared the initial finding on X the same day, and the story reached the Hacker News front page within hours — one write-up published on the morning of 17 November recorded 523 points and 361 comments on the thread.

Li's starting point was unusual: he asked an AI assistant to look through recent Windows binaries for undocumented features. One lead stood out immediately — a list of coding-agent executable names embedded in a Windows system DLL, sitting next to familiar developer tools such as python.exe, dotnet.exe, gcc.exe and code.exe. He then opened the DLL in Binary Ninja and followed the references to see what the code actually did with those names.

That method matters for judging the finding. This was not a guess from a settings screen or a line in a privacy policy. It was a read of the compiled code, followed by live tests on a real Windows 11 virtual machine: the researcher renamed a harmless program to match one of the watched names, ran it, and watched the corresponding bit flip in the component's saved state and then appear in the diagnostic report.

It is also worth saying what kind of evidence this is: one researcher, one virtual machine, one Windows 11 build. The finding proves the mechanism exists and runs. It does not prove how widely Microsoft has switched it on, which editions or regions are affected, or what Microsoft intends to do with the data.

The component doing the logging: MissionControlAggregator

The component is MissionControlAggregator.dll, a Windows system library. It is loaded by a process called AggregatorHost.exe, which is started by DiagTrack — the Connected User Experiences and Telemetry service, the Windows service that handles diagnostic data. On the researcher's test machine, DiagTrack was already running with its startup type set to Automatic, and the chain from service to child process to loaded DLL was confirmed with standard administrative commands.

The agent checks sit behind a feature flag with an unusually direct name: CLICodingAgentUsage, feature ID 62580080. Feature flags like this are how Microsoft stages new behaviour: the code ships in the operating system, and the flag decides whether it runs. On Li's test virtual machine the flag was already enabled as a service experiment, without him touching it. On a clean install built from the public ISO, the same checks were disabled by image default. That split is the clearest sign that this is a staged rollout, not a switch flipped everywhere at once.

The DLL version on the test machine was 10.0.26100.9818, running on a Windows 11 ARM64 virtual machine, build 26200.9974. The version numbers matter because they pin the finding to a specific, recent Windows 11 build — this is not an old behaviour that someone only just noticed, and it is not something demonstrated on Windows 10. Everything Li tested was Windows 11.

The design has two details worth noting. First, the component does not scan your running processes. It subscribes to event streams Windows already produces, and matches the application identities inside those events against its built-in lists. Second, some of the matches also verify the publisher: the Node.js slot expects the publisher "Node.js Foundation", Python expects "Python Software Foundation", and the .NET and Visual Studio slots expect "Microsoft Corporation". The compiler and debugger group, which covers GCC and Clang, has no publisher check on that path.

Exactly what Windows records: the slots

The component organises what it watches into three groups: four command-line agent slots, five desktop agent slots, and twelve broader developer-tool categories. The internal reporting function is literally named ReportCohorts<12,4,5> — twelve categories, four CLI agents, five desktop agents. When a match succeeds, the component sets a slot in its saved state; later, the report turns those slots into bitsets.

Group Slots Executables watched What the bit means
CLI coding agents 4 copilot.exe, claude.exe, codex.exe, agy.exe (plus extensionless equivalents on the WSL route) A matching command-line agent was observed
Desktop agents 5 github.exe, claude.exe, codex.exe, cursor.exe, antigravity.exe A matching desktop agent was observed
Developer-tool categories 12 Named in the write-up: node.exe, python.exe, dotnet.exe, gcc.exe, clang.exe, gdb.exe, javac.exe, devenv.exe, code.exe, among others A tool in that category was observed (one shared bit per category, not per program)

A few details in that table deserve emphasis. The CLI list includes agy.exe alongside the three big names, and the desktop list includes github.exe and antigravity.exe next to Claude, Codex and Cursor. The same executable can appear in more than one group: claude.exe and codex.exe are checked on both the CLI and the desktop routes, because the same tool can run as a terminal program or as a desktop app.

The twelve developer-tool categories work differently from the agent slots. A category shares one bit across several programs: the compiler and debugger category, for example, covers GCC, Clang and related tools under a single bit. The report can therefore say "a compiler was used" without saying which one. The agent slots, by contrast, are one bit per named agent.

Where does the component learn these names? It consumes three event sources. For Windows programs it reads KernelProcess.AppStateChangeSummary and Win32kTraceLogging.AppInteractivitySummary, which carry application identities. For the Windows Subsystem for Linux it reads Microsoft.Windows.Lxss.Manager.Exec and ExecCritical, using the imageName field. A native Windows codex.exe reaches the CLI checks through the app-state route, so the collection covers ordinary Windows programs too, not just WSL.

What Windows does not record

The most important sentence in the researcher's write-up is a negative one: "Although I found no prompt text or source code in this report, collecting which agents and developer tools someone uses already has significant privacy implications." Both halves are doing work. The first half limits the finding; the second half explains why it still matters.

Here is what is not in the report. There are no executable names in the payload — only bit positions. There are no prompts, no code, no file names, and no session durations. The source events Windows produces do include runtime, time in focus, interaction counters and resource usage, but this component appears to discard those details and record only the bitmap of which agents were observed. The report's timestamps describe the collection window, not how long each agent was used.

That is a narrow payload, and it is the reason the scariest versions of this story are wrong. A bitmap cannot reveal what you asked Claude to write, what Codex changed in your repository, or what your Cursor session contained. Anyone claiming otherwise has not read the report — or has read it and chosen to exaggerate it.

But narrow is not harmless. A list of the tools you use still reveals your stack: which AI assistant you prefer, whether you work in Python or .NET, whether you use Visual Studio or VS Code. For a team, the same bits aggregated across machines reveal which vendor is winning inside a company. For someone under a client NDA, even a one-bit-per-tool record is a disclosure they did not agree to make. The researcher says so plainly, and he is right: the privacy question is real even though the payload is small.

How the report gets to Microsoft

The report is an Event Tracing for Windows (ETW) event — the diagnostic tracing system built into Windows. In this case it is a self-describing TraceLogging event named Segments.SubCohorts, emitted by the provider Microsoft.Windows.MissionControlAggregator. Roughly once a day, a timer fires and the aggregator converts its saved slots into the bitsets described above.

The state is saved on disk first, at C:\ProgramData\Microsoft\Diagnosis\AggregatorStorage\MissionControlAggregator. That local file is where the bits accumulate between reports — and what makes the behaviour checkable: the researcher could read the saved state after his test and see the Codex slot set before any report was emitted.

An ETW event alone would not prove anything leaves the machine — ETW is used for plenty of local diagnostics. So Li checked who was listening: a DiagTrack listener session was already subscribed to both ends of the path. Then he inspected DiagTrack's EventStore.db, decompressed the JSON batches in its AsimovEvents table, and found the agent event itself sitting in the upload queue.

That queue entry carried the decisive detail: "isRequired": false. The report is treated as optional diagnostic data. When optional diagnostic data is allowed, Windows queues the coding-agent usage report for upload to Microsoft. When it is not allowed, the report is not queued.

One honest caveat belongs here. Li observed the report in the upload queue; he did not publish a network capture of it leaving the machine. Queued is what is proven; uploaded is what follows from how the queue works. That is a strong inference — DiagTrack's job is to upload its queue — but it is not the same as a packet capture, and this post will not pretend otherwise.

From your terminal to Microsoft: the data pathYou run an agentclaude · codex · cursorWindows eventsapp state + WSL execMissionControlAggregatorname matchersBitset reportSegments.SubCohortsMicrosoftonly if optionaldata is onOne bit per tool, about once a day. No prompts. No code. No file names.The report is queued as optional diagnostic data (isRequired: false).Turn off “Send optional diagnostic data” and the queue entry never forms.Source: Xusheng Li, “Windows 11 Is Collecting Data on Which Coding Agentsand Developer Tools You Use”, 16 November 2026 (xusheng.dev).

The renamed-cmd.exe test: the match is by name only

The most telling experiment in the write-up is also the simplest. Li copied cmd.exe — the ordinary Windows command prompt, which contains no AI code at all — to a file named codex.exe and ran it in an interactive Windows session. The name alone was enough for the component to recognise it as Codex.

He captured an AppStateChangeSummary for the running process and found the aggregator's saved state contained the value 2 in CLI slot 2, the Codex slot. A separate run without an additional ETW trace also set that slot, confirming that detection worked through Windows' existing diagnostic collection rather than through the researcher's own tracing.

To capture the report immediately — it normally follows a roughly daily boundary — he temporarily moved the saved reporting-window start 25 hours into the past and restarted DiagTrack with a trace listening for the aggregator. The resulting Segments.SubCohorts event read: provider Microsoft.Windows.MissionControlAggregator, version 0.1.5, CliAgentSubCohorts: 4, DesktopAgentSubCohorts: 0. The value 4 is 1 << 2: the Codex CLI bit. The name had made it through the input summary, the matcher, the persisted state and the emitted report.

He tested the developer-tool path the same way, using a copy of cmd.exe named gcc.exe. It set developer-tool slot 3, and with optional diagnostic data allowed, the resulting event appeared in DiagTrack's upload queue with SubCohorts: [8] — 1 << 3, the compiler and debugger category shared by GCC, Clang and related tools.

Two lessons follow. First, the system matches names, not behaviour: a renamed command prompt counts as Codex and as GCC. That weakens the data's accuracy — and it also weakens any claim that the data is deeply invasive, since the report cannot tell a real coding agent from a file that shares its name. Second, the test is reproducible by anyone with a Windows 11 machine and administrative rights, which is why the finding spread so fast.

Required vs Optional diagnostic data: which level sends this

Whether the report leaves your machine depends on one setting: your diagnostic data level. Microsoft splits Windows diagnostic data into two tiers, Required and Optional, and its own support documentation describes the difference plainly.

Level What it includes Coding-agent report? Can you turn it off?
Required Device, settings and capability information, and whether the device is performing properly — the minimum needed to keep it reliable, secure and operating normally No — the report is marked isRequired: false No — this is the minimum level on Home and Pro
Optional Everything in Required, plus additional details about the device, its settings, capabilities and health — including device activity (sometimes called usage) and enhanced error reporting Yes — queued for upload when this level is on Yes — Settings → Privacy & security → Diagnostics & feedback → Send optional diagnostic data: Off

Microsoft's documentation already discloses, in general terms, that optional diagnostic data can include device activity — which programs launch, and how they are used. What it does not disclose is that coding agents have their own dedicated slots, distinct from the generic app stream. The behaviour fits inside the published disclosure; the specificity does not appear in it. That gap between the letter of the disclosure and what a reader would understand from it is the heart of the story.

Two further points from Microsoft's own page are worth knowing. First, required diagnostic data is always included when you send optional data, and your device is just as secure and operates normally on Required alone — turning off Optional costs you nothing in security or functionality. Second, Microsoft says it collects some optional data from only a small percentage of devices as a sample, so even with Optional enabled, a given device may not be in the sample for a given data item. That sampling does not change the finding, but it means two identical machines can behave differently.

Is it running on your PC right now?

Honest answer: it depends, and you can check in under a minute. The evidence does not support claiming the feature runs on every Windows 11 machine — and the researcher says so himself.

On Li's Parallels Windows 11 ARM64 virtual machine, build 26200.9974, the CLICodingAgentUsage feature was already enabled as a service experiment, and DiagTrack was running automatically. He had not manually enabled anything. But on a clean installation built from the public ISO, the coding-agent checks were disabled by image default. Two machines, two different states, same operating system.

That is how Microsoft rolls features out: code ships to everyone, a service-side experiment switches it on for a subset, and the subset grows over time. So the question "is it on my PC?" has no universal answer — it must be checked per machine, and the answer can change after an update.

The second half of the question is the diagnostic data setting, and here the default is clearer. On a clean installation of Windows 11 Home ARM64 from the latest official ISO, with the region set to the United States, the Diagnostic data toggle was already enabled and labelled "Include Optional" before Li changed any privacy settings. After completing setup with those defaults, the AllowTelemetry value was 3, confirming optional diagnostic data was enabled. So in the US setup flow he tested, the upload path was open by default — but only for machines where the feature flag itself is switched on.

You can confirm the service chain on your own machine with three administrative PowerShell commands, exactly as the researcher ran them:

Get-Service DiagTrack | Format-Table Name, Status, StartType -AutoSize
Get-Process AggregatorHost | Format-Table ProcessName, Id -AutoSize
tasklist /m MissionControlAggregator.dll

If DiagTrack is running, AggregatorHost is its child process, and the module listing places MissionControlAggregator.dll inside that child, the pipeline is live on your machine. Whether the agent checks themselves are enabled is a separate question — the flag is a service experiment with no settings page. The check that matters most is the diagnostic data toggle, covered below.

What about Windows 10?

Nothing in the research covers Windows 10. Every test Li ran was on Windows 11: the ARM64 virtual machine on build 26200.9974, and the clean Windows 11 Home install from the latest official ISO. The DLL version, 10.0.26100.9818, belongs to the Windows 11 codebase. There is no evidence in the write-up that MissionControlAggregator.dll, the CLICodingAgentUsage flag, or the Segments.SubCohorts report exists on Windows 10 — and no evidence that it does not.

That is the correct way to state it: unknown, not absent. Microsoft's documentation notes that Windows 10 is being updated to the newer Required and Optional categorisation, which shows the company still changes telemetry behaviour on Windows 10 — but a general telemetry update is not this specific feature, and this post will not claim a Windows 10 connection nobody has demonstrated.

If you still run Windows 10, the practical advice is the same as for Windows 11: check the diagnostic data setting, because the Optional tier on Windows 10 already covers app activity in general terms. The coding-agent slots, as far as anyone has shown, are a Windows 11 matter.

Why would Microsoft want this list?

Microsoft has not said. As of the morning of 17 November 2026 (UTC), there was no public statement from the company about the feature, and the write-up's comment sections are full of speculation rather than answers. Here are the plausible explanations, labelled as what they are — theories, not findings.

The most likely explanation is market measurement. Coding agents are the most contested product category in developer tools right now: Microsoft's own Copilot competes directly with Claude Code, Codex, Cursor and Antigravity. Fixed bit positions — one slot per agent, stable across machines and time — are exactly what you build when you want a comparable time series of who is winning. One independent write-up put it bluntly: Microsoft is building census infrastructure for the agent wars.

A second possibility is compatibility and quality work: knowing which developer tools are popular helps Microsoft prioritise app-compatibility testing and feature improvements, which is precisely how its own documentation describes the purpose of optional diagnostic data. That explanation is consistent with the twelve broader tool categories, which look like a general developer census. It fits the compiler and debugger buckets far better than it fits four named AI agents.

A third possibility is product planning for Windows itself: if Microsoft is deciding where to invest in developer-facing features, knowing the shape of the developer population is useful. None of these explanations requires reading anyone's code, and none of them is confirmed. What is confirmed is only the mechanism and the data it carries. Until Microsoft comments, the purpose is an open question — and any post that states a purpose as fact is guessing.

How to turn it off

Switching this off takes about two minutes, and it changes nothing about how your PC works. Microsoft's own documentation says your device will be just as secure and will operate normally on Required diagnostic data alone. These are the steps, in order.

Step 1 — Turn off optional diagnostic data. Open Settings, go to Privacy & security, then Diagnostics & feedback. Under Diagnostic data, set "Send optional diagnostic data" to Off. This is the switch that controls the coding-agent report: with it off, the report is not queued for upload. It also stops the rest of the optional tier — device activity, enhanced error reporting and the rest — so it is a broader privacy win than just this one report.

Step 2 — Decline it at setup, on your next install. During Windows setup you are offered the same choice: the Diagnostic data screen lets you decline optional data before the first boot. On the US setup flow Li tested, the preselected choice was to include it, so this is a screen worth reading rather than clicking through. If you set up a machine for someone else, set it for them.

Step 3 — Check the setting stuck. The registry value behind the setting is AllowTelemetry. On a machine where the policy is set, you can read it with PowerShell:

Get-ItemProperty -Path "HKLM:\SOFTWARE\Microsoft\Windows\CurrentVersion\Policies\DataCollection" -Name AllowTelemetry -ErrorAction SilentlyContinue | Select-Object -ExpandProperty AllowTelemetry

A value of 3 means optional (Full) diagnostic data is enabled. A value of 1 means Required only. Note that the value may simply be absent on a machine that has never had the policy set — in that case, trust the Settings screen, not the registry. And if your machine is managed by an employer, the setting may be greyed out with a "managed by your organization" note; in that case only your IT department can change it.

Step 4 — Delete what is already there. Turning the setting off stops future uploads; it does not erase what has already been sent. In the same Diagnostics & feedback screen, under "Delete diagnostic data", select Delete to remove the diagnostic data stored for your device. Microsoft's documentation is clear about the limit: this does not delete diagnostic data associated with your Microsoft account. For that, visit the Microsoft privacy dashboard at account.microsoft.com and use its data controls.

Step 5 — Look at what your machine actually sends. Microsoft ships a free Diagnostic Data Viewer that shows your device's diagnostic data in real time. It only shows data while the viewer is running — it cannot show history — but it is the most direct way to see what your machine is producing.

One thing not to do: do not rename your coding agents to dodge the matcher. It would work — the test proves the match is by name only — but it would break updates, file associations and every script that calls the tool by name. The settings switch is the correct fix, and it is a supported one.

If you run a team or a company

For an individual, this is a two-minute settings change. For a team, it is a policy question, and the stakes are higher: the same bits that reveal one developer's preferences reveal a company's stack when aggregated across a fleet.

First, inventory your estate. On managed devices the level is controlled by policy, and it can be enforced centrally through Group Policy or mobile device management using the AllowTelemetry policy. If your organisation has not set a diagnostic data policy, the default on a US setup is to send optional data — so the coding-agent report is likely flowing from at least some of your machines today.

Second, treat the tool list as confidential information. Many client contracts and NDAs cover the tools and methods used on a project, not just the code produced. A diagnostics report that names the AI assistants your engineers use is a disclosure of working methods, and it is one your contracts probably did not anticipate. That is a conversation for your legal or compliance colleagues, not just for IT — and it is worth having before a client asks about it.

Third, re-check after updates. Feature flags are service experiments: a machine that has the checks off today can have them switched on by a later update, and a diagnostic data policy is the durable control, not a one-time settings change. Put the diagnostic data level in your standard device checklist, alongside disk encryption and update policy, and verify it on a sample of machines each quarter.

Finally, do not overreact. The report contains no code, no prompts and no client data. It is not a breach, and it is not evidence that Microsoft has seen your source. It is a telemetry behaviour that your organisation did not choose and probably did not know about — which is exactly why it deserves a policy decision rather than a panic.

The bottom line

Windows 11 does record which coding agents and developer tools you run. The mechanism is real, demonstrated end to end, and documented in the operating system's own code: a diagnostics component watches for a fixed list of executable names, sets one bit per tool, and queues a daily report for Microsoft whenever optional diagnostic data is enabled — which, in the US setup flow, is the default.

What it records is narrow: no prompts, no code, no file names, no durations — one bit per tool per day. The scariest readings of this story are wrong. Your choice of tools is your business, and until 16 November 2026, almost nobody knew Windows was keeping a list of them.

The fix is simple, supported and free: turn off "Send optional diagnostic data" in Settings → Privacy & security → Diagnostics & feedback. Do it on your own machines, and if you run a team, set it by policy. Then get back to work — with the tool of your choice, and without Windows taking notes on which one it is.

Frequently Asked Questions

Is Windows 11 really logging which coding agents I use?

Yes, on machines where the feature is switched on. A Windows 11 component called MissionControlAggregator.dll watches for specific executable names — including claude.exe, codex.exe, copilot.exe and cursor.exe — and records a yes-or-no flag for each one it sees. The flag is queued for upload to Microsoft as optional diagnostic data. The feature is a staged rollout, so it is not proven to be active on every Windows 11 PC.

Does Microsoft see my prompts or my code?

No. The researcher found no prompt text, no source code, no file names and no session durations in the report. The report is a bitmap: one bit per tool, saying only that the tool was observed during the reporting window. The source events Windows produces do include runtime and focus details, but this component discards them and records only the bitmap.

How do I turn it off?

Open Settings → Privacy & security → Diagnostics & feedback, and set "Send optional diagnostic data" to Off. That stops the coding-agent report from being queued, along with the rest of the optional diagnostic tier. Your PC stays just as secure and works normally on Required diagnostic data alone, according to Microsoft's own documentation.

Is this happening on Windows 10 too?

Unknown. All of the researcher's tests were on Windows 11 — an ARM64 virtual machine on build 26200.9974 and a clean Windows 11 Home install. There is no evidence the component or the feature exists on Windows 10, and no evidence that it does not. The Optional diagnostic tier on Windows 10 already covers app activity in general terms, so checking the setting is still good advice there.

Can I check whether it is running on my machine?

Partly. You can confirm the pipeline is live with three administrative PowerShell commands: check that the DiagTrack service is running, that AggregatorHost.exe is its child process, and that MissionControlAggregator.dll is loaded in that process. Whether the agent checks themselves are enabled is controlled by a Microsoft service experiment, and there is no settings page that shows it. The check that matters most is the diagnostic data toggle in Settings.

Has Microsoft said anything about this?

Not as of the morning of 17 November 2026 (UTC). No public statement from Microsoft about the feature had appeared at that point. Microsoft's published documentation does disclose in general terms that optional diagnostic data can include device activity such as which programs launch — but it does not mention that coding agents have their own dedicated reporting slots.

Sources

Related reading on Asad Webs: our guide to Claude-BugHunter's 83 AI security skills for Claude Code, our roundup of the best DNS services for privacy and security in 2026, and Apple's privacy features, from Lockdown Mode to Advanced Data Protection. We checked the tools index at asadwebs.com/tools — there is no Windows telemetry or diagnostic-data tool there, so no tool is linked in this article.

Need a website that respects your visitors' privacy? Asad Webs builds fast, search-friendly websites, eCommerce stores and custom web apps. Get in touch for a free consultation.

Category: Technology

Tags: technology

About Asad Ali

Asad Ali is a graphics designer with over 20 years of experience — designing since 2003 — and the founder of Asad Webs, based in Pakistan. He builds the free online tools on this site and writes the guides published here.

LinkedIn Facebook Website
KEEP READING

Next guide

Best Webcam and Microphone for Remote Developers in 2026: Look and Sound Like a Pro on Every Call →

Related articles

Join the conversation

No account needed. Comments appear only after approval. Your email address will not be published.

Be the first to share a useful question or thought.

You won’t be subscribed to any mailing list.
WELCOME BACK

Sign in

Access your orders, saved products and account.