{"id":948,"date":"2026-11-04T13:17:47","date_gmt":"2026-11-04T13:17:47","guid":{"rendered":"https:\/\/blog.voicenotes.com\/blog\/?p=948"},"modified":"2026-11-04T13:17:47","modified_gmt":"2026-11-04T13:17:47","slug":"are-ai-meeting-notetakers-safe-for-confidential-business-meetings","status":"publish","type":"post","link":"https:\/\/blog.voicenotes.com\/blog\/uncategorized\/are-ai-meeting-notetakers-safe-for-confidential-business-meetings\/","title":{"rendered":"Are AI Meeting Notetakers Safe for Confidential Business Meetings?"},"content":{"rendered":"\n

What businesses should know about AI notetaker privacy and what the published policies of Voicenotes, Granola, Fireflies and Otter <\/em>says.<\/em><\/p>\n\n\n\n

<\/p>\n\n\n\n

\"\"
Photo by Compare Fibre<\/a> on Unsplash<\/a><\/figcaption><\/figure>\n\n\n\n

Halfway through a board meeting, someone mentions a possible acquisition. The finance director shares next quarter\u2019s projections. A colleague explains why a major client might leave.<\/p>\n\n\n\n

And the AI notetaker has been recording the whole conversation<\/p>\n\n\n\n

An hour later, you have a tidy summary, a transcript and a list of actions. But there’s something else you’ve created: a digital record of a confidential conversation. It may have passed through multiple systems to become the neat summary sitting in front of you.<\/p>\n\n\n\n

Where did the audio go? Which companies processed it? Could the conversation help train an AI model? And if you delete the meeting tomorrow, what exactly disappears?<\/p>\n\n\n\n

\n
\n
\n

Full disclosure: <\/strong>I\u2019m Dona Sussan, Head of GTM at Voicenotes. Privacy comes up in almost every conversation I have with teams evaluating us, so I decided to compare our commitments with those of Granola, Fireflies and Otter. I wanted to see where we stand, where others are doing better and where we still have work to do.<\/p>\n\n\n\n

I thought it would take a couple of hours. It didn\u2019t. Privacy policies led to help pages, help pages led to fine print, and the fine print sometimes sent me right back to where I started.<\/p>\n\n\n\n

I reviewed each company\u2019s privacy policy and security documentation as of November 2026. I\u2019m sharing what I found because buyers face the same exercises when comparing AI notetakers, and that includes the areas where our competitors are ahead of us.<\/p>\n\n\n\n

Reviewed by Jishnu Krishnan, Data Protection Officer, Voicenotes.<\/em><\/p>\n<\/div>\n<\/div>\n<\/div>\n\n\n\n

Where does the data go?<\/h2>\n\n\n\n

A meeting recording can create several separate records:<\/p>\n\n\n\n

Captured audio \u2192 transcript \u2192 AI-generated notes \u2192 stored meeting \u2192 shared or exported copies<\/strong><\/p>\n\n\n\n

Different platforms may handle different stages. The notetaker might store your recordings, a transcription provider might convert speech into text, and another AI service might generate the summary.<\/p>\n\n\n\n

That does not automatically make the arrangement unsuitable for confidential work. It does mean that approving the app involves approving more than the interface on your screen.<\/p>\n\n\n\n

I would want a review to establish:<\/p>\n\n\n\n

Which platforms receive the audio, transcripts and prompts.<\/p>\n\n\n\n

Where the information is stored and processed.<\/p>\n\n\n\n

How long each platform retains it.<\/p>\n\n\n\n

Who can access the resulting records.<\/p>\n\n\n\n

What happens to copies shared with other systems.<\/p>\n\n\n\n

One detail that stood out was the difference between storage and processing locations.<\/p>\n\n\n\n

Fireflies<\/strong>, for example, says its Enterprise private-storage option can keep meeting data in a customer-controlled storage bucket, while processing still takes place on Fireflies\u2019 servers in the United States. Choosing where information is stored therefore does not necessarily determine where it is processed. Fireflies private-storage guide<\/a><\/p>\n\n\n\n

Granola<\/strong> describes cloud storage for transcripts and notes. Voicenotes also describes cloud storage and external AI processing. Granola data-handling explanation<\/a>, Voicenotes privacy policy<\/a><\/p>\n\n\n\n

Recording and consent: Does everyone actually agree?<\/h2>\n\n\n\n

I wanted to understand how participants are asked about recording in the first place.<\/p>\n\n\n\n

There’s a fundamental difference between informing someone, giving them a chance to object, and asking for their permission.<\/p>\n\n\n\n

Fireflies<\/strong> offers optional pre-meeting emails with an opt-out link. Participants can decline recording, but what happens if someone never opens the email? Unless they object, recording may still proceed.<\/p>\n\n\n\n

Otter<\/strong> makes its Notetaker bot visible and supports meeting notifications. But visibility isn’t consent. Its desktop recording option can also capture meetings without a bot, placing responsibility for informing participants on the host.<\/p>\n\n\n\n

Granola<\/strong> takes a botless approach. Depending on the platform, it can display chat notifications or recording watermarks. But without a visible bot, participants may not even know an AI notetaker is capturing the conversation unless they’re explicitly informed.<\/p>\n\n\n\n

Voicenotes<\/strong> lets teams choose between two optional consent modes.<\/p>\n\n\n\n

Opt-in: <\/strong>Recording requires everyone\u2019s approval<\/p>\n\n\n\n

Before the meeting, each attendee receives an email asking for permission to record. The bot joins only when everyone has explicitly agreed. If anyone declines or doesn\u2019t respond, the bot stays out and the meeting will not be recorded.<\/p>\n\n\n\n

Best for sensitive conversations, external guests, or any meeting where clear consent matters most.<\/em><\/p>\n\n\n\n

Opt-out:<\/strong> Recording proceeds unless someone declines<\/p>\n\n\n\n

Before the meeting, each attendee receives an email letting them know it\u2019s scheduled to be recorded, along with the option to decline. No response is needed to proceed. If everyone is comfortable, the bot joins and records as planned. If anyone declines, the bot stays out and the meeting will not be recorded.<\/p>\n\n\n\n

Best for routine internal meetings where recording is expected, but attendees should still have the final say.<\/em><\/p>\n\n\n\n

Every consent decision is logged for future reference.<\/p>\n\n\n\n

The distinction is simple: silence isn’t treated as consent.<\/strong><\/p>\n\n\n\n

For a routine team catch-up, notification-based recording may be appropriate. But when a board is discussing an acquisition or a client is sharing commercially sensitive information, the stakes are different.<\/p>\n\n\n\n

A privacy team needs more than proof that a notification was sent. It needs to know who agreed, when they agreed, and what happens when someone doesn’t.<\/p>\n\n\n\n

Because the absence of an objection is not the same as permission.<\/strong><\/p>\n\n\n\n

How AI processes a meeting<\/h2>\n\n\n\n

Before comparing training policies, I found it useful to separate three things that are often bundled together: processing, training and storage.<\/p>\n\n\n\n

Imagine asking a notetaker to summarise a client call.<\/p>\n\n\n\n

First, a speech-recognition system converts the audio into text. A language model may then receive that transcript, together with instructions such as \u201cidentify the decisions and action items\u201d. It generates a response, which the notetaker can save alongside your meeting.<\/p>\n\n\n\n

Processing<\/strong><\/h4>\n\n\n\n

When an already-trained model works on a new input, this is usually called inference.<\/p>\n\n\n\n

For a meeting summary, the transcript supplies information the model needs for that task. The model uses its existing learned patterns to generate an output. Ordinary inference does not itself update the model\u2019s learned parameters. IBM\u2019s explanation of AI inference<\/a><\/p>\n\n\n\n

That distinction matters: sending a transcript to an AI system does not automatically mean the system is being trained on it.<\/p>\n\n\n\n

Training<\/strong><\/h4>\n\n\n\n

Training changes a model\u2019s learned parameters using examples. It is a separate activity from generating a particular meeting summary.<\/p>\n\n\n\n

A provider might process your transcript to deliver the service and, depending on its policy and your settings, also use eligible data later to improve a model. You therefore need to check both activities. IBM\u2019s explanation of model training<\/a><\/p>\n\n\n\n

Storage<\/strong><\/h4>\n\n\n\n

A no-training commitment does not, by itself, mean a service keeps no copy.<\/p>\n\n\n\n

The notetaker may save the transcript so you can search it later. An external processor may have a separate retention period for operational, safety or legal purposes.<\/p>\n\n\n\n

Anthropic\u2019s commercial API documentation illustrates the distinction: it describes a standard deletion period for inputs and outputs, alongside exceptions and the possibility of a different contractual arrangement, including zero data retention. That is general API documentation, rather than proof of the arrangement any particular notetaker has negotiated. Anthropic\u2019s commercial data-retention explanation<\/a><\/p>\n\n\n\n

Is your meeting an AI training material?<\/h2>\n\n\n\n

This was one of the clearest differences in the policies I reviewed.<\/p>\n\n\n\n

A statement that external AI providers cannot train on customer data does not necessarily describe what the notetaker itself may do.<\/p>\n\n\n\n

Voicenotes<\/strong> says it does not use notes, recordings, transcripts or imported content to train its own models or those of its providers in any circumstances. Voicenotes information-secuirity policy<\/a><\/p>\n\n\n\n

Fireflies<\/strong> says customer meeting content is never used to train internal or external AI models. Fireflies information-security policy<\/a><\/p>\n\n\n\n

Granola<\/strong> prohibits third-party model training on every plan. Its own model-improvement policy is different: training using data it describes as anonymised <\/strong>is enabled by default on Basic and Business, with an opt-out. Enterprise has an organisation-wide opt-out enabled by default. Granola training policy by plan<\/a><\/p>\n\n\n\n

Otter<\/strong> says it uses de-identified recordings <\/strong>and transcripts to train its own models<\/strong>, while prohibiting training by its external AI service providers. Its Enterprise documentation provides an important qualification: Enterprise workspaces are opted out of model training by default. Otter privacy and security<\/a>, Otter Enterprise controls<\/a><\/p>\n\n\n\n

\u201cAnonymised\u201d and \u201cde-identified\u201d are relevant qualifications. They are still different commitments from excluding customer meeting content from training altogether.<\/p>\n\n\n\n

For confidential meetings, the important question is what opting out of AI training actually changes. Does it apply across the organisation? Does it cover previously collected data? And what happens to information that may already have been used for training?<\/p>\n\n\n\n

Otter<\/strong> and Granola<\/strong> do not provide users with a way to inspect training datasets or identify whether specific meeting content has previously contributed to model training. This makes it difficult to independently verify what an opt-out changes retrospectively.<\/p>\n\n\n\n

I would also avoid assuming that deleting a meeting reverses any earlier training use. That needs its own explicit answer.<\/p>\n\n\n\n

What remains after recording?<\/h2>\n\n\n\n

One of the more useful privacy choices is deciding whether original audio needs to be retained at all.<\/p>\n\n\n\n

Audio preserves voices, tone and background sound as well as the words spoken. Keeping it can help verify a disputed figure or commitment, but it also creates another record to manage.<\/p>\n\n\n\n

Granola<\/strong> says cached audio is deleted after transcription. Transcripts and notes remain, subject to the relevant controls. Crucially, its transcript auto-deletion can remove the underlying transcript while leaving enhanced notes and rough notes available. Granola data lifecycle<\/a><\/p>\n\n\n\n

Fireflies<\/strong> makes a distinction between recording methods. Its desktop \u201cTake Notes\u201d mode generates a transcript and summary without saving a separate replayable audio or video file. Its meeting bot can save recordings alongside the written record. Fireflies recording guide<\/a><\/p>\n\n\n\n

Voicenotes<\/strong> retains the original audio alongside the transcript and summary for as long as the note exists. This allows users to return to the recording when they need to check exactly what was said. Enterprise customers can also arrange automatic recording deletion after a specified period. Voicenotes Enterprise Plan<\/a><\/p>\n\n\n\n

Otter<\/strong> allows Enterprise administrators to remove audio across the entire workspace while preserving transcripts. Once enabled, the setting applies to both past and future meetings, and previously removed audio cannot be restored. Otter Audio Removal<\/a><\/p>\n\n\n\n

These are different choices about what survives a meeting.<\/p>\n\n\n\n

But removing audio does not remove the sensitivity of the written record. A summary naming an acquisition target or describing an employee\u2019s health can still contain information a business needs to protect carefully.<\/p>\n\n\n\n

I would assess the transcript and summary with the same attention as the recording.<\/p>\n\n\n\n

What does deletion mean?<\/h2>\n\n\n\n

This was the part of the research where the wording needed the most care.<\/p>\n\n\n\n

\u201cHow long do you retain data?\u201d sounds like a single question. In practice, several different clocks can be running.<\/p>\n\n\n\n

Normal retention<\/strong> is how long information remains available unless someone acts.<\/p>\n\n\n\n

Automatic retention<\/strong> is a rule that deletes specified content when it reaches a chosen age.<\/p>\n\n\n\n

Manual deletion <\/strong>is what happens when someone presses Delete.<\/p>\n\n\n\n

A recovery period<\/strong> is time spent in a trash folder before permanent deletion.<\/p>\n\n\n\n

Processor retention <\/strong>concerns copies handled by external services.<\/p>\n\n\n\n

Backup retention <\/strong>concerns information held in recovery systems, with any applicable exceptions.<\/p>\n\n\n\n

A 30-day automatic retention policy and a 30-day trash period describe different things. The first runs from the age of the content; the second starts when it is deleted.<\/p>\n\n\n\n

There is another distinction worth understanding: deleting a recording from any AI notetaker does not necessarily mean that information previously sent to an external AI processor disappears at the same time. Those providers have their own retention obligations and arrangements.<\/p>\n\n\n\n

Granola<\/strong><\/h4>\n\n\n\n

Granola says manually deleted notes move to trash and are purged from its servers after 30 days.<\/p>\n\n\n\n

Separately, individuals can configure transcript auto-deletion from one day to one year, while Enterprise customers can establish organisation-wide transcript policies. Remember that transcript deletion does not necessarily delete the notes derived from it. Granola deletion and retention guidance<\/a><\/p>\n\n\n\n

Fireflies<\/strong><\/h4>\n\n\n\n

Fireflies\u2019 security explanation says manual meeting deletion irreversibly removes the meeting from its database. The same article says it does not take snapshots of audio or transcripts, while metadata snapshots may remain for up to one year. Fireflies security explanation<\/a><\/p>\n\n\n\n

Its current privacy policy separately states that account-related personal information is deleted within 30 days of account closure. That is an account-deletion statement, rather than a universal deadline for every meeting-deletion event. Fireflies privacy policy<\/a><\/p>\n\n\n\n

Automatic meeting deletion is another control again. Fireflies documents it as an Enterprise feature, with periods from one week to one year. The meeting owner\u2019s setting governs, and the guide says the policy applies to future meetings. Fireflies automatic-deletion guide<\/a><\/p>\n\n\n\n

Otter<\/strong><\/h4>\n\n\n\n

Otter says ordinary user deletion immediately unshares a conversation and moves it to Trash for 30 days. Users can permanently delete it sooner by clearing it from Trash. Its help page says permanent deletion removes the conversation and associated data from its servers. Otter conversation-deletion guide<\/a><\/p>\n\n\n\n

Enterprise administrator deletion has a separate irreversible workflow. Enterprise retention policies can also cover a full conversation or apply separately to audio\/video and transcripts. Otter administrator deletion<\/a>, Otter retention controls<\/a><\/p>\n\n\n\n

Voicenotes<\/strong><\/h4>\n\n\n\n

Voicenotes retains the original audio for as long as the corresponding note exists. When a user deletes the note, the original recording is immediately removed from active storage. Copies already captured in rolling backups may remain for up to seven days before being removed.<\/p>\n\n\n\n

For enterprise customers, Voicenotes also offers automatic deletion of recordings after a specified period. This allows organisations to set retention periods rather than relying entirely on individual users to delete recordings manually.<\/p>\n\n\n\n

<\/p>\n\n\n\n

The copies outside the app!<\/h2>\n\n\n\n

There is one further complication: the meeting may already have travelled elsewhere.<\/p>\n\n\n\n

A transcript downloaded to a laptop, a summary emailed to a client or notes copied into a CRM need their own retention and deletion process.<\/p>\n\n\n\n

I would not treat deletion from the original notetaker as proof that every version has disappeared. Nor would I assume that a permanent-deletion help page supplies a complete backup schedule for every category of information.<\/p>\n\n\n\n

Where a business needs those assurances, it should ask for them explicitly.<\/p>\n\n\n\n

What SOC 2 tells us?<\/h2>\n\n\n\n

Once I had worked through training and retention, I returned to another familiar reassurance: SOC 2 Type II.<\/p>\n\n\n\n

Voicenotes, Granola, Fireflies and Otter all publish information about their SOC 2 Type II status. Their documentation provides a starting point for requesting and reviewing the relevant reports. Voicenotes Trust Centre<\/a>, Granola Trust Centre<\/a>, Fireflies security documentation<\/a>, Otter security documentation<\/a><\/p>\n\n\n\n

Although often described as \u201ccertification\u201d, SOC 2 is an independent examination and report. A Type II report assesses the relevant controls over a defined period.<\/p>\n\n\n\n

That is useful evidence. I would want to check what system the report covers, the audit period, any findings and the responsibilities placed on the customer. AICPA SOC resources<\/a><\/p>\n\n\n\n

It does not settle whether a particular meeting should be recorded, whether your sharing settings are appropriate or whether you accept a provider\u2019s training policy.<\/p>\n\n\n\n

Encryption deserves similar care. Encryption in transit and at rest does not necessarily mean that only you can read the content. Voicenotes explicitly explains that its AI features require providers to read information for processing. Voicenotes privacy policy<\/a><\/p>\n\n\n\n

I would treat security reports and encryption commitments as evidence to examine alongside the data-handling terms.<\/p>\n\n\n\n

How the tools compare<\/h2>\n\n\n\n

Putting the policies side by side makes the differences easier to see. This table summarises the published positions; the supporting documentation is linked in the discussion above.<\/p>\n\n\n\n

Question<\/th>Voicenotes<\/th>Granola<\/th>Fireflies<\/th>Otter<\/th><\/tr>
Provider\u2019s own AI training<\/td>Says meeting content is never used for internal or external training.<\/td>Anonymised-data model improvement on by default for Basic\/Business; opt-out available. Enterprise off by default.<\/td>Says meeting content is not used for internal or external training.<\/td>Describes de-identified-data training; Enterprise opted out by default.<\/td><\/tr>
Third-party AI training<\/td>Not allowed.<\/td>Not allowed.<\/td>Not allowed.<\/td>Not allowed.<\/td><\/tr>
Original audio<\/td>Kept for playback by default; Enterprise auto-deletion available.<\/td>Cached audio deleted after transcription.<\/td>Desktop app saves no separate recording file; bot capture can retain recordings.<\/td>Recordings retained unless deleted.<\/td><\/tr>
Automatic retention<\/td>Enterprise: custom retention for audio, transcripts, or both.<\/td>All plans: Users can auto-delete their own transcripts. Enterprise: enforced workspace-wide policy. Notes are kept.<\/td>Enterprise: auto-deletes full meetings after 1 week to 1 year.<\/td>Enterprise: custom retention for audio, transcripts, or both.<\/td><\/tr>
Manual deletion<\/td>Deleted immediately when the corresponding note is deleted. Copies in rolling backups are removed within seven days.<\/td>Notes move to trash; server purge after 30 days.<\/td>Meeting removed irreversibly from database, according to its security explanation. Metadata backups are separate.<\/td>Ordinary user deletion moves conversations to Trash for 30 days.<\/td><\/tr>
Consent Mechanism<\/td>Optional opt-out or opt-in consent: Teams can choose opt-in, where the bot joins only if every attendee agrees, or opt-out, where it joins unless someone declines.<\/td>Optional Chat notifications and video watermark; host responsible for partcipant consent.<\/td>Optional Opt-out mechanism: All meeting participants receive a pre-recording email with an option to decline.<\/td>Provides recording notifications and consent-related settings; host responsible for partcipant consent.<\/td><\/tr>
SOC 2 Type II<\/td>Report available<\/td>Report available<\/td>Report available.<\/td>Report available.<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n

Features and defaults depend on the plan, recording method and workspace configuration. This is a comparison of published documentation. Cells are simplified. Caveats such as backup periods and plan differences are covered in the sections above.<\/em><\/p>\n\n\n\n

What I would ask next?<\/h2>\n\n\n\n

Reading these policies did not leave me with a single winner for every business.<\/p>\n\n\n\n

It left me with a more useful set of questions.<\/p>\n\n\n\n

A team may value audio playback because it needs to check the exact wording of a commitment. Another may prefer not to retain original audio. Some organisations need central controls that individual accounts cannot provide.<\/p>\n\n\n\n

Before approving an AI notetaker for confidential work, I would work through these points:<\/p>\n\n\n\n

    \n
  • Define what can be recorded. Identify approved meeting types and discussions requiring separate permission or no recording.<\/li>\n\n\n\n
  • Check Enterprise Settings. Set the workspace settings and contractual protections your decision depends on.<\/li>\n\n\n\n
  • Map the processing. Confirm which platform receive content and where they store and process it.<\/li>\n\n\n\n
  • Check training policies. Examine the notetaker\u2019s own use and the permissions granted to external AI providers.<\/li>\n\n\n\n
  • Choose what to keep. Decide whether the business needs audio, full transcripts, summaries or only agreed minutes.<\/li>\n\n\n\n
  • Separate the retention periods. Document automatic expiry, manual deletion, trash, processor retention, backups and exceptions.<\/li>\n\n\n\n
  • Keep the evidence. Save the relevant agreement, policies and security report, and revisit material changes.<\/li>\n<\/ul>\n\n\n\n

    After the meeting ends<\/h2>\n\n\n\n

    Let’s return to that board meeting.<\/p>\n\n\n\n

    The acquisition discussion is over. Everyone has left. The action items are neatly organised, the summary is ready to share, and nobody had to spend the meeting taking notes.<\/p>\n\n\n\n

    Sounds like a win.<\/p>\n\n\n\n

    But choosing an AI notetaker should involve more than comparing transcription accuracy, summaries, and integrations.<\/p>\n\n\n\n

    Think about it. With so many advanced AI models available today, almost every notetaker can produce a pretty impressive transcript or summary. And as these models keep improving, the differences between them are only getting smaller.<\/p>\n\n\n\n

    So what really sets one apart?<\/p>\n\n\n\n

    Trust, security, privacy, and consent.<\/strong><\/p>\n\n\n\n

    And when your most confidential conversations are involved, we’d say that’s a pretty big deal.<\/p>\n\n\n\n

    At Voicenotes, we take these questions seriously. After all, you’re trusting us with conversations that aren’t meant for everyone. That’s why we believe you should understand how your information is handled and have meaningful control over it.<\/p>\n\n\n\n

    And if you’re recording a confidential client meeting, shouldn’t your clients feel just as comfortable as you do? That’s where consent matters. Giving people a say in whether they’re recorded is about showing them that you value their privacy too.<\/p>\n\n\n\n

    Because confidential conversations don’t stop being confidential when the meeting ends.<\/p>\n\n\n\n

    And before your next board meeting, client call or sensitive internal discussion, there’s one question worth asking:<\/p>\n\n\n\n

    Everyone in that room may have agreed to be recorded. But did they know where their conversation would go, who could access it, and how long it would remain there?<\/strong><\/p>\n\n\n\n

    If not, that’s probably worth finding out before you press record.<\/p>\n\n\n\n

    \u2014 Dona Sussan<\/p>\n\n\n\n

    <\/p>\n","protected":false},"excerpt":{"rendered":"

    What businesses should know about AI notetaker privacy and what the published policies of Voicenotes, Granola, Fireflies and Otter says. Halfway through a board meeting, someone mentions a possible acquisition. The finance director shares next quarter\u2019s projections. A colleague explains why a major client might leave. And the AI notetaker has been recording the whole […]<\/p>\n","protected":false},"author":8,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[25,20,27,29,21,22,26,24,19,23],"class_list":["post-948","post","type-post","status-publish","format-standard","hentry","category-uncategorized","tag-ai-meeting-notes","tag-ai-notetakers","tag-ai-training","tag-confidential-meetings","tag-data-retention","tag-fireflies","tag-gdpr","tag-granola","tag-meeting-privacy","tag-subprocessors"],"yoast_head":"\nAre AI Meeting Notetakers Safe for Confidential Business Meetings? - Insights, Tips & Tricks for Voicenotes<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/blog.voicenotes.com\/blog\/uncategorized\/are-ai-meeting-notetakers-safe-for-confidential-business-meetings\/\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Are AI Meeting Notetakers Safe for Confidential Business Meetings? - Insights, Tips & Tricks for Voicenotes\" \/>\n<meta property=\"og:description\" content=\"What businesses should know about AI notetaker privacy and what the published policies of Voicenotes, Granola, Fireflies and Otter says. Halfway through a board meeting, someone mentions a possible acquisition. The finance director shares next quarter\u2019s projections. A colleague explains why a major client might leave. And the AI notetaker has been recording the whole […]\" \/>\n<meta property=\"og:url\" content=\"https:\/\/blog.voicenotes.com\/blog\/uncategorized\/are-ai-meeting-notetakers-safe-for-confidential-business-meetings\/\" \/>\n<meta property=\"og:site_name\" content=\"Insights, Tips & Tricks for Voicenotes\" \/>\n<meta property=\"article:published_time\" content=\"2026-11-04T13:17:47+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/voicenotes.com\/blog\/wp-content\/uploads\/2026\/11\/Photo-by-Compare-Fibre-on-Unsplash-1024x767.jpeg\" \/>\n<meta name=\"author\" content=\"Dona Sussan\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:creator\" content=\"@voicenotes\" \/>\n<meta name=\"twitter:site\" content=\"@voicenotes\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"Dona Sussan\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"15 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\\\/\\\/blog.voicenotes.com\\\/blog\\\/uncategorized\\\/are-ai-meeting-notetakers-safe-for-confidential-business-meetings\\\/#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/blog.voicenotes.com\\\/blog\\\/uncategorized\\\/are-ai-meeting-notetakers-safe-for-confidential-business-meetings\\\/\"},\"author\":{\"name\":\"Dona Sussan\",\"@id\":\"https:\\\/\\\/blog.voicenotes.com\\\/blog\\\/#\\\/schema\\\/person\\\/5b74c968c5a25a7ef6a585ac49cd6100\"},\"headline\":\"Are AI Meeting Notetakers Safe for Confidential Business Meetings?\",\"datePublished\":\"2026-11-04T13:17:47+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/blog.voicenotes.com\\\/blog\\\/uncategorized\\\/are-ai-meeting-notetakers-safe-for-confidential-business-meetings\\\/\"},\"wordCount\":3322,\"commentCount\":0,\"publisher\":{\"@id\":\"https:\\\/\\\/blog.voicenotes.com\\\/blog\\\/#organization\"},\"image\":{\"@id\":\"https:\\\/\\\/blog.voicenotes.com\\\/blog\\\/uncategorized\\\/are-ai-meeting-notetakers-safe-for-confidential-business-meetings\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/voicenotes.com\\\/blog\\\/wp-content\\\/uploads\\\/2026\\\/11\\\/Photo-by-Compare-Fibre-on-Unsplash-1024x767.jpeg\",\"keywords\":[\"AI meeting notes\",\"AI notetakers\",\"ai training\",\"Confidential meetings\",\"Data retention\",\"Fireflies\",\"GDPR\",\"Granola\",\"Meeting privacy\",\"Subprocessors\"],\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"CommentAction\",\"name\":\"Comment\",\"target\":[\"https:\\\/\\\/blog.voicenotes.com\\\/blog\\\/uncategorized\\\/are-ai-meeting-notetakers-safe-for-confidential-business-meetings\\\/#respond\"]}]},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/blog.voicenotes.com\\\/blog\\\/uncategorized\\\/are-ai-meeting-notetakers-safe-for-confidential-business-meetings\\\/\",\"url\":\"https:\\\/\\\/blog.voicenotes.com\\\/blog\\\/uncategorized\\\/are-ai-meeting-notetakers-safe-for-confidential-business-meetings\\\/\",\"name\":\"Are AI Meeting Notetakers Safe for Confidential Business Meetings? - Insights, Tips & Tricks for Voicenotes\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/blog.voicenotes.com\\\/blog\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/blog.voicenotes.com\\\/blog\\\/uncategorized\\\/are-ai-meeting-notetakers-safe-for-confidential-business-meetings\\\/#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/blog.voicenotes.com\\\/blog\\\/uncategorized\\\/are-ai-meeting-notetakers-safe-for-confidential-business-meetings\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/voicenotes.com\\\/blog\\\/wp-content\\\/uploads\\\/2026\\\/11\\\/Photo-by-Compare-Fibre-on-Unsplash-1024x767.jpeg\",\"datePublished\":\"2026-11-04T13:17:47+00:00\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/blog.voicenotes.com\\\/blog\\\/uncategorized\\\/are-ai-meeting-notetakers-safe-for-confidential-business-meetings\\\/#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/blog.voicenotes.com\\\/blog\\\/uncategorized\\\/are-ai-meeting-notetakers-safe-for-confidential-business-meetings\\\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/blog.voicenotes.com\\\/blog\\\/uncategorized\\\/are-ai-meeting-notetakers-safe-for-confidential-business-meetings\\\/#primaryimage\",\"url\":\"https:\\\/\\\/voicenotes.com\\\/blog\\\/wp-content\\\/uploads\\\/2026\\\/11\\\/Photo-by-Compare-Fibre-on-Unsplash-1024x767.jpeg\",\"contentUrl\":\"https:\\\/\\\/voicenotes.com\\\/blog\\\/wp-content\\\/uploads\\\/2026\\\/11\\\/Photo-by-Compare-Fibre-on-Unsplash-1024x767.jpeg\"},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/blog.voicenotes.com\\\/blog\\\/uncategorized\\\/are-ai-meeting-notetakers-safe-for-confidential-business-meetings\\\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/blog.voicenotes.com\\\/blog\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Are AI Meeting Notetakers Safe for Confidential Business Meetings?\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/blog.voicenotes.com\\\/blog\\\/#website\",\"url\":\"https:\\\/\\\/blog.voicenotes.com\\\/blog\\\/\",\"name\":\"Insights, Tips & Tricks for Voicenotes\",\"description\":\"\",\"publisher\":{\"@id\":\"https:\\\/\\\/blog.voicenotes.com\\\/blog\\\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/blog.voicenotes.com\\\/blog\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Organization\",\"@id\":\"https:\\\/\\\/blog.voicenotes.com\\\/blog\\\/#organization\",\"name\":\"Insights, Tips & Tricks for Voicenotes\",\"url\":\"https:\\\/\\\/blog.voicenotes.com\\\/blog\\\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/blog.voicenotes.com\\\/blog\\\/#\\\/schema\\\/logo\\\/image\\\/\",\"url\":\"https:\\\/\\\/blogv2.voicenotes.com\\\/wp-content\\\/uploads\\\/2026\\\/07\\\/cropped-vnlogo.png\",\"contentUrl\":\"https:\\\/\\\/blogv2.voicenotes.com\\\/wp-content\\\/uploads\\\/2026\\\/07\\\/cropped-vnlogo.png\",\"width\":512,\"height\":512,\"caption\":\"Insights, Tips & Tricks for Voicenotes\"},\"image\":{\"@id\":\"https:\\\/\\\/blog.voicenotes.com\\\/blog\\\/#\\\/schema\\\/logo\\\/image\\\/\"},\"sameAs\":[\"https:\\\/\\\/x.com\\\/voicenotes\"]},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/blog.voicenotes.com\\\/blog\\\/#\\\/schema\\\/person\\\/5b74c968c5a25a7ef6a585ac49cd6100\",\"name\":\"Dona Sussan\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/4c578336290b99da86ea32ef0d11f6cbaaf2d6b8c617f7dd8734aba1a78ce3ab?s=96&d=mm&r=g\",\"url\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/4c578336290b99da86ea32ef0d11f6cbaaf2d6b8c617f7dd8734aba1a78ce3ab?s=96&d=mm&r=g\",\"contentUrl\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/4c578336290b99da86ea32ef0d11f6cbaaf2d6b8c617f7dd8734aba1a78ce3ab?s=96&d=mm&r=g\",\"caption\":\"Dona Sussan\"},\"url\":\"https:\\\/\\\/blog.voicenotes.com\\\/blog\\\/author\\\/dona\\\/\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"Are AI Meeting Notetakers Safe for Confidential Business Meetings? - Insights, Tips & Tricks for Voicenotes","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/blog.voicenotes.com\/blog\/uncategorized\/are-ai-meeting-notetakers-safe-for-confidential-business-meetings\/","og_locale":"en_US","og_type":"article","og_title":"Are AI Meeting Notetakers Safe for Confidential Business Meetings? - Insights, Tips & Tricks for Voicenotes","og_description":"What businesses should know about AI notetaker privacy and what the published policies of Voicenotes, Granola, Fireflies and Otter says. Halfway through a board meeting, someone mentions a possible acquisition. The finance director shares next quarter\u2019s projections. A colleague explains why a major client might leave. And the AI notetaker has been recording the whole […]","og_url":"https:\/\/blog.voicenotes.com\/blog\/uncategorized\/are-ai-meeting-notetakers-safe-for-confidential-business-meetings\/","og_site_name":"Insights, Tips & Tricks for Voicenotes","article_published_time":"2026-11-04T13:17:47+00:00","og_image":[{"url":"https:\/\/voicenotes.com\/blog\/wp-content\/uploads\/2026\/11\/Photo-by-Compare-Fibre-on-Unsplash-1024x767.jpeg","type":"","width":"","height":""}],"author":"Dona Sussan","twitter_card":"summary_large_image","twitter_creator":"@voicenotes","twitter_site":"@voicenotes","twitter_misc":{"Written by":"Dona Sussan","Est. reading time":"15 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/blog.voicenotes.com\/blog\/uncategorized\/are-ai-meeting-notetakers-safe-for-confidential-business-meetings\/#article","isPartOf":{"@id":"https:\/\/blog.voicenotes.com\/blog\/uncategorized\/are-ai-meeting-notetakers-safe-for-confidential-business-meetings\/"},"author":{"name":"Dona Sussan","@id":"https:\/\/blog.voicenotes.com\/blog\/#\/schema\/person\/5b74c968c5a25a7ef6a585ac49cd6100"},"headline":"Are AI Meeting Notetakers Safe for Confidential Business Meetings?","datePublished":"2026-11-04T13:17:47+00:00","mainEntityOfPage":{"@id":"https:\/\/blog.voicenotes.com\/blog\/uncategorized\/are-ai-meeting-notetakers-safe-for-confidential-business-meetings\/"},"wordCount":3322,"commentCount":0,"publisher":{"@id":"https:\/\/blog.voicenotes.com\/blog\/#organization"},"image":{"@id":"https:\/\/blog.voicenotes.com\/blog\/uncategorized\/are-ai-meeting-notetakers-safe-for-confidential-business-meetings\/#primaryimage"},"thumbnailUrl":"https:\/\/voicenotes.com\/blog\/wp-content\/uploads\/2026\/11\/Photo-by-Compare-Fibre-on-Unsplash-1024x767.jpeg","keywords":["AI meeting notes","AI notetakers","ai training","Confidential meetings","Data retention","Fireflies","GDPR","Granola","Meeting privacy","Subprocessors"],"inLanguage":"en-US","potentialAction":[{"@type":"CommentAction","name":"Comment","target":["https:\/\/blog.voicenotes.com\/blog\/uncategorized\/are-ai-meeting-notetakers-safe-for-confidential-business-meetings\/#respond"]}]},{"@type":"WebPage","@id":"https:\/\/blog.voicenotes.com\/blog\/uncategorized\/are-ai-meeting-notetakers-safe-for-confidential-business-meetings\/","url":"https:\/\/blog.voicenotes.com\/blog\/uncategorized\/are-ai-meeting-notetakers-safe-for-confidential-business-meetings\/","name":"Are AI Meeting Notetakers Safe for Confidential Business Meetings? - Insights, Tips & Tricks for Voicenotes","isPartOf":{"@id":"https:\/\/blog.voicenotes.com\/blog\/#website"},"primaryImageOfPage":{"@id":"https:\/\/blog.voicenotes.com\/blog\/uncategorized\/are-ai-meeting-notetakers-safe-for-confidential-business-meetings\/#primaryimage"},"image":{"@id":"https:\/\/blog.voicenotes.com\/blog\/uncategorized\/are-ai-meeting-notetakers-safe-for-confidential-business-meetings\/#primaryimage"},"thumbnailUrl":"https:\/\/voicenotes.com\/blog\/wp-content\/uploads\/2026\/11\/Photo-by-Compare-Fibre-on-Unsplash-1024x767.jpeg","datePublished":"2026-11-04T13:17:47+00:00","breadcrumb":{"@id":"https:\/\/blog.voicenotes.com\/blog\/uncategorized\/are-ai-meeting-notetakers-safe-for-confidential-business-meetings\/#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/blog.voicenotes.com\/blog\/uncategorized\/are-ai-meeting-notetakers-safe-for-confidential-business-meetings\/"]}]},{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/blog.voicenotes.com\/blog\/uncategorized\/are-ai-meeting-notetakers-safe-for-confidential-business-meetings\/#primaryimage","url":"https:\/\/voicenotes.com\/blog\/wp-content\/uploads\/2026\/11\/Photo-by-Compare-Fibre-on-Unsplash-1024x767.jpeg","contentUrl":"https:\/\/voicenotes.com\/blog\/wp-content\/uploads\/2026\/11\/Photo-by-Compare-Fibre-on-Unsplash-1024x767.jpeg"},{"@type":"BreadcrumbList","@id":"https:\/\/blog.voicenotes.com\/blog\/uncategorized\/are-ai-meeting-notetakers-safe-for-confidential-business-meetings\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/blog.voicenotes.com\/blog\/"},{"@type":"ListItem","position":2,"name":"Are AI Meeting Notetakers Safe for Confidential Business Meetings?"}]},{"@type":"WebSite","@id":"https:\/\/blog.voicenotes.com\/blog\/#website","url":"https:\/\/blog.voicenotes.com\/blog\/","name":"Insights, Tips & Tricks for Voicenotes","description":"","publisher":{"@id":"https:\/\/blog.voicenotes.com\/blog\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/blog.voicenotes.com\/blog\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Organization","@id":"https:\/\/blog.voicenotes.com\/blog\/#organization","name":"Insights, Tips & Tricks for Voicenotes","url":"https:\/\/blog.voicenotes.com\/blog\/","logo":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/blog.voicenotes.com\/blog\/#\/schema\/logo\/image\/","url":"https:\/\/blogv2.voicenotes.com\/wp-content\/uploads\/2026\/07\/cropped-vnlogo.png","contentUrl":"https:\/\/blogv2.voicenotes.com\/wp-content\/uploads\/2026\/07\/cropped-vnlogo.png","width":512,"height":512,"caption":"Insights, Tips & Tricks for Voicenotes"},"image":{"@id":"https:\/\/blog.voicenotes.com\/blog\/#\/schema\/logo\/image\/"},"sameAs":["https:\/\/x.com\/voicenotes"]},{"@type":"Person","@id":"https:\/\/blog.voicenotes.com\/blog\/#\/schema\/person\/5b74c968c5a25a7ef6a585ac49cd6100","name":"Dona Sussan","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/secure.gravatar.com\/avatar\/4c578336290b99da86ea32ef0d11f6cbaaf2d6b8c617f7dd8734aba1a78ce3ab?s=96&d=mm&r=g","url":"https:\/\/secure.gravatar.com\/avatar\/4c578336290b99da86ea32ef0d11f6cbaaf2d6b8c617f7dd8734aba1a78ce3ab?s=96&d=mm&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/4c578336290b99da86ea32ef0d11f6cbaaf2d6b8c617f7dd8734aba1a78ce3ab?s=96&d=mm&r=g","caption":"Dona Sussan"},"url":"https:\/\/blog.voicenotes.com\/blog\/author\/dona\/"}]}},"_links":{"self":[{"href":"https:\/\/blog.voicenotes.com\/blog\/wp-json\/wp\/v2\/posts\/948","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/blog.voicenotes.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/blog.voicenotes.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/blog.voicenotes.com\/blog\/wp-json\/wp\/v2\/users\/8"}],"replies":[{"embeddable":true,"href":"https:\/\/blog.voicenotes.com\/blog\/wp-json\/wp\/v2\/comments?post=948"}],"version-history":[{"count":23,"href":"https:\/\/blog.voicenotes.com\/blog\/wp-json\/wp\/v2\/posts\/948\/revisions"}],"predecessor-version":[{"id":972,"href":"https:\/\/blog.voicenotes.com\/blog\/wp-json\/wp\/v2\/posts\/948\/revisions\/972"}],"wp:attachment":[{"href":"https:\/\/blog.voicenotes.com\/blog\/wp-json\/wp\/v2\/media?parent=948"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/blog.voicenotes.com\/blog\/wp-json\/wp\/v2\/categories?post=948"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/blog.voicenotes.com\/blog\/wp-json\/wp\/v2\/tags?post=948"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}