A privacy checklist for dictating client work on Mac

Client work tends to collect sensitive details in ordinary places.

The rough note after a call, the draft reply you have not polished yet, the context behind a proposal, the internal reason for a decision, the name of the person who is frustrated, the budget number you should not casually paste into another tool. None of those details need to be dramatic to deserve care.

That is why dictation can feel awkward for client work. Speaking is faster than typing, but voice capture can introduce a new question: where does the audio go before it becomes text?

You do not need a heavy compliance process for every note. You do need a small habit for deciding what is safe to dictate, where the transcript lands, and what should happen to the recording afterward.

1. Separate capture from sharing

The most useful privacy habit is also the simplest: treat dictation as a capture step, not a sharing step.

When you speak a client note, you are often saying more than the final message needs. You might include context, uncertainty, names, alternatives, or the reason you are leaning one way. That rough version is useful for thinking, but it may not belong in an email, ticket, AI prompt, or shared document.

Use voice to get the full thought down. Then edit before the text leaves your Mac.

That gives you a cleaner workflow:

1. Dictate the messy version.
2. Review the transcript locally.
3. Remove names, numbers, or internal reasoning that do not need to travel.
4. Paste or send only the finished version.

The extra minute is worth it. Privacy mistakes often happen when rough context moves too quickly from a private workspace into a shared one.

2. Prefer local transcription for raw client audio

Raw voice contains more than the words you meant to preserve. It can include background comments, side notes, emotional tone, and details you would naturally remove from a written summary.

For client work, a local transcription workflow reduces one unnecessary handoff. The audio can be processed on your Mac before you decide what to do with the text.

SpeakLane is built around that pattern: local dictation, local file transcription, and local history. The app does not host your recordings or transcripts on SpeakLane servers, and the privacy policy explains the local-first boundary.

That does not make every downstream use private. If you paste the finished text into a cloud note app, CRM, email tool, or AI service, that app has its own privacy terms and retention behavior. Local dictation helps with the capture step. You still own the decision about where the final text goes.

3. Decide where the transcript should land

Auto-insert is convenient, but convenience should match the sensitivity of the material.

If you are dictating a normal reply into a trusted text field, auto-insert can be the fastest path. Click into Mail, Notes, Slack, Notion, Cursor, or the document you are already using, hold the hotkey, speak, release, and let the text appear.

If the note includes raw client context, use a review space first. Dictate into a local note, draft document, or scratchpad where you can clean it up before moving it anywhere else.

The practical split looks like this:

• Use auto-insert for low-risk drafts you are comfortable reviewing in place.
• Use clipboard or a scratch document for client-sensitive notes that need cleanup first.
• Use file transcription for recordings that already exist, such as voice memos, call clips, or research interviews.

You can adjust auto-copy and auto-insert behavior in Settings. The goal is not to make every workflow slower. It is to make the risky ones deliberate.

4. Keep the recording only when it has value

Saving every recording feels safe until it becomes a private archive you never meant to maintain.

Some client recordings are worth keeping because the original audio has value: an interview clip, a quoted comment, a research session, or a record you need to revisit. Many quick dictations are different. Once the transcript has been cleaned and moved into the right place, the audio may no longer be useful.

Check your retention settings before dictation becomes a daily habit.

In SpeakLane, local history stores each session with the transcript and audio when audio is available. The history folder is useful as a safety net, but it should not become a permanent dumping ground by accident. Use the retention setting that fits the way you work, and periodically clear anything you do not need.

A good default is:

• Keep transcripts long enough to recover from mistakes.
• Keep source audio only when the audio itself matters.
• Delete throwaway captures after the final note has been moved.
• Store long-term client records in the system you already use for client work.

That last point matters. A dictation history folder is not a replacement for your project archive, document management process, or client file.

5. Be careful with names, numbers, and exact quotes

Dictation is excellent for getting the shape of a thought onto the page. It is not a reason to stop checking important details.

Before you reuse a client transcript, scan for the things that are expensive to get wrong:

• Client names and company names.
• Dates, prices, deadlines, and contract terms.
• Product names, account names, and technical terms.
• Quotes you plan to attribute to someone.
• Anything that sounds like advice, approval, or a commitment.

If the wording matters, listen back or verify against the source. Local transcription reduces upload exposure; it does not turn rough audio into a legal record or an approved client summary.

This is especially important when the dictated text will become a proposal, support reply, invoice note, product decision, or customer-facing message.

6. Do not overstate what a privacy tool does

A private dictation app is one part of a larger workflow.

It can help you avoid uploading raw audio just to get text. It can keep transcripts and recordings on your Mac. It can let you inspect the rough version before sharing anything.

It cannot decide whether a detail belongs in a client email. It cannot control what happens after you paste text into another service. It does not make regulated work compliant by itself. If your client work has formal legal, medical, financial, or contractual requirements, follow those requirements first.

That boundary is useful. It keeps the promise realistic: reduce unnecessary exposure during capture, then make a better decision about what leaves your machine.

A small routine for client dictation

Use this routine for the next client-sensitive note:

1. Open a local scratch document.
2. Dictate the full thought with a clear beginning and end.
3. Fix names, dates, numbers, and product terms.
4. Remove context that helped you think but does not need to be shared.
5. Move the cleaned text into the final app.
6. Keep or delete the recording based on whether the source audio still has value.

That is enough for most everyday client work.

The point is not to make voice feel risky. The point is to make it usable without being careless. When the raw capture stays local and the sharing step is intentional, dictation becomes a practical way to draft client notes, replies, and plans without creating extra data handoffs.