Dictate a note, or polish what you typed
Two buttons sit under the boxes you write in most: Dictate and Enhance. Dictate records what you say and puts the tidied words into the box a couple of seconds after you stop. Enhance rewrites what you have already typed. Neither one saves anything, and Undo puts the box back the way it was.
They appear under these boxes, on every screen the box appears on:
- +Time entry: the Summary box, and the Internal notes box.
- +Notes and activities on a company, on a deal, and on a website lead.
- +Emails you write yourself: an Outreach email, an email step in a sequence, the body you override when you launch one, and a paragraph block in a campaign.
- +LinkedIn steps in a sequence: the connection note and the message.
- +Handwritten card messages.
- +Business review text, in the review builder, the editor and the magazine layout.
Nothing else about the box changes. Your Save button stays where it is, character counters still count, and the keyboard shortcut that saves the form still saves it. On a phone the two buttons wrap onto their own line under the box.
Dictating
- 1Press DictateThe first time, the browser asks whether Current can use your microphone. Allow it once and it remembers for this site.
- 2TalkThe button becomes Stop, with a running timer and a small level meter so you can see your voice being picked up. Press Esc while it is recording to throw the recording away without sending it.
- 3Press StopThe line under the box reads Transcribing, then Polishing. A couple of seconds later the text lands in the box, already written for that box rather than as raw speech.
- 4Read it, then keep it or undo itThe line reads Added, with an Undo beside it. Undo restores the box exactly as it was. It stays for about twenty seconds, or until you start typing.
If the box was empty, the dictation becomes the whole box. If you had already written something, the new text is inserted where your cursor was, with the spacing sorted out, and your cursor ends up after it. A box with a character limit, such as a LinkedIn connection note, is filled up to that limit and the line under the box says the text was shortened to fit.
What Enhance does, and what it will not do
Enhance takes what is in the box and rewrites it for that box: an invoice summary reads like an invoice summary, a LinkedIn note reads like a LinkedIn note. It is the same rewrite that runs at the end of a dictation, which is why dictated text never lands as raw speech-to-text.
The rules it is held to:
- +It never invents a fact. If you did not write it or say it, it is not in the result.
- +Names, numbers, ticket ids, IP addresses, file paths, links, dates, and merge tags such as {{first_name}} come back exactly as they were.
- +Plain text comes back: no markdown, no headings, no bullet characters you did not write, and no quotation marks wrapped around the answer.
- +It writes in the language you wrote or spoke in.
- +An empty box, or something it cannot make sense of, comes back unchanged rather than invented.
- +What you wrote is treated as material to rewrite, never as instructions to obey. A line in your note that reads like a command to the AI is rewritten like any other line.
Time entries: one dictation, two boxes
A technician finishing on site says one thing, and a time entry needs two: a summary the partner reads on the invoice, and the technical detail your own team needs later. Dictate into Summary and you get both. The summary is written into Summary, and when Internal notes is still empty the technical part of what you said is dropped in there, with the line under the box saying so. One Undo puts both boxes back.
Enhance is narrower on purpose: pressing it on Summary rewrites Summary and never touches Internal notes. Internal notes has its own Dictate and Enhance, and that one keeps every command, path and error code exactly as you said it.
What happens to your voice
- +The audio goes to the speech-to-text provider on Current's own account, is transcribed, and is not kept there.
- +Current stores no recording and no transcript. The only thing that is saved is what you leave in the box and then save yourself.
- +Current records the size of each job for its own cost meter: seconds of audio and token counts. Your words are never part of that record.
- +A password or key you say out loud is written as [credential] rather than into the box.
- +There is no voice profile and nothing is learned from your recordings.
Everyone on a staff role sees them: Tenant Admin, Project Manager, Engineer, Account Executive and Sales Manager. Partner Viewers never do, in the portal or anywhere else, and the server refuses a request from one even if the buttons somehow reached the screen. Nobody sees them in a workspace where a Tenant Admin has switched AI writing tools off, which is covered in "Turn AI writing tools on or off, and teach the transcriber your names."
When something goes wrong
The message appears under the box in red, never as a pop-up, and the box is left exactly as it was.
| What it says | What to do |
|---|---|
| Microphone is blocked. Allow it in the browser's address bar, then try again. | Your browser is refusing the microphone for this site. The fix is per browser, below. |
| Didn't catch anything. Try again closer to the mic. | Nothing usable came back from the recording. Usually the headset is muted at its own switch, or the browser is listening to a different input. Check the microphone, then dictate again a little closer to it. |
| Recording is limited to 3 minutes. | The recording reached the cap and stopped. Dictate the rest in a second go. |
| AI writing tools are not set up yet. | The key that powers them has not been added on the platform side. Contact ITPartners+; nothing in your own Settings will fix this one. |
| A message about the AI budget | There is a limit per person per hour and a larger one per workspace per day, so one runaway page cannot spend the workspace's whole allowance. Try again after the hour turns, or ask a Tenant Admin. |
Microphone is blocked: the fix, per browser
- +Chrome: click the permission icon at the left of the address bar, set Microphone to Allow, and reload the page.
- +Edge: click the permission icon at the left of the address bar, open the permissions for this site, set Microphone to Allow, and reload.
- +Safari: open the Safari menu and choose Settings for This Website, then set Microphone to Allow. If it is still refused, open System Settings, Privacy & Security, Microphone, and check that Safari is switched on there too.
- +Firefox: click the crossed-out microphone icon in the address bar, clear the blocked setting, then reload and press Dictate again.
A blocked microphone is a browser setting, not a Current setting, so no admin here can unblock it for you. If the buttons work on your laptop but not on the machine in the workshop, that is the first place to look.
