Skip to main content
After voiceprint recognition is enabled, DingTalk can identify your speech in meetings based on voiceprint information. After the meeting ends, the speaker names in the transcript, To Dos, and minutes summary of the meeting recording (AI Minutes) will be accurately matched. This solves the question of “who said what” during meetings and helps organizations retain useful meeting data.

Scenario Highlights

1.1 Accurate Transcript Review

After voiceprint recognition is enabled, the AI Minutes transcript shows the speaker information for each sentence, including the speaker avatar and contact name. This makes it easier to find important opinions by speaker name when reviewing a meeting, improving information retrieval efficiency.

1.2 Smarter Key-Point Summaries

Smart minutes and chapters identify the source of each opinion by speaker, helping readers quickly understand who made each important conclusion. In workplace collaboration, this makes responsibility clearer and supports more accurate review and traceability.

1.3 Accurate To Do Matching

AI Minutes automatically links meeting To Dos to specific owners based on voiceprint information. It also supports one-click syncing to DingTalk To Do, giving post-meeting tasks clearer owners and improving follow-through efficiency.

Operation Flow

Entry Point for Enabling Voiceprint Recognition

Method 1: Manually Enable Voiceprint Recognition From Settings

On the AI Minutes home page, click in the upper-right corner, then choose Enable voiceprint recognition. 截屏2026-06-11 10.54.16.png

Agree to Enable

Because voiceprints are sensitive biometric information, DingTalk only applies voiceprint recognition after the individual user agrees to enable it. DingTalk protects voiceprint information strictly according to its privacy statement. Privacy agreement: https://terms.alicdn.com/legal-agreement/terms/privacy_policy_simple/20250710191027797/20250710191027797.html If a user wants to clear their voiceprint after agreeing, they can clear it at any time.

Collect a Voiceprint

After clicking Agree to enable, you can collect your voiceprint in either of the following ways.

Method 1: Read a Passage Aloud to Collect Immediately

In the Voiceprint collection pop-up window, click Start reading, then read the text at a normal pace in a quiet environment. After you finish reading the full passage, click Done to complete voiceprint collection. Desktop: 截屏2026-06-11 10.55.58.png Mobile: IMG_4668.png

Method 2: Automatically Collect From Meetings

If you do not have a voiceprint yet, DingTalk will automatically collect your voiceprint from AI Minutes generated by cloud recordings of meetings you previously attended or will attend later. After collection is complete, the voiceprint recognition switch will show My voiceprint exists below it. Click Clear to clear the existing voiceprint.

Voiceprint Recognition Results

Voiceprint recognition results will be shown in the next AI Minutes file after voiceprint collection is complete. Your speech will automatically be matched to your name and avatar. If the voiceprint recognition result is inaccurate, click the speaker avatar, then click Change speaker to modify the recognition result. Names can be changed in two ways:
  • Single change: only the corresponding entry is changed.
  • Global change: all instances of the same recognized speaker are changed.

Common Q&A

Q: In which scenarios is voiceprint recognition currently used? A: Offline meeting scenarios and AI Minutes started directly. Q: An employee has already collected a voiceprint. Why was their speech not recognized? A: Only participants in the current meeting can be matched by voiceprint recognition, including scheduled meeting participants or ad hoc meeting participants.