AI Outbound Privacy Policy (iOS)
Last updated: November 14, 2026 Effective: November 14, 2026
This Policy applies to the AI Outbound app on iPhone, iPad and Mac (the "App"). We process your personal information only to the extent necessary to provide the App's features. The App does not collect the advertising identifier (IDFA), device serial number, contacts, call history or SMS content. It records a visit location only once, when you save an on-site recording, and never tracks your location in the background. The App does not include any advertising, analytics or third-party sign-in SDKs.
This English version is provided for convenience. If it differs from the Chinese version, the Chinese version prevails.
1. Who We Are
The App is developed and published by the individual developer Zhao Jiaqi ("we" or "us"). Business services such as accounts, outbound tasks, customer and call data are provided by the AI Outbound business backend, and the data is stored on the backend's servers (domains such as yunkecn.com and et-b.cn).
Contact email: zhaojiaqi@yunke.ai
2. What We Collect and Why
2.1 Sign-in and Account
- You sign in with your phone number and an SMS verification code, or with your phone number and a password. We collect your phone number, verification code or password to verify your identity.
- The App does not offer sign-up. Your account is created by your company administrator or the business partner, and information such as your name and company name is provided by your company.
- When you sign in, the App reports the device model, OS version and App version for compatibility and troubleshooting.
- After you sign in, the credential issued by the server is stored in the device Keychain and is used only to keep you signed in.
2.2 Outbound Tasks and Customer Management
- Customer numbers, names, companies, notes and other information you import or enter are uploaded to the business backend to create outbound tasks and manage customers.
- Recognizing numbers from photos: With your permission to use the camera, or after you pick an image with the system photo picker, the image is uploaded to the business backend to recognize the numbers in it. The system photo picker only gives the App the images you select; the App cannot access anything else in your photo library.
- File import: For spreadsheets or text files picked with the system file picker, the App reads only the files you select to extract numbers.
- Knowledge Base: Knowledge base files you upload are stored on the business backend so the AI can look up answers during calls.
2.3 AI Calls
- Consent to AI data sharing: AI calls, the AI assistant, AI Live Assist, AI summaries and Q&A for on-site recordings, and voice cloning for AI calls need to send data to the AI services listed in Section 3 for processing. These AI features are the core of the App, so the first time you open the App, before you sign in, the App shows a notice listing what data is sent and who receives it, and asks for your permission. The sign-in screen also has a matching checkbox where you can review the notice at any time. You can sign in only after you agree. If you do not agree, the App does not send this data and you cannot sign in to the App. Your choice is stored on the device. You can review the notice or withdraw your consent at any time in Me > AI Data Sharing; after you withdraw, the App stops sending this data and signs you out, and you need to agree again before signing back in.
- Call audio: When AI takes over a call, both parties' voices are sent to the App through the AI earphones or AI GO device, and streamed in real time to the large speech model service (see Section 3) for recognition and synthesis, so the AI can talk with the customer.
- Call recordings and transcripts: AI calls are recorded on the device and transcribed into text. Recordings and transcripts are uploaded to the business backend for call history, intent analysis and playback; recording files are stored in Alibaba Cloud Object Storage Service (OSS).
- Call history: The number dialed, customer name, dial / connect / hang-up times, call duration, whether AI took over, and similar details are saved on the device and uploaded to the business backend for statistics and billing.
- Call status: The App uses the system CoreTelephony interface to detect when calls on the device are placed, connected and ended, so it knows when the AI should start and stop. This interface does not provide the other party's number, and the App does not read the system call history.
- AI Live Assist: If your company has enabled this feature and you keep it on, during a manual call recorded through the earphones the transcribed text of what the other party says is sent to an AI agent on the business backend, which returns talking-point suggestions on the call screen. After the call, the other party's words and the AI suggestions are saved to the business backend.
- Voice cloning: The reading sample you record or upload, together with its text, is submitted through the business backend to the Volcano Engine Doubao speech service to create your cloned voice, which is used for speech synthesis in AI calls. Apple's system speech recognition is used to turn the sample into text. "Clone my voice" in the AI assistant is a separate feature that runs entirely on your device; see Section 2.8.
- Risky number check: If you turn on "Block Risky Numbers", the number is sent to the business backend before dialing to check whether it is blocklisted or risky.
- Hang-up SMS: If you turn on hang-up SMS, the business backend sends the message you configured to the customer's number after the call ends.
2.4 AI Earphones and AI GO Devices
- The App needs Bluetooth permission to connect AI work earphones or AI GO for call recording and AI voice injection. When connecting, it reads device information such as the device name, serial number, firmware version and battery level for device binding, authentication and status display.
- Some AI earphones must be registered and activated with the Volcano Engine device service on first use, which uploads the device's identifiers.
2.5 Feedback and Logs
- The App keeps runtime logs on the device for troubleshooting. Phone numbers in the logs are masked, and sign-in passwords are never logged.
- Only when you actively submit "Feedback" are your problem description and the device logs packaged and uploaded to the business backend.
2.6 Live Activities (Dynamic Island / Lock Screen)
While a call or outbound task is in progress, the App shows the call status, customer name (a masked number if there is no name) and the latest line of conversation in the Dynamic Island and on the Lock Screen. This content is shown only on your device and may be seen by people near you. You can turn it off in Settings > AI Outbound > Live Activities.
2.7 On-site Recording
- If your company has enabled on-site recording for you, you can use the phone's microphone to record in-person meetings with customers. Recording only starts when you start it yourself: by tapping Start in the App, or through Control Center / Lock Screen controls, Siri and Shortcuts, the Action button, or by long-pressing connected AI earphones (when not in a call). The recording status stays visible in the App, the Dynamic Island and the Lock Screen while recording. Recording continues when the screen is locked until you stop it or the 180-minute limit is reached. Please tell everyone present and obtain their consent before recording.
- After recording ends, the recording file is uploaded to the business backend together with the recording topic you entered, the linked customer (name, mobile number, company), sales stage and notes, and the start and end times, so that you and your company administrator can listen and follow up in Yunke.
- Visit location: When you save a recording, with your permission to access location "While Using the App", the App gets your current location once (latitude and longitude, and the address resolved from them) and uploads it with the recording. Addresses are resolved using Apple system services. You can decline the permission or edit the location manually; the recording can still be saved.
- If an upload fails, the recording and the information you entered are kept on the device and uploaded again automatically later. Recordings you choose to "Finish Later" are saved only on the device as drafts and are uploaded only after you complete and save them.
- Local history: Saved recordings, transcripts, AI summaries and Q&A are kept on the device as history, even after a successful upload, so you can view and replay them later. You can delete individual records at any time (this deletes them only from the device and does not affect recordings already uploaded to Yunke), or use "Clear Audio" on the records page to delete the audio of uploaded recordings. Everything is deleted when you delete your account.
- Live transcription (iOS 26 and later): While recording, the App uses Apple's built-in speech recognition to transcribe the conversation on the device. Recognition works offline, and no audio is uploaded for recognition. Transcripts are stored on the device and are not uploaded with the recording file; they are deleted when you delete your account.
- AI summaries and Q&A for recordings: When you generate a summary, ask for changes to it, or ask questions about a recording, the App sends the transcript together with the recording time, location, linked customer and sales stage (plus the request or question you entered) to the Qwen large model on Alibaba Cloud Model Studio. The audio itself is never sent. Summaries and Q&A are stored only on the device; only the notes you confirm are uploaded to Yunke. You can stop generation or edit the notes at any time.
- Highlights and on-site photos: During recording you can mark key moments (on the recording screen, in the Dynamic Island, or by long-pressing the AI earphones), and with your permission to use the camera you can take on-site photos such as quotes and business cards. Photos are stored only on the device and are not uploaded with the recording. The App uses Apple's built-in text recognition to extract text from photos on the device, offline. Marked times and the text recognized from photos are included in the transcript used for AI summaries and Q&A (see the previous item); the photos themselves are never sent.
- Background generation and notifications: After recording ends, the App keeps generating the AI summary even if you leave the screen or switch to the background. If the network is unavailable, the request is queued on the device and retried automatically. With your permission for notifications, the App sends local notifications when generation is complete or when you need to return to the App to start recording. Notifications are created on the device and do not go through a push server.
- Adding to-dos to Calendar / Reminders: You can add appointments and to-dos from a summary to the system Calendar or Reminders. For Calendar, a system screen asks you to confirm before saving, and the App does not read your calendar. Adding to Reminders requires your permission to access Reminders, and the App only adds the item you selected. These items are stored in your system Calendar / Reminders (and synced by Apple if iCloud sync is on).
- Sharing summary images: The long image is created on the device and is not uploaded to a server. It is passed to another app only when you choose to share it or save it to Photos. Saving to Photos requires your permission to "Add Photos Only".
2.8 AI Assistant
- The AI assistant starts only after you tap its floating button. You can hold to talk or type to have it look up customers, call records, outbound tasks and statistics, or prepare actions such as creating a task or changing scripts and settings. Actions that create or change something are first shown on a confirmation card and run only after you confirm.
- Voice input: While you hold to talk, the App uses the phone's microphone to capture your spoken request and Apple's system speech recognition to turn it into text (on the device where supported). The audio is not saved and is not sent to the AI services listed in Section 3; capture stops when you release.
- What is sent to the model: The text you speak or type, your name and company name, the screen you are on, and what the assistant looks up from the business backend or the device to fulfil your request (customer name, company, notes, follow-up status and scheduled follow-up times; call records, call summaries and conversation text; the title, time, location, linked customer, meeting summary, to-dos, Q&A and transcript excerpts of on-site recordings; outbound tasks and statistics; AI call minutes and service expiry status; script names and prompts; model settings and hang-up SMS templates; device connection status) are sent to Volcano Engine's Doubao model (see Section 3) to understand the request and generate a reply. The middle digits of customer phone numbers are hidden before sending; numbers you speak or type yourself are sent as entered.
- Daily briefing: Using fixed rules, the App puts together a briefing from today's and tomorrow's scheduled follow-ups, AI call minutes and service expiry status, outbound task progress, unsaved drafts, recordings that failed to upload and meeting-summary to-dos due today from on-site recordings, today's interested customers and call volume, and device connection status. It is shown on the assistant's first screen, and the floating button shows the number of unread items. The briefing is produced on the device without using a model, and nothing is sent to the AI services listed in Section 3 to produce it; it never opens or reads itself aloud automatically. Only when you ask the assistant to talk you through the briefing is its content sent to the model as described in the previous item. You can turn the briefing off or choose what it includes in "Me - AI Assistant".
- Timed reminders: After you turn them on in "Me - AI Assistant" and grant the notifications permission, the App schedules local notifications from the most recently fetched follow-up times, scheduled task time windows and service expiry date. Timed reminders are off by default. Notifications are created on the device and do not go through a push server; turning a switch off cancels the notifications already scheduled.
- On-site recordings: When you ask the assistant about your on-site recordings or what was discussed in a meeting, the assistant reads the recording title, time, location and linked customer stored on the device, together with the meeting summary, to-dos and your earlier Q&A, and sends them to the model as described under "What is sent to the model" above. Only when you ask for a detail that is not in the summary does it also read the parts of the transcript related to your question (the full transcript when it is short). Recording audio is never sent. When you ask the assistant to start an on-site recording, the assistant closes and opens the recording screen to start recording; the recording itself is handled in the same way as when you start it manually.
- Spoken replies: The assistant reads its replies aloud by default. You can choose an online or an on-device voice in Me > AI Assistant. With the online voice, the reply text is sent to Volcano Engine's Doubao speech synthesis service (see Section 3) to generate speech. With the on-device voice, speech is generated on your phone by a speech model downloaded to the device, and the reply text is not sent to the speech synthesis service; the model files are downloaded from Alibaba Cloud Object Storage Service, no personal information is uploaded during the download, and you can delete the model on the same screen at any time. If the on-device voice is not ready yet, or the phone cannot generate speech fast enough at that moment, the App switches to the online voice, and the reply text is then sent to the speech synthesis service as described above. In both cases the speech is played on your device and not saved. You can turn spoken replies off at the top right of the assistant screen; once off, reply text is no longer sent to the speech synthesis service.
- Clone my voice (on-device): When you use the on-device voice, you can read a passage of about 20 seconds to your phone so that the assistant reads its replies in a voice close to yours. Recording uses the phone's microphone. The recording is processed only on your device: the App computes a voice feature from it (a set of numbers describing the timbre of the voice) and then deletes the recording immediately. Neither the recording nor the voice feature is uploaded to the business backend or sent to the AI services listed in Section 3. A voice feature reflects the speaker's voice and is sensitive personal information; it is stored on the device under the signed-in account, is excluded from iCloud and computer backups, and is used only for the AI assistant to read replies aloud on the device, not for AI calls. You can delete it at any time in the voice list; it stays on the device after you sign out and is removed when you uninstall the App. Please record only your own voice, or a voice whose owner has given you permission.
- Live conversation: After you tap "Live Conversation" on the assistant screen, the App keeps the phone's microphone on and streams the captured audio to MOSI's MOSS Live real-time voice model (see Section 3), which recognizes what you say and answers by voice. Your name and company name, the screen you are on, and what the assistant looks up to fulfil your request (the same scope as the previous item) are also sent to it. The audio and the answers are not saved on the device. You can turn the microphone off or end the live conversation at any time; when you close the assistant, the App moves to the background, a call comes in or a recording starts, the App stops capturing and disconnects. When you do not use live conversation, the App sends no data to MOSI. The App does not enable MOSS Live's cross-session memory; the conversation context is used only within the current live conversation.
- The conversation is kept only for the current session. It is not saved to local history or uploaded to the business backend, and is cleared when you start a new conversation, after a long period of inactivity, or when you sign out or switch accounts. To support multi-turn conversations, Volcano Engine temporarily keeps the context of the current conversation under its service rules.
3. Third-Party Services
To provide the features above, the App accesses the following third-party services directly. The App does not include these providers' SDKs; it only sends the data needed for each feature through their network APIs.
| Service | Provider | Purpose | Information involved |
| Doubao real-time speech model and speech synthesis | Beijing Volcano Engine Technology Co., Ltd. | Speech recognition, conversation and speech synthesis in AI calls; voice cloning | The voices of both parties on the call; scripts and knowledge base content, plus the customer name, company, notes, phone number and last-call digest referenced in the script; for voice cloning, the voice sample you record and its text |
| Doubao model and Doubao speech synthesis (AI assistant) | Beijing Volcano Engine Technology Co., Ltd. | Understanding your requests to the AI assistant, looking up information and generating replies; reading replies aloud | The text of what you say to the assistant or type (audio is not sent); your name and company name; what the assistant looks up to fulfil the request: customer name, company, notes, follow-up status, scheduled follow-up times and phone numbers with the middle digits hidden; call records, call summaries and conversation text; the title, time, location, linked customer, meeting summary, to-dos, Q&A and transcript excerpts of on-site recordings (recording audio is not sent); outbound tasks and statistics; AI call minutes and service expiry status; script names and prompts; model settings and hang-up SMS templates; device connection status |
| MOSS Live real-time voice model (AI assistant live conversation) | Shanghai MOSI Intelligence Technology Co., Ltd. (上海模思智能科技有限公司) | Speech recognition, conversation and speech synthesis in the AI assistant's live conversation | Sent only after you start a live conversation: audio captured by the phone's microphone during the live conversation; your name and company name; what the assistant looks up to fulfil the request, the same scope as the row above |
| Volcano Engine device service | Beijing Volcano Engine Technology Co., Ltd. | Registering and activating some AI earphones | Device identifiers |
| Object Storage Service (OSS) | Alibaba Cloud Computing Co., Ltd. | Uploading and downloading call recordings; downloading the on-device speech model files | Recording files (no personal information is sent when downloading the speech model) |
| Qwen large model on Alibaba Cloud Model Studio | Alibaba Cloud Computing Co., Ltd. | Speech recognition, conversation and speech synthesis in AI calls (when the corresponding work model is selected); AI summaries and Q&A for on-site recordings | AI calls: the voices of both parties, scripts and the customer details they reference. On-site recordings: transcripts, text recognized from photos, recording time, location, linked customer, sales stage, and the requests or questions you enter (the audio itself is not sent) |
| Yunke AI services (real-time voice, AI Live Assist, intent analysis) | Beijing Lianyun Tianxia Technology Co., Ltd. | Speech recognition, conversation and speech synthesis in AI calls (when the corresponding work model is selected); AI Live Assist; call intent analysis and summaries | AI calls: the voices of both parties, scripts and the customer details they reference. AI Live Assist: transcribed text of what the other party says during a call. Intent analysis: call recordings and conversation text |
The data of each AI call is sent only to the one service behind the work model you choose. Before sending data to the AI services above, the App explains this in the App and asks for your permission (see 2.3). We provide only the data needed for each feature, and require these providers to protect your information to a standard no lower than this Policy and not to use it for any purpose other than providing the features above.
Volcano Engine privacy policy: https://www.volcengine.com/docs/6256/64902
MOSI privacy policy: https://mosi.cn/page/privacy-policy
Alibaba Cloud privacy policy: https://terms.alicdn.com/legal-agreement/terms/privacy_policy_full/20240202100310511/20240202100310511.html
4. System Permissions
| Permission | Purpose | If not granted |
| Bluetooth | Connect AI earphones and AI GO | AI calls are unavailable |
| Camera | Take photos of numbers to recognize them; take on-site photos during on-site recording | You cannot import by photo or take on-site photos; other features are unaffected |
| Microphone | Record in-person meetings during on-site recording; record your voice when cloning a voice for AI calls or for the AI assistant; capture your speech when you hold to talk or have a live conversation with the AI assistant | On-site recording, recording a voice sample and the AI assistant's voice features are unavailable (you can still type); other features are unaffected |
| Location (While Using the App) | Record the visit location when saving an on-site recording | You enter the location manually; the recording can still be saved |
| Photos (Add Only) | Save on-site recording summary images to Photos | You cannot save directly to Photos, but can still share the image |
| Notifications | Notify you when an AI summary finishes in the background, or when you need to return to the App to start recording; after you turn on the AI assistant's timed reminders, remind you at scheduled follow-up times, scheduled task time windows and before the service expires | No reminders; summaries are still generated and the briefing can still be viewed in the assistant |
| Reminders | Add to-dos from summaries to Reminders | You cannot add to Reminders, but can still add to Calendar |
Picking images from your photo library, picking files and adding to Calendar are all confirmed by you through system screens and require no permission. Each permission is requested only when you use the related feature, and you can turn it off at any time in Settings > AI Outbound.
5. Sharing, Transfer and Public Disclosure
- Apart from the services listed in Section 3 that are necessary for the features, we do not sell, rent or share your personal information with third parties.
- We do not publicly disclose your personal information, except with your separate consent, or as compelled by laws and regulations, litigation or government authorities.
- Customer information you enter in the App is business data belonging to you or your company. Do not provide customer information to others without the customer's consent.
6. Storage and Security
- Your personal information is stored within the People's Republic of China and is not transferred abroad.
- Local data (call history, recordings, transcripts, customer cache, logs) is stored on the device separately for each account. Signing out does not delete it, and you can view it again after signing back in to the same account. It is deleted when you delete the records or uninstall the App. Data on the business backend is kept while your account exists, and is handled as described in Section 7 after account deletion.
- Network traffic is encrypted with HTTPS and similar methods, and the sign-in credential is stored in the system Keychain. Although we take reasonable security measures, the internet is never completely secure, so please keep your account password safe.
- If a personal information security incident occurs, we will inform you by in-app notice, SMS or email, and report it to the authorities as required.
7. Your Rights
- Access and correction: You can view your account information, call history and customer information in the App and edit them on the relevant screens. You can change your sign-in phone number and password in Me > Account & Security (your company may disable phone number changes, in which case contact your company administrator).
- Deletion: You can delete local call history and recordings in the App. To delete data on the business backend, contact us by email.
- Withdrawing permissions: You can turn off Bluetooth, Camera, Microphone, Location, Notifications, Reminders, Live Activities and other permissions in system Settings. The related features will stop working, but processing carried out before withdrawal is not affected.
- Account deletion: Employee accounts are handled by the company administrator in the backend. You can also email zhaojiaqi@yunke.ai with your sign-in phone number to request deletion. We will process the request within 15 working days after verifying your identity. After deletion, your personal information will be deleted or anonymized, except as otherwise required by law.
8. Minors
The App is a work tool for businesses and sales staff and is not intended for minors under 14. If you are a minor, please use it with the supervision and consent of your guardian.
9. Updates to This Policy
When this Policy is updated, we will notify you prominently in the App, for example with a pop-up. For material changes (such as changes to the purposes or methods of processing, or to third parties), we will ask for your consent again.
10. Contact Us
If you have any questions, comments or complaints about this Policy or how your personal information is handled, email zhaojiaqi@yunke.ai. We usually reply within 15 working days.