DATA PRIVACY
Privacy policy
How Talenda protects the personal data of candidates, employees, students and our customers, including when AI works on their behalf.
01About this policy
Talenda provides recruitment, HR (HRIS) and interview-practice software to employers, recruitment firms and schools (“organisations”). This policy explains how we handle personal data under the Personal Data (Privacy) Ordinance (Cap. 486) (“PDPO”) and its six Data Protection Principles (“DPPs”). It also explains how we follow the Privacy Commissioner’s guidance on protecting personal data in the use of agentic AI.
We act in two roles:
- For organisations’ data. The organisation that uses Talenda decides why the personal data of its candidates, employees and students is collected. It is the data user. We process that data on its behalf, as its data processor, only on its instructions and under contract.
- For our own data. For people who hold a Talenda account, visit this website or contact us, we are the data user.
If you are a candidate, employee or student, contact the organisation that invited you first. Its own Personal Information Collection Statement also applies to you. We will help it answer your request.
02Personal data we hold
DPP 5
- Account data: name, work email, role, organisation, language preference and sign-in records.
- Candidate data: CVs and their contents, applications, job preferences, interview schedules, recruiter notes and match results.
- Interview data: audio and transcripts of AI voice interviews, AI summaries and scores. Video is recorded only when the organisation turns it on and you agree.
- Pre-employment and HR data: e-contract signatures, documents you upload for onboarding, and HR records such as employment details, leave, attendance, compensation, MPF and payroll information. Some of this is sensitive, such as identity documents and bank account details. It is collected only when the organisation needs it to employ or pay you.
- School data: student roster details, practice interview recordings, transcripts and feedback.
- AI agent notes: short notes an AI recruitment agent keeps about preferences a candidate has shared, and its message history.
- Technical data: security and audit logs, AI usage records (feature, model, time and volume) and the cookies described below.
03How we collect personal data
DPP 1
We collect personal data only for a lawful purpose directly related to the service, and only what is necessary and adequate for that purpose. We do not collect data because it might be useful later.
We collect it from you directly (for example a CV upload, a form or an interview), from the organisation (for example an import or HR record), and from public pages a user asks us to read. Talenda’s AI features are not designed to need identity document numbers, bank account numbers or passwords, and we ask organisations not to supply them for AI tasks.
When we collect data from you, we explain the purpose, who it may be shared with, whether supplying it is voluntary and your right to access and correct it. For example, you agree to an AI interview before it starts, and pre-employment forms record the consent you give and the version you saw.
04How we use personal data
DPP 3
We use personal data to run recruitment, onboarding, HR administration and interview practice for organisations. We also use it to provide accounts and support, to keep the service secure and to meet legal obligations.
- We do not use personal data for a new purpose without your prescribed consent.
- We do not use personal data for direct marketing without your consent, and we do not sell it.
- We do not use organisations’ data to train AI models, and one organisation’s data is never used for another.
05How we use AI, including agentic AI
DPP 1(3), DPP 3
AI helps organisations do the groundwork. People make the decisions. This table shows each AI feature, the personal data it uses and how people stay in control.
| Feature | What it does | Personal data used | Human control |
|---|---|---|---|
| CV reading and profiles | Reads an uploaded CV and fills in a structured candidate profile. | CV contents: contact details, work history, education, skills. | Recruiters review and can edit every field. Reprocessing never overwrites staff checklists. |
| Candidate matching | Suggests candidates for an opening and explains the reasons. | Profile, preferences and the job description. | Suggestions only. People decide who to shortlist, interview or reject. |
| AI voice interviews and practice | Runs a structured screening or mock interview, then transcribes, summarises and scores it. | Voice audio, transcript, answers, and video only if the organisation enables it. | You see a notice and give consent before the interview starts, and recording is shown on screen. Staff review the transcript. A score informs a decision but never makes it. |
| Ask Anything assistant | Answers staff questions about their own workspace and prepares changes they ask for. | Records the staff member is already allowed to see. | Every create, edit or delete is held as a draft on our server and runs only after the staff member presses Confirm. Drafts expire after 30 minutes. A typed "yes" or the model's own output can never approve an action. |
| AI recruitment agents (agentic AI) | When an organisation switches them on, agents follow up with candidates by email or WhatsApp, suggest suitable openings, send interview preparation and ask for feedback. | Profile, applications, message history and short notes on preferences the candidate has shared, such as availability or preferred contact channel. | Agents act only within routines the organisation enables, with daily limits. Staff can pause the agent for any candidate. Agents do not make hiring decisions, and deleting an agent's saved notes needs human confirmation. |
| Web reading | Reads a public web page, such as a job advertisement, when a user asks. | The public page only. | Pages are opened in an isolated browser service that has no access to workspace data or credentials. |
Our agentic AI safeguards
- Human in the loop. Decisions with a significant effect on you, such as hiring, rejection, offers and HR changes, are made by people. High-risk, irreversible or unusual actions need an explicit human confirmation. These include deleting records, sending data outside the workspace and changing settings.
- Minimum access. Every AI action runs with the permissions of the signed-in user and is confined to that organisation’s workspace. AI never gets administrator or system rights. Provider keys stay on our servers and are never sent to the browser.
- Accuracy. AI answers are grounded in the workspace’s own records, matches show their reasons, and staff can review and correct AI output.
- Traceability. We log each AI call (feature, model, time, organisation and user), agent tasks and messages, and changes to HR records, so actions can be traced and audited.
- Vetted tools only. Users cannot install third-party plugins or skills into Talenda’s agents. We build and review every agent tool ourselves and use official, pinned versions of software and models.
- Ongoing risk assessment. We test each AI feature before release and keep reviewing its privacy risks in use, taking account of the data involved, whether its actions can be reversed and the possible impact on people.
06Accuracy
DPP 2(1)
We take practicable steps to keep personal data accurate. AI can make mistakes, so we reduce the risk by grounding AI in the workspace’s records, showing reasons and sources, and keeping people in review. If you think data about you is wrong, ask for it to be corrected (see Your rights).
07How long we keep personal data
DPP 2(2)
We keep personal data no longer than necessary for its purpose. Organisations set most retention periods for their own data. The service also applies these limits:
- Interview audio and video recordings are deleted automatically after 30 days. Transcripts and summaries stay with the candidate record until the organisation deletes it.
- Voice notes spoken to the Ask Anything assistant are transcribed and not stored.
- Unconfirmed assistant drafts expire after 30 minutes.
- Organisations can set pre-employment documents to be deleted automatically three months after the employment start date.
- Deleting a candidate removes their interview recordings. Agent notes can be reviewed and deleted by staff at any time.
- Account data is kept while the account is active. When an organisation’s contract ends, we return or delete its data within a reasonable period, unless the law requires us to keep it.
09Security
DPP 4
- All traffic is encrypted in transit. Access is controlled by role, and every query is limited to the user’s own organisation.
- Voice interviews use single-use signed tickets, and browsers never connect to an AI provider directly.
- Document conversion and web reading run in isolated, restricted services. Links sent for e-contracts and pre-employment forms are signed and expire.
- We apply guardrails before, during and after AI processing to limit leaks of personal data, keep audit logs, back up data and grant staff the minimum access they need.
- If a data breach occurs, we act to contain it, inform affected organisations promptly and help them notify you and the Privacy Commissioner where appropriate.
10Your rights
DPP 6
You can ask for a copy of your personal data and ask for it to be corrected. Send your request to the organisation that holds your data, or to us at the address below. We will reply within 40 days. We may charge a fee for a copy, which will not be excessive. If we cannot meet a request, we will tell you why, as the PDPO allows.
You can also ask the organisation to stop AI agent messages to you. Its staff can pause the agent for you at any time.
If you are not satisfied with our response, you can complain to the Office of the Privacy Commissioner for Personal Data.
12Governance and training
A named person at Talenda is responsible for personal data privacy and has the authority to stop or change an AI feature. Staff who handle personal data or build AI features are trained on the risks of agentic AI. We review this policy and our AI safeguards at least once a year, and whenever we add a significant AI feature.
13Changes and contact
We will post any changes to this policy on this page and update the effective date. If a change materially affects how we use your data, we will tell account holders before it takes effect.
For questions or requests about personal data, contact our privacy team at privacy@talenda.ai.