Effective 6 October 2026

Privacy notice

This notice explains how we handle information in the Tempworks platform.

Who we are and how to contact us

Tempworks AI Ltd provides the Tempworks platform. Contact Antony at antony@tempworks.ai about privacy or your information.

Your organisation decides which business records to use in the platform and who may access them. We process those records on its instructions. Your organisation is responsible for explaining its purposes and lawful basis to the people whose information it supplies. We are responsible for the information we use for our own platform security and for meeting our own legal obligations. You can contact us about either type of processing; we will involve your organisation where it is responsible for the request.

Information we use and why

We use account details and organisation memberships to authenticate you and control access. We process job/project records, messages, relevant conversation context, uploaded files and filenames to deliver the features you use. We keep limited operational information to maintain the service, and business audit records to identify who performed recorded actions.

Information comes from you, your organisation and the business records connected to your workspace. It can include names, business contact details and the identities of engineers carrying out checks. The service is intended for job and business work, rather than collecting health, criminal or other sensitive personal records. Your organisation chooses its uploads and must ensure they are appropriate and covered by its instructions; uploading something does not remove the privacy protections applying to it.

We use customer chats and files only to provide and support the service. We do not sell them, use them for advertising or use them to train our own AI models.

For our own account-security, abuse-prevention and service-diagnostic processing, we rely on legitimate interests: providing secure access, preventing misuse and investigating failures, with restricted access and limited retention. For processing necessary to meet our own data-protection obligations, we rely on the applicable legal obligation. Customer-directed business processing follows your organisation's instructions and lawful basis. These are distinct purposes; we do not treat an employer's subscription as your individual consent.

Account and membership details are needed to provide authenticated access under the service arrangement. Without them, we cannot provide access to your organisation's workspace. Chats and file uploads are optional; not supplying them means the corresponding chat or file action cannot be carried out. Your organisation can explain any separate obligation to provide its job records.

Access, AI and engineering checks

Access to shared jobs and saved files follows your organisation's permissions. Chats are private to their user within the active organisation; organisation administrators do not receive a special permission to read them. Authorised infrastructure operators and providers may process information for service operation and support under access controls and processing arrangements.

Messages, relevant conversation context and tool-retrieved records are sent to Amazon Bedrock. The configured model is gpt-oss-120b, hosted and invoked through Amazon Bedrock in AWS London. Currently, uploaded binary file contents are not sent to the model, although filenames may be. AI responses can be wrong. A designated engineer must review outputs before they are used for engineering or safety decisions. Each recorded engineering check must identify its engineer, job/revision, decision and time in the business audit trail, separately from the chat. This is a human review requirement, not a claim that AI certifies engineering work.

Providers and locations

AWS provides the application's hosting, identity service, database, file storage, AI processing and operational monitoring. The application database, file storage and configured model processing are in London. This does not mean every provider support or service operation is restricted to the UK.

AWS's service terms incorporate its Data Processing Addendum and applicable UK GDPR terms. Those terms address subprocessors and international transfers, including contractual safeguards where applicable. Contact us for information about relevant safeguards and providers. We also disclose information where required by applicable law. We do not represent London configuration alone as proof that no international processing occurs.

Retention and deletion

Chats
30 days from creation; continuing a chat does not extend this period. You can delete a chat earlier.
Temporary uploads
A separate 30 days from upload registration. Deleting a chat does not itself delete its uploads.
Saved job/project files and engineering/business audit
Separate business records, retained under the organisation's approved schedule. Explicit job/project closure starts the applicable file/audit period; holds can delay deletion. Saving a file preserves it after chat and temporary-upload expiry. Company activity has its separately agreed schedule.
Identity and memberships
While needed for active access or unresolved account obligations. Removing access to one organisation does not necessarily remove an account used in another organisation.
Privacy-request and approval records
While needed to resolve the request, verify the action, handle a specific complaint/hold or explain a retention decision that still applies. Retain only evidence needed for the documented purpose and dispose of redundant material when that need ends.
Deletion/recovery evidence
Minimum identifiers and decisions needed to prevent deleted information returning from backups or older clients, until relevant recoverable copies and recreation risks have been resolved. This does not justify retaining the deleted chat or file contents.
Chat-service diagnostic logs
30 days. Other incident or support evidence follows its documented purpose and retention need.
Automated database backups
A further seven-day recovery window. Separate retained recovery copies require their own expiry/hold decision.

Existing organisations without an approved permanent-record schedule are not assigned an invented deletion deadline. Contact your organisation or us for the schedule applicable to your records. Automatic permanent-record cleanup is not yet enabled; the approval and deletion-verification requirements still apply.

Deleted or expired chat content becomes unavailable through the service. Physical expiry cleanup runs hourly; failures can delay physical removal and are retried. Recoverable backup copies are restricted to recovery, with sanitisation and deletion replay required before restored access. We do not describe pending or unverified file deletion as complete. Copies you download are controlled by the recipient, and offline devices must reconnect to receive changes.

Your rights and choices

Settings lets you export or delete your saved chats in the current organisation. This does not export every business record or file, or erase all personal data held across your organisation's systems. Contact us for access, correction, erasure, restriction or portability requests where those rights apply.

You may object to processing based on legitimate interests. Contact antony@tempworks.ai so we can assess your request. You may complain to the Information Commissioner's Office.