Privacy policy
How Telematric collects, uses, stores and shares information across device infrastructure, software connections and AI services.
1. Scope and contact
This notice covers the Telematric website, accounts, APIs, developer console and services operated by Sakura Inc. Privacy enquiries and requests can be sent to hello@telematric.com. For technical help, contact support@telematric.com or use our contact page.
For account administration, subscriptions, security and our own business records, we determine the purposes of processing. For operational information a customer supplies or authorizes us to access, we generally process on that customer’s instructions. A customer may itself act for another organization. The applicable agreement defines those roles and responsibilities.
If your information was supplied by a company using Telematric, contact that company first about its purposes and records. We can help identify the appropriate contact and assist with requests concerning the information we process. This notice does not replace the privacy notices of customer applications or third-party systems.
2. How we handle data
Telematric stores operational information needed to provide its services. Connecting a system does not mean we copy its entire database. Different services use live reads, selected stored snapshots or hosted operational records.
| Information | Current handling |
|---|---|
| Account and workspace information | We store account names, email addresses, country, calling code and phone number when supplied, workspace memberships, roles and service access. Registration and account details may be reviewed by authorized platform staff. |
| Device and installation information | We store customer/site/device references, optional contact and location information, configuration, device identity, accepted readings and their measurement/receipt times. Readings may reveal activity or location when linked to identifiable people. |
| Alerts, commands and firmware | We store rules, incidents, command and delivery outcomes, webhook payloads/attempts, firmware artifacts and update records as applicable to the enabled service. |
| SAP and IBM Maximo | Supported equipment, location and work-order operations read the source on demand. We retain connection settings, encrypted credentials, explicit identity links and operational/audit metadata, rather than a complete asset inventory mirror. |
| GIS and laboratory connections | Selected tracked Giswater records and mapped SENAITE results can be retained as timestamped snapshots and history. The lims.science bridge currently reads supported samples/results on demand and stores configuration, mappings and operational metadata. |
| Water records and calculations | We store submitted utility identities, meters/readings, consumption records, tariff versions, reviewed calculations and export/audit evidence. Customer financial systems remain separate. |
| CRM and merchant connections | Supported operations can retain identity links, explicitly requested handoffs/checkouts, delivery/reconciliation records and status snapshots. They may send the information needed for an authorized operation to the selected provider. |
| Weather and planning | Upstream weather responses use a short server-side cache, currently valid for ten minutes. Stateless planning APIs process supplied scenarios without creating a customer installation record. Saved sites, training samples, security logs and other explicitly enabled features have separate storage behavior. |
| AI training | Authorized samples, training policies and audit records are stored by workspace. Customer-specific model artifacts, versions and evaluations use private storage separated by workspace and supported service. |
| Billing, enquiries and diagnostics | We store subscription/payment references, amounts, status and receipts, support messages, request/usage metadata and security records. Payment providers handle payment instruments; do not send card details or secrets in support messages. |
Source applications remain authoritative for their records and workflows. A laboratory snapshot describes a particular source state and collection/publication time; it is not a live water-quality measurement. Stored or cached records may become stale. Source references, timestamps and freshness should be checked before relying on them.
3. Why we use information
We use information to create and secure accounts, authenticate devices and integrations, deliver subscribed operations, maintain history, evaluate configured alerts, deliver webhooks, process subscriptions, respond to support enquiries, prevent abuse and investigate failures.
Depending on the activity and applicable law, the basis is performing a contract or taking requested steps before one, complying with legal obligations, pursuing legitimate interests such as service security and administration, or consent where required. When relying on legitimate interests, we consider the effect on individuals and their rights. Customer-directed operational processing follows the customer’s instructions and the applicable processing agreement.
Account and service emails include verification, password recovery, welcome, subscription and expiry information, and supported workspace connection notifications. Registration can generate an internal signup notification for our authorized operations team. We do not treat account creation or payment as permission for unrelated marketing or model training.
4. AI, predictions and training authorization
A workspace owner or administrator separately authorizes supported customer-specific training and selects the permitted services, sources and retention setting. We do not automatically access a customer’s business database or combine different customers’ training samples into a shared dataset.
We do not sell customer operational data or use it to train another customer’s model without separate authorization. Reusable software and training methods do not require sharing customer datasets. Current model selection is scoped by workspace, service and policy revision.
Some responses use rules or physical calculations when an eligible trained model is unavailable. Predictions and recommendations are estimates, not guarantees or automatically authorized equipment actions. The customer remains responsible for appropriate human review and decisions in its application.
An authorized workspace manager can disable training through the available policy controls. This removes stored training samples from the active training store and stops selection of the affected trained models under the current policy. Private-artifact cleanup is attempted and may require a later maintenance run if the runtime is unavailable. Other telemetry, billing, audit and backup records are handled separately. Contact us about a complete data request.
6. Retention, exports and deletion
We retain information for the purposes described here, taking account of the subscribed service, agreed instructions, operational history, security, disputes and applicable legal obligations. Different records have different lifecycles; there is currently no single automatic retention period covering all telemetry, snapshots, audit records and backups.
For authorized ML samples, the supported policy setting is 7–365 days, with a default of 90 days. Applying that policy and cleaning up artifacts requires the relevant maintenance/runtime operations. Backup copies and separate audit records are not covered by the training-sample setting. The weather cache has a ten-minute validity period; expiration of validity does not mean every stored copy is immediately erased.
Cancellation, disconnection and deletion are different requests. Stopping a subscription renewal does not automatically erase records, disconnect every source account, cancel dependent subscriptions or remove provider-owned records. Archiving a site or disabling a device preserves history. Source credentials may also need to be revoked at the provider.
Use the documented read APIs for supported data access. For a broader export, correction, account closure or erasure request, contact hello@telematric.com. These requests are currently support-assisted; a universal self-service workspace export or erase button is not provided. We verify authority, explain the applicable scope and timing, and handle the request in accordance with applicable rights and legal requirements.
Backups may temporarily retain information after removal from active systems. We will explain applicable backup handling when arranging deletion. Necessary financial, security or dispute records may require separate retention with restricted use. A future plan’s retention period or proposed roadmap target applies only when it is expressly included in the applicable service terms; internal plans are not current guarantees.
7. Security and your controls
The platform uses authenticated access, workspace and API-key permissions, hashed account passwords and device/API credentials, and encryption for selected retrievable connector secrets. These protections do not mean every stored field is individually encrypted or that no security incident can occur.
Keep workspace and provider keys on your backend, limit permissions, rotate compromised credentials, secure your hardware and maintain your own necessary records. Workspace owners control membership and selected connections; an integration’s permissions do not grant rights at a third-party system that the source account does not have. We investigate incidents and provide notifications required by applicable law and agreements.
8. Browser storage and third-party pages
The developer console uses tab-scoped session storage for its session and workspace selection. Verification/reset links may contain one-time tokens that the console removes from the visible URL after reading them. These are functional mechanisms, not permission for unrelated tracking. Enabled reCAPTCHA and hosted payment pages have their own data practices. Follow their notices when using those features.
This notice does not govern independently operated customer applications or third-party websites. Changes to optional tracking or new processing purposes will be disclosed, with consent obtained where required.
9. Your choices and rights
Depending on applicable law, you may request access, correction, erasure, restriction, portability, withdrawal of consent, objection to processing, or human review of certain automated decisions. These rights may have exceptions. Withdrawing consent does not make earlier lawful processing unlawful.
Contact hello@telematric.com with enough information to identify the relevant account or records. We may need proportionate identity or authority verification; do not email passwords, API keys or unnecessary identity documents. We respond within applicable legal time limits and explain material restrictions. You may also complain to your relevant data-protection authority, including the Nigeria Data Protection Commission where applicable.
10. Updates
We update this notice when our services or processing change and revise the date above. Material changes are communicated through appropriate service channels. A new notice does not substitute for consent or other authorization required for a new purpose. Read our Terms of use for customer-data responsibilities.