Trust Center

SmartX HUB Security and Trust Center

This page exists to make your security review faster. SmartX HUB security architecture, access control, encryption, audit logging, deployment options and data handling are documented here in one place, so your IT and information security teams do not have to assemble the picture from a dozen pages and a sales call.

Where something is not published here, it is because it belongs in a signed agreement or a controlled document rather than on a public page. Ask and we will provide it under the appropriate terms.

Security Architecture

How SmartX HUB Security Is Built

Security is enforced at the platform core rather than per module. Because tracking, maintenance, supply chain, sensors and worker safety all resolve to the same record, they also inherit the same identity model, the same permission checks and the same audit trail. There is no module with its own user list, and no side door where one capability family applies weaker rules than another.

Authentication

  • Two-factor authentication via time-based one-time codes, email codes or WebAuthn passkeys
  • Single sign-on through SAML 2.0, Azure AD, Google Workspace and LDAP or Active Directory
  • Just-in-time provisioning, so accounts follow your directory instead of being maintained twice
  • Deprovisioning through the same directory, so a leaver loses access where you already manage it

Authorisation

  • Twelve role profiles covering the standard separation of duties, extended where needed
  • Group-level data filters, so a user sees their site or area rather than the whole estate
  • Field-level access control, so sensitive attributes can be hidden without hiding the record
  • Multi-tenant isolation, so one organisation's data is never reachable from another's session

Encryption

Traffic between users, devices and the platform is encrypted in transit with TLS 1.3 as the enforced minimum, terminated at the Cloudflare edge. Data at rest sits on Google Cloud Platform, where storage is encrypted by default. Stored credentials and integration secrets are additionally encrypted at application level with AES-256-GCM, an authenticated cipher, so tampering is detectable rather than silent. Outbound webhooks are signed with HMAC-SHA256, which lets your receiving systems verify that a payload genuinely came from the platform and was not altered in transit.

Audit and logging

  • Tamper-evident audit logging with a configurable retention policy
  • Export to your own security monitoring platform, so logs live where your SOC already looks
  • A security report covering account, permission and authentication activity
  • Operational history retained alongside the record it belongs to, not in a detached log file

Application and integration surface

Integrations run over a documented REST surface published with OpenAPI, which means your team can review exactly what is exposed before anything is connected. Outbound events are signed, and API credentials are scoped and revocable independently of user accounts — so an integration can be shut off without disabling a person, and a person can leave without breaking an integration.

API token lifetime is set per deployment rather than fixed by us, so it can be aligned with your own credential policy instead of forcing an exception to it.

The wider architecture, including how one record serves every capability family, is described on the enterprise platform page. Where the platform locates people rather than assets, the governance and consent considerations are set out on connected worker RTLS.

Deployment and Data Residency

Where Your Data Lives Is Your Decision

The same platform runs in three deployment shapes. The choice is usually driven by data residency obligations, network latency to the shop floor, or an internal policy that predates the project. All three run the same code and the same security model — what changes is who operates the infrastructure underneath.

  Cloud Private cloud On-premise
Infrastructure operated by SmartX HUB, on Google Cloud Platform Your cloud tenancy, deployed as a virtual machine Your data centre
Data leaves your network Yes Stays within your tenancy No
AI inference Platform service Platform service or local Can run locally, so content never leaves the site
Patching and updates Managed for you Coordinated with your team Scheduled with your change process
Typical reason to choose it Fastest to start, least internal effort Cloud policy or existing enterprise agreement Residency rules, air-gapped sites, latency

The managed cloud runs on Google Cloud Platform, with Cloudflare in front of it handling edge termination and protection. Both are disclosed as sub-processors, so your legal team reviews a named supply chain rather than an unnamed one.

Hosting regions are available in the United States and the Middle East, and the region is selected per deployment rather than assigned by us. Backups are held in the same region as the data they protect, which means residency survives disaster recovery instead of quietly breaking at the point it matters most. It is worth checking that answer with every vendor you evaluate, because a backup copied to a convenient region undoes the residency commitment made on the main deployment.

The full sub-processor list and the terms covering access to customer data are set out in the data processing agreement, available on request.

Resilience, Availability and Vulnerabilities

Continuity and Response

Backup, recovery objectives, availability commitments and vulnerability handling are governed by the agreement covering your deployment, and they differ between a cloud tenancy we operate and an on-premise installation your own team runs. The sections below must be completed with the commitments your contracts actually make.

Backups run daily and are held in the same region as the data they protect. Recovery objectives, availability commitments and the remedies attached to them are defined in the agreement covering your deployment, because they necessarily differ between a cloud tenancy we operate and an on-premise installation your own team runs.

Customers affected by a security incident involving their data are notified by email from the platform. If you believe you have found a vulnerability, contact us directly rather than disclosing it publicly and we will respond — reach the team here.

Documents Your Team Will Ask For

Everything below is either published or provided on request under the appropriate terms. Nothing here requires a sales conversation first.

On certifications, one distinction is worth stating plainly. SmartX HUB runs on Google Cloud Platform, which operates its own audited infrastructure programme. Certification of the underlying cloud is not certification of an application running on it, and we will not present it as though it were. If your review requires a specific framework, tell us which one and we will answer precisely about what we can evidence today rather than implying coverage we do not have.

Send Us Your Security Review

If your team has a standard questionnaire, send it. If your architecture board needs a session with our engineers rather than a document, we will arrange one. Security review is part of the evaluation, not an obstacle to get past.

FAQ

Security and Trust — Common Questions

The questions information security, IT and procurement teams raise most often during a SmartX HUB security review.

Can we run SmartX HUB entirely inside our own network?
Yes. On-premise deployment runs the same platform inside your data centre, and the artificial intelligence layer can also run locally rather than calling an external service — which matters when the content being analysed cannot leave the site at all. A private cloud option sits between the two, deploying into your own cloud tenancy so data stays within a boundary you already govern.
Does it integrate with our identity provider?
Single sign-on is supported through SAML 2.0, Azure AD, Google Workspace, and LDAP or Active Directory, with just-in-time provisioning so accounts are created from your directory rather than maintained separately. Where single sign-on is not used, two-factor authentication is available through time-based codes, email codes or WebAuthn passkeys. The practical benefit of directory integration is deprovisioning: a leaver loses access where your team already manages it, instead of relying on someone remembering a second system.
How granular are permissions?
Three layers. Role profiles establish what a category of user may do, group-level data filters restrict which records they see at all, and field-level access controls hide individual attributes within a record they are otherwise allowed to open. That last layer is what lets a supervisor work with a person record without seeing personal data they have no business reason to access — a requirement that comes up in almost every deployment involving worker location.
Can we feed audit logs into our own SIEM?
Yes. Audit logging is tamper-evident with a configurable retention period, and events can be exported to your own security monitoring platform so your security operations team correlates them alongside everything else rather than logging into another console. A security report covering account, permission and authentication activity is also available inside the platform for teams without a central monitoring stack.
How is data separated between customers?
Multi-tenant isolation applies at the platform core, so one organisation's records are not reachable from another organisation's session regardless of which module is being used. Because every capability family resolves to the same underlying record, the isolation boundary is enforced once rather than reimplemented per module — which is the failure mode worth asking any vendor about, since a single module with its own data path is enough to undermine the whole model.
How do we verify a webhook really came from the platform?
Outbound events are signed with HMAC-SHA256, so your receiving endpoint can confirm both the origin and that the payload was not modified in transit. Integration credentials are stored encrypted with AES-256-GCM, an authenticated cipher, meaning tampering with stored secrets is detectable rather than silent. The full integration surface is documented with OpenAPI, so your team can review exactly what is exposed before connecting anything.
What happens to our data if we stop using the platform?
Your operational data is exportable through the documented REST surface throughout the life of the agreement, so extraction is not a favour you have to request at the end. Identification is built on open GS1 standards rather than a proprietary numbering scheme, which means the tags already attached to your assets remain readable by other systems.
We track people. How is that handled from a privacy standpoint?
Two design choices matter here. Analytical views derive from position data the platform already records, so adjusting what a supervisor can see does not change what is stored. And visibility is bounded by the same role profiles, group filters and field-level permissions as everything else, so access is scoped to an area rather than to the entire workforce. Purpose limitation, retention periods, consent and consultation with worker representatives remain your organisation's responsibilities as data controller — we recommend settling them before enrolment rather than after. The connected worker RTLS page covers this in more depth.
Will you complete our security questionnaire?
Yes. Send it through and we will complete it. A pre-completed standard questionnaire is also available for download from this page, which often answers most of what your template asks and saves a round trip. If your architecture review board would rather talk to engineers than read a document, we will arrange that session instead. Get in touch and tell us which format suits your process.

Explore our blog for insightful articles, personal reflections and ideas that inspire action on the topics you care about.

error: Content is protected !!