Frequently asked questions

How long does a first rollout take, and what do we need to provide?

Most programs start with one line or one inspection flow and go live in 2–4 weeks. You provide read access to the relevant MES, ERP, historian or inspection data and a named process owner; our team handles connector configuration, characteristic mapping and validation.

Do we have to replace our MES, ERP or historian?

No. The platform sits above your existing systems as a quality intelligence layer. It reads from MES, ERP, PLC/OPC UA, historians, cameras and gauges, and writes back only where you explicitly enable it, so your systems of record stay unchanged.

Where is our production data stored, and can it stay in our region?

Data residency is configurable per tenant across EU, US and APAC regions. Enterprise agreements also support private cloud and on-premise deployment where data cannot leave the plant network. Process parameters and recipes are never used to train models for other customers.

How does AI defect detection work in a validated environment?

Models run on your existing camera or edge inspection hardware and always operate as decision support. Every inference keeps the image, model version, confidence score and reviewer verdict attached to the record, so inspection decisions stay explainable and reviewable during audits.

Which standards and compliance programs does the platform support?

Customers use the platform to support ISO 9001, IATF 16949, AS9100D, ISO 13485, VDA 6.3 and FDA 21 CFR Part 11 evidence workflows, including electronic signatures, review history and immutable audit trails. Certification always depends on your configured process.

How does access control work across plants and suppliers?

Role-based access control is scoped by site, line, product family and supplier. It integrates with your identity provider through SAML or OIDC single sign-on with SCIM provisioning, so joiners, movers and leavers follow your existing IT process.

What does commercial engagement look like?

Pricing is based on connected lines, sites and modules rather than named users, so plant-wide visibility is not penalised. Enterprise agreements include security review support, a named success team, SLA-backed availability and a defined escalation path.

How do you prove value before a plant-wide rollout?

Every engagement starts with a scoped baseline: current first pass yield, scrap cost, investigation time and audit effort. The pilot report compares those baselines against measured results, which is the same evidence used for internal capital approval.

Can we run the platform without any outbound connectivity?

Yes. An air-gapped deployment runs the entire stack inside the plant network, including model serving. Updates arrive as signed offline bundles that are verified, staged and applied inside your change window.

What happens to inspection if the platform is unavailable?

Inspection continues. The edge gateway keeps evaluating against specification and control limits locally and buffers records for up to 14 days. The platform is never a control-loop or availability dependency for the machine.

How do you handle mixed-model production on one line?

Limits, sampling rules and control charts are bound to the product and operation, not the line. When the running order changes, the applicable inspection plan and limits change with it, and the chart context follows the product.

Do you support short-run and low-volume statistics?

Yes. Individual and moving-range charting, pooled variance across configurations and short-run standardised charts are available, which matters in aerospace and high-mix electronics where classical subgroup methods do not apply.

Can we export our data if we leave?

Yes. Full tenant export of master data, measurements, records, audit trail and attachments is available in open formats on request, and is written into enterprise agreements as an exit provision.

How are limits and specification revisions handled over time?

Every limit and revision is versioned with an effective period. Historical analysis is evaluated against the revision that was in force at production time, so a specification change never silently rewrites past capability.

Does the platform write back to MES or ERP?

Only where you explicitly enable it, per integration. Write-back actions are logged as events, attributable to a user or service account, and reversible. Connectors are provisioned read-only by default.

What is required to add a second site?

A gateway host, read access to that site's sources and a local process owner. Characteristic definitions, alert routing patterns and dashboards are reused from the template site, which is why later sites go live in days rather than weeks.

Security and validation questions go to security@trustqualityassurance.com. Integration and architecture questions go to support@trustqualityassurance.com. A solution architect answers directly; there is no ticket queue for evaluation work.