MetricSpark is designed to sit quietly alongside your operations — reading the signals your team already produces and turning them into answers, without changing your systems or taking your data anywhere it doesn't belong. This page explains how the product is built and the principles we hold ourselves to.
MetricSpark observes; it does not act on your systems. Connections to your source systems request read-only scopes wherever the platform supports them. The product does not create, edit, close, or delete records in your PSA or other tools — your source of truth stays exactly where it is.
Each integration asks only for the narrowest set of permissions it needs to do its job. Access is scoped per connection and can be revoked by you at any time from the source system. We do not request broad administrative rights we don't use.
Data is encrypted in transit using TLS, and encrypted at rest by the underlying cloud platform. Secrets and access tokens are stored in managed secret storage, never in plain text in application code or logs.
We keep only what a feature needs and no longer than necessary. Where the product can work from aggregates or recent windows rather than full history, it does. On account closure, connected-system data can be purged on request.
Sign-in supports single sign-on via Microsoft Entra. Access inside the product is role-based, so people see the views appropriate to their role. Administrative actions are limited to authorised users.
MetricSpark runs on reputable managed cloud infrastructure with network isolation, automated patching, and platform-level protections. We rely on the provider's physical and infrastructure security controls rather than operating our own hardware.
Every answer the AI gives is grounded in your own data and shows the numbers behind it — it does not invent figures. If the data to answer a question isn't available, the product says so rather than guessing.
Everything shown publicly on this website — screenshots, the interactive demo, and any sample deliverables — uses 100% synthetic, fictional data. It represents no real organisation, customer, or dataset.
If you believe you've found a security issue, please email support@metricspark.info. We welcome good-faith reports and will work with you to confirm and address valid findings.