Where does our data live?
SOC 2 criterion C1.1 covers identifying and keeping confidential information, so a customer reading your report asks which stores hold their records, copies included.
SaaS. Technology. IT services.
Your customers ask where their data sits and who can reach it. IRIS answers from scans of the stores you connect, with each finding’s date.
Four paid engagements and three proofs of concept.
SOC 2 criterion C1.1 covers identifying and keeping confidential information, so a customer reading your report asks which stores hold their records, copies included.
ISO/IEC 27001:2022 control A.8.33 has test information selected, protected and managed. A staging database restored from production is one place that control applies.
Criterion CC6.1 asks for logical access security over protected information, and the customer’s questionnaire turns that into a list of named roles.
Records for a SaaS company’s stores, from production to support exports, with one candidate rejected and one deployment left out.
| Source | Location | Detected type | Why | Sensitivity | Decision |
|---|---|---|---|---|---|
| Production database | app.users.email | Email address | The customer’s end users, by tenant | Confidential | Keep; reads limited to the support_tier2 role |
| Staging database | staging.users.email | Email address | Restored from a production backup for a load test in September | Confidential | Replace with seed data, then re-scan; A.8.33 applies |
| Support tool export (Amazon S3) | support-exports/tickets_2026-10.json | Payment card number | A customer pasted a card number into a ticket, and the nightly export kept it | Restricted | Delete the export; the support owner redacts the ticket |
| Analytics warehouse | events.page_views.url | Email address | Emails carried in URL query strings into the event log | Confidential | Strip the parameter at collection; purge the history |
| Seed data repository | fixtures/users.csv | Email address (candidate) | Addresses at example.com written for automated tests | None | Rejected by the reviewer: test fixtures, not customer data |
| Customer-hosted deployments | Single-tenant installs in customers’ clouds | Not scanned | The customer runs these; your team has no standing access | Unknown | Excluded: the customer scans its own install, and your answer says so |
The AICPA’s Trust Services Criteria set what a SOC 2 report tests: CC6.1 for logical access security, and C1.1 for confidential information, where the confidentiality category is in scope. The 2017 criteria still apply; the AICPA put a 2027 revision out for comment in October 2026.
ISO/IEC 27001:2022 includes A.8.33, test information, and A.8.11, data masking (new in 2022), which the protection check on sampled values supports once it is activated on your install. A questionnaire may quote neither, but its questions on data location, test data and access map onto them.
Our position: answer each location question with the store, the scan date and the coverage record, so the customer can see what you checked and what you didn’t.
Fill the questionnaire’s data-location and access questions from the latest scan, and attach its date and coverage record instead of a narrative.
Lists each store holding customer data, the roles that can read it from table and view grants, and the stores it couldn’t read.
Scan staging and test databases beside production after each restore, so real customer records in a test environment are found the same week.
Scans run on a schedule or on demand, and each dated run shows whether staging holds personal data again.
Reads a sample of each table across the 12 database and warehouse engines, including PostgreSQL, MySQL, SQL Server, Snowflake, BigQuery, Databricks and SAP HANA.
Reads JSON, CSV and text exports in connected cloud storage, where ticket exports and log files land.
Checks sampled values for masking, hashing or encryption, once the check is activated on your install, so a staging copy left in plain form is flagged.
IRIS can be deployed in your own infrastructure, on-premises or in any cloud, so your data stays within your environment.
Capabilities vary by edition.
IRIS shows where customer data sits and who can reach it. It doesn’t answer the questionnaire’s control questions on your behalf, such as key management or incident response, and it doesn’t clean a staging database.
Some growth-stage companies Dipal Panchal speaks with have no chief information security officer (CISO) yet: a virtual CISO, or the chief technology officer covering everything. The evidence then has to be readable by whoever signs the questionnaire.
What each framework or law asks of the customer data a SaaS company holds.
For SOC 2, ISO/IEC 27001 and UK GDPR, IRIS maps its evidence to their controls through its unified compliance framework. A scored readiness view for these in SaaS isn’t in v1; ask us if you need one.
Connect the production database and its staging copy with read-only accounts, and run one scan across both. The how-it-works page follows that scan, and the architecture page answers your customer’s questions about IRIS itself.
Yes, as an export of the evidence register or a report, with each scan’s date and coverage. Decide first which tenants’ findings a customer may see.
Some, while it runs, because any scanner reads through queries. IRIS reads a sample of each table and is designed to limit load on production sources; to keep it off the primary, point it at a read replica.
Yes, in memory, to classify it. It keeps findings and coverage records on your host unless you turn on emailed reports. Stored samples are masked by default wherever personal data sits in a column.
Bring the last security questionnaire a customer sent. We’ll show which answers IRIS evidence supports.