Point-in-time is enough when
You need a baseline, or a dated answer for one audit or customer review, and the sources in scope change slowly.
Category comparison. Scan timing.
Three ways to time a data scan, and what each lets you show an auditor over a 12-month period.
By Dipal Panchal, co-founder and CEO · Updated November 8, 2026
A point-in-time scan records what was in a source on the date it ran. A scheduled scan repeats that on a fixed rhythm, so you build a dated series. An event-driven scan runs when the source reports a change, such as a new file or a changed permission.
Each one shows only what it observed. NIST’s guide to assessing controls leaves the frequency of assessments to the organization, so the rhythm is a choice you make and defend. Match it to how fast the data can change.
Explore scan coverageCells describe the typical tool for each scan-timing model.
| Aspect | Point-in-time | Scheduled | Event-driven |
|---|---|---|---|
| Acts on | Each source in scope, once | Each source in scope, on every run | The changes a source reports |
| Main strength | A baseline, or a dated answer for one review | A dated series an auditor can sample | Fast notice of a high-risk change |
| What it cannot do | Show anything that changed after the scan date | Show what happened between two scans | Show the full state, or changes the source never reported |
| Best fit | A first assessment, or a check before an audit or customer review | Showing that the same check repeats on a known rhythm | A file shared outside the company, or a changed permission |
| A typical scenario | A pre-audit scan of the billing database | A weekly scan of the same database through the audit period | An alert when a folder is shared with anyone who has the link |
| Where it runs | Wherever the scanner runs, started by a person | On a scheduler that starts the scanner | A service that receives notices from each source |
| What it needs | Read access at scan time | Read access, plus a scheduler and capacity for each run | A source that emits events, a live subscription and a service ready to receive them |
| Load on the source | One read of each source in scope | A read of each source on every run | Small reads per event, plus a first scan to set the starting point |
| Gaps to plan for | It ages from the day after the scan | The interval between runs is the gap | Notices can arrive late or more than once, and subscriptions can expire |
Take an audit period of January to December 2026, covering one database that holds personal data. The auditor sets the questions.
| Audit question | Point-in-time (Dec 1) | Scheduled (monthly) | Event-driven |
|---|---|---|---|
| Where was personal data in March? | Cannot say; the record starts December 1 | Answered by the March scan, with its date | Answered only if a baseline scan came first |
| A permission added in June and removed in July | Invisible | Seen only if a scan ran between the two dates | Seen if the source reported both changes |
| A source unreachable for two weeks in August | Not in the record | A failed run on record, with the reason | A silent gap if the subscription lapsed |
| Does the rhythm match the policy? | No rhythm to show | Twelve dated runs against the policy | Depends on what the source emits |
| What the auditor can retest | One dated state | Any of the dated states, by reading the source again | The reported changes, not the full state |
You need a baseline, or a dated answer for one audit or customer review, and the sources in scope change slowly.
An auditor wants the same check to repeat on a known rhythm, and the gap between runs is acceptable for the data in scope.
A single change matters fast, such as a file shared with anyone who has the link. Events show change, not the full state, so pair them with scans.
You need a baseline and a dated series for audits, plus fast notice of the few changes that can’t wait for the next run.
Change something on a test source, then check what each model recorded.
IRIS runs point-in-time scans that you schedule or run on demand.
Scans run on a daily or weekly schedule, or on demand before an audit. Every scan is dated, and a re-scan changes only what it covered.
For every scan, IRIS records what was in scope, scanned, failed or excluded. Any file it couldn’t read appears with a reason, and unknowns show as unknown, never zero.
Evidence is generated from scan data and collected in a register. You can review it and export it as CSV or XLSX.
After your team fixes a finding in the source, a re-scan records the new condition with its own date.
Capabilities vary by edition.
What IRIS reads, what it keeps and how each scan runs, from connection to evidence.
What a scan covered, what it couldn’t read and why, and why “not measured” is not zero.
A scope record, an evidence register and framework reports, with every finding linked to its source.
Set a daily or weekly schedule, or run a scan on demand. Every scan is dated.
You do. NIST leaves the frequency of control assessments to the organization, so set the rhythm in your policy and keep to it. Your auditor checks that you did.
On the source emitting an event for the change, and on the subscription staying active. Notices can arrive late or more than once, so a full scan still sets the record.
Yes, for many questions, when the rhythm matches the risk and every run is dated. Where one change matters fast, add alerts from the source itself.
Change a test permission between two runs, then book a demo to see a dated re-scan.