Accountability: what the figures show
Three topics with two figures each over a month or a quarter, the explanation per provision, measurement coverage, and the report you freeze as a PDF.
Accountability brings together what you can substantiate with BeeSensible's measurements, and which legal obligations that is relevant to. It passes no judgement: the page shows figures, an explanation panel says why they matter legally, and the report holds both.
Switching it on
Accountability is off until an admin switches it on, under Organisation > Settings with the switch Accountability dashboard: "What you can substantiate towards the GDPR and other frameworks with existing measurements." The switch needs the Realtime Privacy and AI modules; without those two there is nothing to measure. Accountability then appears in the menu.
Accountability in the dashboardThree topics, measurement coverage, the provisions and the reports.The period
You pick a month or a quarter; the last full quarter is selected by default. Pick the running period and the page says: "This period is still running: the figures go up to today."
Three topics
Each topic shows two figures and one sentence about what they show, and what they do not.
| Topic | Figures | What it says |
|---|---|---|
| Visibility of AI use | AI tools observed and Tools with a decision ("x of y observed tools") | Which AI tools staff visited, and for how many of them a decision is recorded. Under review does not count as a decision. |
| Protection of sensitive data | Critical detections and Adjusted ("x of y masked, replaced or removed") | Only in the apps where BeeSensible helps; what happens elsewhere is not included. |
| Supporting staff | AI notices shown and Alternative opened ("after x of y notices") | These figures do not show whether staff are sufficiently trained. |
Explanation per provision
At each topic, Explanation opens a panel with three questions per provision: What does the law say?, What does BeeSensible contribute? and What falls outside? Below that, separate from our explanation, the Literal text with the Official source (EUR-Lex, wetten.overheid.nl).
| Topic | Provisions |
|---|---|
| Visibility of AI use | NIS2 Article 21(2)(a) (policies on risk analysis and information system security) and (d) (supply chain security) |
| Protection of sensitive data | GDPR Article 5(1)(c) (data minimisation) and Article 32(1) (security of processing) |
| Supporting staff | NIS2 Article 21(2)(g) (cyber hygiene and training) and AI Act Article 4(1) (AI literacy) |
| Supporting information for the measurements | GDPR Article 5(2) (accountability) and Article 24(1) (responsibility of the controller) |
For NIS2 the Dutch implementation is shown as well: the Cyberbeveiligingswet, Article 21(3). Those duties apply only to essential and important entities.
The overview of AI tools and the decisions about them is input for your record of processing activities (GDPR Article 30) and for a DPIA (Article 35). It does not replace them.
Measurement coverage
Below the topics the page shows how large the part is that BeeSensible measures: Seats with BeeSensible activated (how many seats in the subscription have reported in at least once with the extension or the desktop app) and AI visits to an app where BeeSensible helps, an estimate. How is this calculated? explains that the names of the observed AI tools are compared with the apps where BeeSensible helps, and that visits are counted, not text. Coverage is context for the figures, not a condition.
A figure, or honestly no figure
Every figure is a value, No measurement data or Not measurable, with the reason: a source that did not respond, no account activated yet, statistics off, a module off, AI notices off, or nothing to calculate a share from. A missing figure is never shown as 0; a true zero reads "Nothing observed in this period."
Below the privacy threshold there are no figures and no report: "Figures appear from 5 active users."
Reports
With Create report you fix a period. "The report is stored and does not change afterwards, not even when decisions or settings change." It holds the three topics with their figures, the coverage and how it was calculated, the Scope, per provision the explanation and the literal text with its source, the Source of the decisions (as they stood at the end of the period, where the register history allows it) and the Settings when compiled: where detection runs, the apps where BeeSensible helps, the modules and statistics.
Every report carries a Checksum (SHA-256), shown in the dashboard under Accountability > Reports and in the PDF; with it you show the report has not been changed. If there is an earlier report of the same kind, month beside month or quarter beside quarter, the Comparison with the previous report is included, on the figures, without judgement. Download PDF gives you the file.
See also Notices at AI tools for where the notice figures come from, and Sensitive data for the markings per app and data type.
