Your endpoint and gateway tools never see a ChatGPT prompt go out. BeeSensible highlights sensitive data in the text people type in AI tools, email, and chat, and shows you in numbers where the risk sits.
Who this is for
18
apps with a highlight while you type: AI tools, email, chat, and social
0 sec.
The text you type is processed and discarded at once
EU
Our own servers at Scaleway and Hetzner, no US cloud
Looking for the wider story for your organisation? That lives on AI governance for security and IT teams. DLP sits on the endpoint and on the gateway. But sensitive data increasingly goes into a text field in a browser tab: a prompt, an email, a chat message. Your existing stack does not watch there. BeeSensible highlights that data while someone types, so the user can remove it before it leaves the organisation.
From the field
An analyst has an unclear alert and wants to know quickly what it means. He pastes a chunk of log into ChatGPT and asks for an explanation. That log holds internal hostnames, an IP range, and a username with an email address. Triage gets faster, and internal data has now been processed by an outside tool.
Policy names two approved AI tools. In practice people use ten: for summaries, code, translations, email. You never approved them and you cannot see what data goes in. A ban only moves it to personal laptops.
A report comes in that customer data ended up in an AI tool. Management wants to know what exactly was shared, through which tool, and how often this happens. You have an AI policy and a training, but no numbers on what actually happens at the moment of input.
The catalog tracks 865 AI tools. For close to a hundred of them the vendor's ownership is opaque, and dozens sit in high-risk jurisdictions; those carry a high or critical risk profile before a single prompt is entered. When someone in your organisation opens one, you see it appear. Then you decide: allow it, allow it with a condition of your own, or not allow it. Anyone opening the tool afterwards sees that decision in the browser, with the approved alternative leading.
Engineers can remove credentials before asking an AI tool or teammate for help.
A log export, a threat intelligence write-up, a postmortem: open it in the Mac or Windows desktop app, and IP addresses, email addresses, usernames, and credentials are marked for you to remove before you share the document or paste it into AI.
First click at 09:14 by d.hofman, from 145.94.28.117. The message arrived at [email protected].
The analysis showed the submitted password Zomer2026! was also in use on two other services.
An impression of the anonymisation screen. Detection runs on the document you open yourself.
Why this is hard
Sensitive data today travels through AI tools, email, and chat in the browser. Endpoint and gateway tools miss that moment, because the text sits in a text field, not in a file or an upload.
Staff use dozens of AI tools you never approved. Without a signal in the browser, you cannot see what data enters them, and you cannot tell what risk you are carrying.
A hard block frustrates people and pushes work to channels you cannot see. People need help at the moment they type, not a rule they route around.
GRC, auditors, and NIS2 do not ask for policy on paper, they ask for evidence that a technical control is actually active. That evidence is hard to deliver if you do not see the moment of input.
Across the risk surface
The same gap shows up in different forms, from Shadow AI to incident response.
Shadow AI
Unapproved AI tools people use every day, without security knowing what data goes in.
The moment of input
The text field in a browser tab that endpoint and gateway DLP cannot see.
Incident response
Logs, threat intelligence, and alerts that analysts paste into AI tools to triage faster.
Awareness
A signal at the moment of input that sticks, unlike a training from six months ago.
Oversight and reporting
Aggregate counts by tool and category that GRC and auditors can use.
How BeeSensible helps
Sensitive details get a highlight while staff write. They decide what to remove, replace, or mask.
Sensitive data gets a highlight in the text field while someone types in AI tools, email, and chat. No gateway, no endpoint agent required.
Runs in Chrome and Edge, in the web apps your staff open every day, plus a desktop app for macOS and Windows for document redaction. No proxy, no network change.
On a highlight the user chooses: remove, replace with a realistic alternative, or mask. The extension never changes text on its own and never blocks sending.
The dashboard shows where the most detections sit by tool and category. The text people type is never stored and cannot be read by anyone.
For CISO, GRC, and security leadership
BeeSensible covers the moment of input in the browser and delivers the numbers GRC and NIS2 ask for: markings adjusted or sent anyway, what happened after a tool notice, and what share of AI use runs through approved tools. Without looking over your own people's shoulders.
Total detections
12,438
Top apps
Example dashboard. Counts and types only, never content.
CISO
Show GRC, auditors, and the board that an active control sits at the moment of input, backed by figures per period: markings adjusted or sent anyway, a recorded decision per AI tool, and what people did with the notice.
GRC and compliance
The dashboard shows no text and no single people: only aggregate counts, with no per-employee drill-down anywhere. Insight into patterns, not surveillance of people.
Security architecture and IT
No proxy and no new application. The extension runs in Chrome and Edge and rolls out centrally. Detection runs on our own servers at European providers, with no US parent, and all traffic is encrypted in transit.
Honest answers
Straight answers to the questions people actually ask before rolling this out.
The extension analyses text in the input fields of supported tools to highlight sensitive data. To be clear about the mechanics: that text travels to a BeeSensible server at a European provider, is processed in working memory, and is discarded at once. The content is never stored and cannot be read by anyone, not even an administrator.
No, BeeSensible blocks nothing. That is deliberate: a hard block drives workarounds. The user sees a highlight and chooses what to do, and you get aggregate insight into where the risk sits.
No tool makes you compliant on its own. BeeSensible helps with GDPR and NIS2 by covering the moment of input in the browser and backing up that the control is active with real numbers. Your organisation stays the controller, BeeSensible is the processor, and a processing agreement is signed.
Yes. The detection engine handles Dutch and English reliably and recognises both personal and technical data, such as names, account numbers, email addresses, and credentials that show up in logs and incidents.
Limited. There is no proxy to configure, and you change nothing on the network. The desktop app for document redaction is optional. The extension runs in the browser your organisation already uses and rolls out centrally through your management console. You see your first detections in minutes.
Compliance
GDPR
Covers the moment personal data is entered and backs up accountability with real numbers.
NIS2
A demonstrable technical control against unwanted data sharing through AI tools, email, and chat.
Processing agreement
A processing agreement is signed with every customer. A product DPIA is available on request.
EU processing
Detection runs on our own servers at Scaleway in the Netherlands and Hetzner in Germany: European companies with no US parent, both ISO 27001 certified.
BeeSensible runs in the browser your organisation already uses. No proxy, no network change, and you see your first detections in minutes.
Rens, BeeSensible
I am happy to give you a twenty-minute tour. Leave your details and I will find a moment that suits you. Calling or emailing works just as well.
ChatGPT