Kayzen Copilot
Understand an alert in one sentence
Kayzen Copilot answers in plain language: why this alert, which computers are out of date, what needs to be done. It suggests actions, but a human always approves.
The problem
A security alert is only useful if someone understands it and knows what to do. In a small organisation, that person is not an analyst.
What the module does
Natural-language questions
“Why this alert?”, “Which computers are out of date?”, “Prepare this month's report”.
Human approval
Any suggested action (isolate, approve an elevation) must be confirmed by an authorised person.
Strict separation
The assistant only accesses your organisation's data, and every tool call is logged.
Daily summary
A short, understandable summary for non-technical managers.
How it works
Question
You ask a question in the console.
Search
The assistant consults your data through read-only tools.
Suggestion
It answers and suggests an action, which you approve or not.
What you should know
- The module is in development and will be evaluated against a set of questions and injection tests before release.
- Your data is not used to train AI models.
- Analysed content (emails, files) is treated as data, never as instructions.
Frequently asked questions
Questions about this module
Which AI model is used?
Copilot relies on Anthropic's Claude API. This provider will appear in the list of sub-processors before the module opens.
Want to try this module?
The pilot programme gives you access to modules as they are delivered, free of charge and with no commitment.