What is the Claude plugin?
Your people are already pasting development reports into Claude. This is the version your security team can review. Wave Connect Develop returns validated development insight inside Claude, with permissions enforced by us rather than by the assistant.
How does it work?
Claude calls the plugin as a tool. The answer is built by our platform from completed assessment profiles, so what appears is our validated language and not the model forming its own view of a colleague.
What people ask it varies. Their own results, their team's, or an aggregate view across a population. Every answer is grounded in a real assessment, which is the thing a general model cannot do however well it is prompted.
- Development insight drawn from that person's completed behavioral assessment
- Only what that same person could already see by logging into the platform
- Our own validated language, generated on our side rather than by the model
- The same answer if you switch assistant or the model behind Claude changes
Who sees what?
Three scopes, enforced by us. An employee sees their own profile in full. A manager sees themselves and their own team, and a request for a direct report's full profile is refused server side, every time, including upward. A power user sees themselves and aggregates, with a floor of three completed people on every aggregated view so that filtering down to one person does not work.
What will it not answer?
It does not answer on anyone without a completed assessment, and it infers nothing from a job title. It refuses anything beyond your permission. It declines any group below the floor of three, and says why. It will not rank people, because a profile describes how someone works rather than how well.
Claude is one of the assistants we test against. The plugin connects to Wave Connect Develop only: it does not carry hiring assessments, and it does not supply learning content.
Is any of our data used to train an AI model?
No. Nothing in the chain uses your data to train a model, ours or Claude's. The plugin passes validated language from a completed assessment profile into the context window, and the answer itself is generated by our platform.
Does it make decisions about someone's job?
No. It does not rank people, it does not expose an individual profile beyond permission, and a profile describes how someone works rather than how well.
Where does the data go?
Profile data enters Claude's context window. That is how it appears in the conversation. What Claude keeps is governed by your agreement with them, exactly as it is for every other document your people put in.
Talk to us
If your people use Claude and you want to put this in front of your security team, book a call. Tell us what your reviewers need and we will bring the right people, including the documents that do not sit on a web page.