I’m Building BoletusOps

Rene Rivkin went to jail because inside information was shared with him. Today, people put sensitive company information into AI tools every day:

• internal financial and legal papers
• sensitive emails
• spreadsheets with client private details
• source code (manageable risk for me, but for you?)
• and sometimes even passwords or payment data

Usually not because they are careless. Usually because they are under pressure to work faster, and today it is hard to stay productive without using AI.

That is why I think a serious AI-related incident inside many companies is only a matter of time. Not because AI is bad. But because AI adoption is moving faster than controls.

So I started building BoletusOps.

The goal is simple: help organisations use AI without losing visibility and control.

BoletusOps is designed to route AI requests through approved paths, block or reroute unsafe ones, support private local models where needed, and keep an auditable record of what actually happened.

In plain English: staff should be able to use AI, but the organisation should know where sensitive data is going, which model is handling it, and what should be blocked or allowed.

BoletusOps is currently available for a limited number of pilot engagements with Australian organisations that are already feeling this tension: productivity on one side, security, privacy, compliance, and client confidentiality on the other.

If this pressure sounds familiar inside your organisation, message me.

I am looking for early partners who want to shape the product with real-world requirements.