BoletusOPS

Alas, 45 days after announcing the launch of BoletusOps, my corporate AI usage safeguard, I have not found a single potential client willing to become an early adopter. And, as my experience tells me, that is usually an indicator of an idea’s failure.

The product was meant to help organisations make compliance happy, keep sensitive information more secure, and reduce AI costs by using an API proxy to major AI platforms instead of buying expensive subscriptions for everyone. That still feels like a real problem to me, especially for companies where staff are already using third-party AI tools without much visibility or control. BoletusOps was designed exactly for this: policy-based routing, audit trails, sensitive prompt controls, and cost governance.

I attribute the failure to my exceptionally bad sales skills. But the market is probably not ready either. There have not yet been enough major security incidents involving corporate AI usage, so not many people feel the risk strongly enough. Larger organisations are also building their own AI tools and safeguards.

I also reached out to my contacts, looking for a partner or co-founder, but found no real interest. And this is the kind of project where you need to focus either on building an exceptional product or selling it. Not both at the same time. That is why a co-founder matters. Unfortunately, everyone I reached out to is either busy with their own projects and jobs, or has other reasons.

So I am putting BoletusOps on hold.

If you are interested, reach out and chat to me. Maybe you want to become a co-founder and know prospective clients who might care about corporate AI usage. Or maybe you want to build something like this internally, and I can help as a consultant.

Meanwhile, I am going to tidy up the code, shut down demo server, and move on to something else. Fortunately, I treated BoletusOps as an MVP, so not much work went into it, and the only layoff is myself.

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.