How Seamium's Focus on Energy Governance Drives SaaS Startup Success

I worked directly with the operators for nearly a year, examining energy intelligence; only then did the real issue become apparent. The problem wasn't with energy intelligence; it was with compute energy governance. The necessary components are already available, as are the energy and compute power. What is lacking is proper governance between them, and none of the current options are suitable. Manually managing the energy is too slow, traditional software is static, and AI-powered tools add intelligence without giving operators any real control. What the operators needed was a governance engine that operated according to their policies, not one that was autonomous, not one that was manual, not one that was static; it should be self-learning, but only by learning from the infrastructure it governs. Crosentis is the product we developed at Seamium, the company I founded to address this very issue, and it has been trained on actual energy signals and operational data rather than on general assumptions. That's it, no new metric, just a new layer.

I went for suitability over quantity; in every case, two pilot partners, each with a real, specific problem, performed better than six generic ones. I never ensured proper alignment at an early stage; instead, I approached any operator willing to talk to us. There was no common problem and no mutual sense of urgency. As a result, those conversations achieved nothing and wasted weeks, which I did not have. It was clear that if a pilot partner doesn't feel the problem, they will hesitate, delay, and in the end withdraw; partners who are aligned make better progress than six who aren't, and they tell you the truth about what is wrong. At that time, I replaced the filter. Nowadays, before conducting any pilot conversation, I look for genuine alignment rather than just interest. That single change has been more effective for our pipeline than any of the outreach techniques we've tried.

Most solutions are designed to replace the operator; ours isn't, and that is precisely what our customers have repeatedly told us they need. The most common complaint we receive is that the available tools show data without helping people take action. They offer dashboards and charts without suggesting any decisions. Operators have to interpret the information themselves, on top of everything else they are responsible for. What they actually desire is simplification, not replacement; they want at least 80% of the manual work eliminated while still maintaining full control over how their infrastructure functions. Not automation, which takes away the power to make decisions, but governance that uses policies and data from operating infrastructure to make decisions while the operator remains in control. Which is why Crosentis was designed with that division in mind, to provide full visibility, offer real support for decision-making, and give the operator 100% control. Crosentis uses a proprietary machine learning model to take action, and it's always improving by learning from infrastructure and operator policies.

Most people think the buyer is the one who falls in love with the solution first, but in this case, that is usually not the case. The individual who is enthusiastic about Crosentis is usually not the one who has the final say as to whether the product enters production. Since it is a critical system, making a mistake here could cause the infrastructure teams that rely on it to suffer. For this reason, even if a buyer is truly interested in it, they must have it go through the approval stages, passing the review of those who weren't present for the initial demo and who must approve before anything is applied to a live environment. The lesson I have drawn is to address these two questions at the beginning, namely, who needs it and who has the authority to approve it. These two individuals are often completely different. The quickest way to get both parties to agree is to be transparent. You must clearly explain why the solution is better and, equally clearly, why it poses no risk to their infrastructure. It is only then that the solution moves from being something an interested buyer considers to being one that is actually approved for deployment.

We are concentrating on the largest infrastructure environments in the world. In this stage, the emphasis is on providing proof, not merely on deployment; that is, we are showing that Crosentis doesn't just detect problems but actually takes action, all within the actual constraints that operators have to work with. Two pilot projects are already in production. In the following phase, this proven approach will be expanded to more complex, larger-scale, and higher-stakes environments, without sacrificing the level of control operators need.

I had no direct involvement in the energy industry, but I did have direct contact with it. Before I began my work, I conducted nearly a year of research to determine which of the problems in this area actually needed to be solved. It isn't the case that every obvious problem does. The research proved to be useful thanks to my experience as an IT engineer. For many years, this role has kept me in close contact with infrastructure teams and energy operators. I was familiar with the way they operated, with what they would accept and what they wouldn't. It was this close association that turned the research from a mere guess into a real product decision.

I attempted to develop a product for several teams simultaneously, believing that broad appeal would be the safer option at that stage. That was not the case. When the product was spread too thin across so many different applications, it lost its effectiveness in all areas and failed to solve any of the problems well. Therefore, I changed my approach to be more focused, I decided to develop a single product that is truly excellent at addressing one particular problem in compute energy governance, rather than creating an average product for a variety of uses. Begin with a narrow scope and expand later, after you've justified it.