Kenneth Eversole holds a bit of a rare perspective when it comes to artificial intelligence. The CEO and co-founder of OpsCompanion, a St. Louis-based AI-focused startup, doesn’t parrot the same unrelenting hype about the technology that leaders of big and small companies alike have been trumpeting in recent years.
In fact, he’s fairly skeptical of OpenAI, Anthropic, and other players that provide the most widely used AI tools at the moment, especially their goal of creating artificial general intelligence.
Keep up with local business news and trends
Subscribe to the St. Louis Business newsletter to get the latest insights sent to your inbox every morning.
“This concept of machine learning, where if I get enough [computing power] and enough training data, I can do anything in the world, kind of breaks down in practical reality,” Eversole says. “If you were to make a superhuman from AI and I use all of your personal data, where does your input get to be delivered into that? Where [is] your consent really at the end of the day? That’s what we’re starting to see in 2026, that consent starts to break down.”
Out of a small office in the One Metropolitan Square building in downtown St. Louis, Eversole and his lean team at OpsCompanion are building technology to respond to concerns about control and data sovereignty in AI use—an area of growing focus.
“People are starting to wake up, and businesses woke up to it really quickly,” he says.
The company’s main product, Standpipe, is an application available for download or remote installation by an IT department. It provides a final check on AI prompts before a user sends them from a physical computer into the cloud. Eversole says it offers the ability to automatically redact details in a prompt, like a customer email or name, or block others that contain highly sensitive information, such as a Social Security number or the password to a key database.
He explains Standpipe runs in the background of a device, scouring for signatures of AI activity in outgoing web requests and enforcing rules and policies set by an individual company, no matter whether the prompt is being sent on a website, in an app, or by an agent.
The company has raised over half a million dollars from notable investors including Cultivation Capital, Missouri Technology Corporation, Move Ventures Capital, and the chief technology officer of Cloudflare, according to coverage earlier this year from the St. Louis Business Journal.
Eversole says Standpipe comes with a monthly cost of $5 per endpoint (any device, like a phone, laptop, or virtual machine that connects to a computer network and can send or receive data) for a version of the software that can allow or block specific pieces of sensitive information. There’s also a beefier version of it for $12 per endpoint per month that regulates how coding agents interface locally stored files, as those agents can be much more invasive.
Since OpsCompanion’s product is oriented as a paid service, Eversole says it inoculates the company from concerns over data privacy that its clients might have. “Our financial interest is deeply aligned with your personal protection, sovereignty, and security,” he says.
In July, OpsCompanion onboarded its first two companies onto the tool, with more following this month (Eversole declines to name them, citing concerns for their security and privacy). He says the rollout has been going “really well,” with clients in the financial space relieved to have a solution to what had been a glaring security hole.
And the company is getting sizable interest from other enterprises that want to use AI but also want to keep themselves protected from potential data breaches. “We have [chief technology officers] of extremely big companies here in town being like, ‘You need to continue this right now,’” Eversole says.
Standpoint is actually the result of a pivot Eversole says that OpsCompanion made earlier this year after its original idea of providing AI-powered site reliability engineers wasn’t finding much traction among local companies. He says he kept hearing the sentiment that the AI rollout at larger enterprises focused on health care, finance, or law was moving at a glacial pace.
“Honestly for about 30 days I was confused because I just came from big tech. Like, I thought everyone was super into AI. I use it every day,” Eversole says. “We’re a team of three or four people [and] we’re able to build faster than we ever could imagine. We would probably have to be 20 people five years ago. Why aren’t you using it?”
Before launching his own company, Eversole was a software engineer with Boeing, then joined Balto right before the pandemic supercharged growth for the local AI startup. He later worked as a reliability and systems engineer for the internet infrastructure and website security company Cloudflare.
When he stepped back to listen to what was holding local firms back from using AI more widely, he heard concerns over compliance and cybersecurity if sensitive information escaped secure servers.
“What happens with patient data? Customer data? Client-lawyer data?” he says. “The risk is not worth the reward even though their boards and CEOs are like, ‘We must adopt AI at all cost.’ The practical reality is people in the org being like, ‘Yes, but no.’”
Indeed, IBM pegs the cost of an AI-related data breach running into the millions of dollars, blaming poor AI governance and implementation as potential culprits. Eversole says OpsCompanion offers a solution that could spare companies that fate for a fraction of the cost, even if they have thousands of endpoints to manage.
Eversole sees Standpipe as just the starting point for OpsCompanion, as he envisions the company blossoming into a Midwestern tech company focused on keeping humans in control of AI. While the company’s offerings today focus on identifying and regulating how AI is used on a computer, he sees OpsCompanion growing into a platform that determines what an AI agent can access and how it may use that information.
“This is the Wild West, and we just need to put some fences up,” Eversole says. “Putting [this] on your laptop allows me to save on costs from the business angle, like you pay for the compute. The flip side of that is because I’m on your laptop, I can see the prompt at the most critical, vulnerable moment before it leaves a device, and we can do an enormous amount with that.”
Eversole is also excited about the prospect of developing specialized AI systems that run entirely on a local machine, where OpsCompanion could function as a type of clearing house. He says locally run AI models are the way of the future, which puts him at odds with the current fervor around the need for many more massive data centers.
“This hyperscale buildout is devoid of logic,” he says. “We’re watching this massive buildout purely because of hype.”
He points to companies like AMD, NVIDIA, and Apple all wanting to make local AI models easier to run on their hardware, which individuals can then purchase.
“I have yet to hear a super technical person that I deeply respect say, ‘Yeah, we’re not going to [move to] local models,’” Eversole says. “They all go, ‘Yeah, it’s definitely going to happen. We’re just not there yet.’”
