What’s going on with your tech stack? Catio’s AI copilot knows

CTOs and chief architects have often considered the tech stack more of an art than a science, a realm of experience and tribal knowledge. 

But enterprise architectures are growing increasingly complex and opaque, requiring a whole new approach, according to Toufic Boubez, co-founder and CTO of Catio. As he put it, infrastructure building and management is being transformed into a data-driven science. 

Catio aims to be a leader in this transformation. The one-year-old startup, which is currently in closed beta, offers an AI copilot for tech stack architecture. Boubez and Co-Founder and CEO Boris Bogatin presented on their unique platform at the Innovation Showcase at this year’s VB Transform in San Francisco. 

Boubez described the company’s platform as “the next evolution of the stack.”

Providing a digital twin, or blueprint of the enterprise tech stack

Catio’s platform essentially acts as a digital twin, creating blueprint-like visualizations of an enterprise’s architecture. The system can plan new tech stacks and continuously evaluate architecture to serve as an “ongoing AI advisor,” Bogatin explained. 

This visualization provides strategic observability into subsystems incorporating thousands of points of context. AI sitting in the middle of the stack processes that data and presents it in a contextually relevant way said Bogatin. 

“If you think about Catio, it’s AI for CTO,” he said. 

The company is nearing commercial availability — that’s expected for the fall — and its eight-person team is working with larger startups and Fortune 100s. It has raised $4 million in pre-seed funding and expects another fundraise before its commercial release. 

“We’re kicking the tires on this aggressively with major design partners,” Bogatin explained. 

Going way beyond whiteboards

Enterprise leaders need to have observability into their systems so that they can make strategic decisions and investments, Bogatin noted. But tech stacks are becoming increasingly more complex as organizations integrate more and more SaaS and cybersecurity apps and components. 

Often, CTOs don’t know what their architecture is for, and probably most importantly, why they need to invest significant dollars into it, he pointed out. They treat tech stacks as a cost center, as they can’t directly tie investments and business outcomes. 

“They don’t know how to measure it,” said Bogatin. “They’re all looking for help. Many folks don’t want to be beholden to figure out the best architecture.”

This is because traditionally, enterprises have relied on high-priced architects and expensive consultancies to build their tech stacks, Bogatin pointed out. But often, those designs “live in people’s heads.”

“Human brains will synthesize that and put it on whiteboards; that was the most logical way to represent what was going on in their heads,” he said. 

But that hasn’t been easily translated, thus often making tech architectures opaque, or black boxes. “It’s not been very data-defined, very accessible,” said Bogatin. 

Boubez pointed out that, if you asked software developers 3 years ago whether they would use AI, they most likely would have responded that software development is a skill that AI couldn’t handle. But now, everyone uses GitHub Copilot, Tabnine or other AI platforms. 

“This is the same kind of approach and perspective,” he said of his company’s product. 

Relying on a sophisticated workflow of AI agents

Catio creates a “canonical view” of architecture requirements with dashboard-level analytics. A 24/7 AI advisor provides recommendations across the stack so that leaders know exactly what they’re dealing with (and what they need). 

“It’s finding exactly what you have in your tech stack,” said Bogatin. “You need to understand what you have, you need to be able to evaluate it, you need to have analytics.” 

Catio’s platform is built on a multi-agent system, what Boubez described as a “relatively complex, sophisticated workflow of AI agents that collaborate with each other to create design proposals.” 

This consists of a chief architect agent who communicates with retrieval agents and agents specialized in areas such as data, messaging and security. The AI lower down in the stack performs analysis and makes recommendations — and provides reasoning for those suggestions — then crafts design proposals that are filtered back up to the chief AI architect. That agent then analyzes proposals based on business context, summarizes them and presents them to human users. 

“This context-aware use of foundational models, and agentic use of foundation models is extremely powerful,” said Bogain. “This is like using scalpels to go into a foundational model to retrieve things that are super, super precise as opposed to asking a prompt.”

Making CTOs’ lives a lot easier

Catio is similar to SaaS in that users integrate their tech stack components and create policies around them. The platform then forms a codebase that codifies the entire architecture. Users can see what changed from snapshot to snapshot and gain a true understanding of their architecture. 

The view is “exactly the way it is at the point of production,” Bogatin explained, and users can see any information about any component, move, filter and rearrange them and analyze raw data to understand its parent and child relationships. They can also save various teams as views. 

Boubez pointed to a use case of a chief architect of a large biotech company overseeing several architecture and project teams. They needed to be able to have a view of each of those tramps and wrap their hands around everything being built. Catio allowed the architect to communicate on a common platform with teams to assess and discuss different recommendations, make plans and improve architectures. 

Jordan Rosen, ex-SVP and GM at Disney, explained that he ran the profit and loss (P&L) for the company’s over-the-top (OTT) video streaming platform utilized by HBO, WWE, Fox, Sony, Discovery, MLB, NHL and others 

“As an organization, it was very challenging getting quantifiable visibility on cost performance of our cloud operation,” he said. He noted that if Catio can provide dashboard-level visibility and recommend paths to improvement, “it would be a go-to tool to track and jointly evaluate roadmaps and performance.”

Satish Raghunath, VP of engineering at Salesforce, pointed out that infrastructure engineers continually grapple with a complex set of issues. 

“Some of the key pain points are in picking the appropriate software components and integrating them with the rest of the stack,” he said. He added that it is “heartening” to see Catio tackle these problems. “It can make the CTO’s life a lot easier.”