The most interesting part of Dezion’s pitch is the part it doesn’t have to invent. For nearly two decades, the company has been building interfaces for complex systems. That’s the foundation it’s now using to sell something far more ambitious: an execution control plane for AI. The bet is that the hardest part of governing automated decisions isn’t the policy logic, but the interface that makes it visible, trustworthy, and safe for the people who have to sign off on it.
According to its primary domain, Dezion is now positioning itself as “the execution control plane for AI systems, ensuring secure, policy-compliant, and trustworthy automated decisions” [dezion.in, 2024]. This is a significant pivot from its established identity as a UX/UI design agency, which its LinkedIn profile still describes as creating “premium mobile apps, web applications, and interfaces for complex systems” [LinkedIn, 2024]. The core of the new product appears to be a software layer that sits atop AI workflows, designed to enforce guardrails and provide audit trails.
The Wedge: Interface as Governance
The move from services to product is a well-trodden but perilous path. Dezion’s apparent wedge is its legacy. A design studio that has spent years mapping user journeys through intricate B2B software has a built-in intuition for where oversight breaks down. The control plane product seems to be an encapsulation of that knowledge, a set of visual tools and policy surfaces that make AI’s black-box decisions legible to human operators. The procurement argument would be straightforward: you already trust us to design the system, now let us sell you the software that keeps it in check.
An Early-Stage Reality Check
Public information on Dezion’s new direction is sparse, which points to a very early-stage or quietly bootstrapped operation. There is no record of institutional funding, named customers, or a leadership team beyond founder Tomer Fierstein [LinkedIn, 2024]. The company appears to be a solo endeavor leveraging its design agency’s reputation as a proof-of-concept lab. This brings a set of natural constraints:
- Go-to-market friction. A services background provides deep customer insight but doesn’t automatically translate into scalable sales pipelines.
- Product scope. An “AI execution control plane” is a broad category that overlaps with established observability, security, and compliance platforms.
- Resource intensity. Building and selling a platform in this category requires significant engineering and sales investment.
By targeting the specific interface and policy layer problems it already understands from its design work, Dezion could avoid a head-on collision with larger infrastructure players and instead sell a designer-built overlay. For now, the ideal customer profile is likely a mid-market company that has already engaged Dezion’s design studio for other projects and is now experimenting with internal AI automation.