Jetty Raises $2M to Build the Safety Net for Agentic AI

MONTREAL, QC – May 06, 2026 – As enterprises race to deploy autonomous AI, a critical and costly gap has emerged between dazzling demonstrations and real-world reliability. Montreal-based startup Jetty Solutions Inc. has just secured over $2 million in pre-seed funding to tackle this problem head-on, building what it calls the essential infrastructure for reliable agentic AI.

The funding round, led by AQC Capital and Hidden Layers Capital, with participation from Mila Ventures and strategic angel investors from AI powerhouses like Google and Meta AI, signals a significant investment in solving one of the most pressing challenges in artificial intelligence today: its inherent fragility.

The 'Fragility' Crisis in Production AI

The promise of agentic AI—autonomous systems capable of executing complex, multi-step tasks—is vast. Yet for many organizations, the reality has been one of frustration. An AI agent that performs flawlessly in a controlled demo can fail unpredictably when faced with the messy, complex, and dynamic nature of a live production environment.

"Most AI systems today are still fragile - they work in isolation but break under real-world complexity," said Jonathan Lebensold, Founder and CEO of Jetty, in a statement. This fragility manifests in numerous ways that stall enterprise adoption. Industry experts point to a "reset to zero" problem, where a failure in one step of a long workflow forces the entire process to restart, wasting time and computational resources. Agents often struggle with integrating disparate software tools, handling API timeouts, or parsing inconsistent data formats, leading to cascading errors.

Furthermore, the cost of these failures can be immense, not just in terms of wasted resources but also in potential security vulnerabilities. As AI agents are granted access to production systems, a small conceptual error can be amplified into a significant system-level incident. This risk, combined with the difficulty of debugging intermittent and context-dependent failures, has made many business leaders hesitant to deploy autonomous AI for mission-critical tasks.

A Mission, Not Just a Prompt

Jetty's approach is to move beyond simply giving AI agents tasks and instead provide them with a robust framework for execution, self-assessment, and improvement. The company is building a foundational layer that combines three core components to create a more resilient and trustworthy AI ecosystem.

First, Jetty utilizes structured runbooks, which provide AI agents with a "mission, not just a prompt," according to Lebensold. This gives the AI a clear, structured workflow to follow, reducing ambiguity and providing a stable path for complex operations. Second, it provides isolated execution environments, or sandboxes, where the AI can run its tasks. This allows the agent to interact with tools and data in a secure, contained space, preventing any potential errors from impacting the broader production infrastructure.

The third and most crucial component is a loop of rigorous evaluation and continuous learning. Within the sandbox, Jetty's infrastructure enables an AI agent to run a workflow, assess its own performance against a defined quality bar, and identify its own mistakes. With a human in the loop for oversight and guidance, the system can then iterate on the process, learning from its failures until it can reliably achieve its mission. This creates a powerful feedback mechanism that allows AI systems to adapt and improve over time, something that has been critically absent in many agentic deployments.

Forged in Montreal's AI Powerhouse

The company's emergence and funding are a testament to the strength of Montreal's world-renowned AI ecosystem. The investment is led by local firms AQC Capital and Hidden Layers Capital, reflecting a deep-seated confidence in homegrown talent to solve global technology challenges.

The involvement of Mila, the Quebec Artificial Intelligence Institute, as a backer provides significant institutional validation. "Enterprises need a way to deploy their models with the kind of audit trail required to build trust with their customers," noted Alex Shee of Mila Ventures. "Jetty is solving this critical pain point to allow AI to be reliably deployed."

This connection to Montreal's academic and research community runs deep. Doina Precup, a luminary in the field who is a Professor at McGill University, a CIFAR AI Chair, and a research lead at DeepMind, has also endorsed the company's mission. "As AI systems become more autonomous, ensuring they behave reliably in complex environments becomes a central challenge," Precup stated. She sees Jetty's work as "critical infrastructure" that is essential for moving agentic AI into real-world applications. This endorsement is particularly resonant given that Jetty's founder, Jonathan Lebensold, is a PhD student under Professor Precup's supervision at McGill, grounding the company's commercial vision in cutting-edge academic research on AI safety and reliability.

A Veteran Team to Navigate the Frontier

While the technology is forward-looking, the team building it is grounded in proven experience. Jetty is helmed by a leadership trio with a track record of scaling both technology and businesses. Alongside Lebensold's deep AI expertise, the company boasts Roberto Cipriani, former CTO and COO of the massively successful ed-tech platform Paper, and Tracy Milner, a serial founder-CEO with an "inception to exit track record."

Cipriani’s experience at Paper, where he was instrumental in scaling the platform to serve millions of students, provides invaluable insight into building and managing robust, large-scale technology infrastructure. Milner’s operational and strategic acumen from building and successfully exiting previous ventures ensures that Jetty is not only developing a powerful product but is also building a sustainable business around it. This blend of academic rigor, technical depth, and seasoned operational leadership gives the pre-seed company a level of maturity that belies its early stage.

The new infusion of capital will be used to accelerate product development, expand the engineering team, and deepen engagement with enterprise customers who are eager to move their agentic AI initiatives from the lab to production. By focusing on the unglamorous but essential work of building guardrails, audit trails, and learning mechanisms, Jetty is positioning itself not as another tool to build AI agents, but as the foundational layer that will allow organizations to finally trust them.