From sandbox to scale: Jentic powers agentic integration with Amazon Bedrock AgentCore

For enterprises across industries, agentic AI is rapidly moving beyond boardroom discussion to deployment. MIT research suggests that adoption has already reached 35 percent, while McKinsey reports 62 percent of organizations are actively experimenting with the technology. But for AI agents to deliver meaningful value, they must be able to securely connect to enterprise systems, interact with operational software, and execute tasks within real-world business processes.

Jentic enables enterprises to connect AI agents to their bespoke systems and build deterministic workflows with out-of-the-box security and compliance. The company provides the missing layer between APIs and AI, helping businesses simulate and test agentic capabilities before deploying them to production environments. Most recently, Jentic expanded that approach with Jentic Mini, a free, open source, self-hosted offering designed to help developers connect AI agents to real systems more safely. The team is collaborating closely with Amazon Web Services (AWS), and utilizing scalable services like Amazon Bedrock AgentCore to help enterprises prepare for a future where operating fleets of AI agents is the norm.

Connecting AI agents to the systems that run enterprises

Founded in 2024 by three Dublin-based entrepreneurs, Jentic helps enterprises safely integrate and scale agentic AI. “Every enterprise has its own bespoke systems that actually run the business,” says Sean Blanchfield, CEO. “To gain an AI advantage, they have to build agents that safely connect to those systems. That’s a hard integration challenge to navigate, and it’s even harder to do it safely, with security, fine-grained access control, observability, and governance.”

“Enterprises trying to figure out how to deploy AI are often stuck because they only have a production environment, and they cannot deploy AI straight into production,” says Blanchfield. “It could be health data, it could be bank accounts, it could be insurance—whatever it is, you can't just connect a large language model (LLM) and let it start updating records.” “The first thing we set up for customers is a simulation sandbox, a digital twin of their existing infrastructure. It acts as a 24/7 training area for agents,” says Blanchfield.

Businesses can use the sandbox to test agents and develop deterministic workflows that can later be used in production environments. These workflows are repeatable and give businesses control over how AI behaves and interacts with critical systems. “It’s super-fast because it’s deterministic, and it’s explainable, which is very important from a regulatory perspective,” says Blanchfield.

Crucially, Jentic equips enterprises with a unified, scalable foundation that helps them think beyond small pilot projects and prepare for operating large fleets of agents across their systems. “Having three pilots that work is fine, but our customers are already planning for a future where they could potentially manage thousands of agents,” says Blanchfield. To support those ambitions, Jentic is working closely with AWS to build an enterprise-grade, secure, and globally scalable platform.

From experimentation to execution at AI speed

Jentic uses Amazon Bedrock to simplify how it builds, deploys, and manages AI at scale with access to high-performing foundation models from multiple providers. “The fact that we can have access to all these models in the same place as our infrastructure, simplifies compliance and removes the need to manage multiple providers,” says Cordner. It also makes it easy to evaluate and switch between different models depending on the task being performed. For example, certain workloads may require more advanced reasoning capabilities, while others benefit from faster or more cost-efficient inference.

Jentic uses Amazon Bedrock AgentCore, an agentic platform for building, deploying, and operating effective agents securely at scale, as the runtime layer within its sandbox environment. With features like persistent memory, complete session isolation, and agent performance tracking, Amazon Bedrock AgentCore allows Jentic to build, deploy, and monitor agents across complex, multi-step workflows. “We’re one of the first companies to deploy agents directly on AgentCore,” says Cordner. “It gives us things like session-scoped execution and persistent state management, so we don’t need to build those ourselves.”

Jentic recently launched Jentic Mini, a free, open source, self-hosted offering that gives developers and teams a practical way to connect universal agents to real systems more safely. It provides a controlled execution layer between agents and external APIs, with fine-grained permissions, runtime credential injection, and a single killswitch to revoke agent access instantly, while enabling access to a catalog of more than 10,000 APIs. For Jentic, it is a natural extension of the company’s broader platform vision: helping organizations move from experimentation to real-world agent deployment with stronger operational control.

Behind the scenes, Jentic relies on a broad set of AWS services to run its platform infrastructure including Amazon Elastic Container Service (Amazon ECS), AWS Fargate, and Amazon Simple Storage Service (Amazon S3). Together, these services allow Jentic to run its platform within a single, unified environment, reducing operational complexity while maintaining the reliability and security required for enterprise deployments. “It’s not just models, it’s the infrastructure, the security, the compliance—the fact that we have everything in one place is really valuable,” says Cordner.

Turning technical innovation into commercial traction

Beyond technology, AWS has helped Jentic expand its commercial reach. “The AWS team in Dublin and in the UK have been very proactive in trying to introduce us to prospective customers too, which has been great,” says Dorothy Creaven, COO. “Working with AWS also helps us have more confidence walking into customer meetings because when they know we work with AWS, they know we’re working with a very reliable supplier.”

Jentic also participated in the AWS Generative AI Accelerator, a global eight-week hybrid program supporting generative and agentic AI startups. Participants can learn from a seasoned network of AWS experts, receive up to $1 million in AWS Credits, connect with a community of innovative startups, and take advantage of exclusive go-to-market opportunities. “The accelerator brought us closer to AWS technical expertise and the entire AWS ecosystem from the start,” says Creaven. “As it’s a global program, we were able to meet other startups in the same phase of growth as us, and the fact that we are on the accelerator also shows potential customers that we have the backing of a company like AWS.”

AWS events have also provided valuable exposure and opportunities to connect with customers and partners. “Those kinds of events are really valuable because you get to meet potential customers and partners who are all interested in the same problems. We did a panel at Cloud Day Dublin, which gave us exposure to a much bigger audience,” says Creaven. “We were also involved in the Gen AI Loft events, which were great for meeting other startups and people building in the same space. It’s helpful to be part of the ecosystem and to be in those conversations, because that’s where people who are actively exploring this technology are.”

Listing its offerings on AWS Marketplace has also helped Jentic streamline how it sells to customers. “Getting listed on Marketplace is a major first step and there was a lot of direct help given to us to make that happen fast,” says Blanchfield. “There’s a certain endorsement of quality that comes with that listing and there’s practical things like streamlining contracting and procurement. As a startup, being able to have a fast-track in place with all of our ideal customers around the world is really valuable.” The company is now working towards enabling one-click deployments of its services into its customers’ own virtual private clouds.

A platform built for agentic fleets running across the enterprise

As enterprises continue to explore agentic AI, the focus is shifting from experimentation to long-term operation. Jentic is building its platform with that future in mind, one where AI agents are embedded across most—if not all—core business systems. “What we’re really building towards is a world where enterprises are running large fleets of agents,” says Blanchfield. “That only works if you can operate and manage those agents in a controlled way.”

To support that transition, Jentic now offers a structured six-week onboarding program to help enterprises make the jump from agentic strategy to production. “We take them from not really knowing what to do with agents to having something working,” says Creaven. “We set up the sandbox, connect into their systems, and start building workflows. By the end of it, they have a working foundation  they can build on for production.”

Going forward, collaboration with AWS will continue to support Jentic’s vision and growth. “With AWS, we get to work with hard-working, super smart people at the top of their game,” says Blanchfield. By building its platform on scalable AI infrastructure, the company can support enterprise customers from the deployment of their first agent to their first agentic fleet. “AWS allows us to take an enterprise posture without necessarily having the resources of an enterprise,” says Michael Cordner, CTO. “It gives us the complete package.”

Ambitious startups like Jentic are finding success on AWS. Programs like AWS Activate have helped over 350,000 disruptors get up and running, attract investor interest, and turn their ideas into marketable products with direct funding in the form of AWS Credits. If you want to build boldly, scale without limits, and launch with confidence, apply today.