From Dublin to Vegas: Jentic Joins AWS GAIA 2025

We started the year by raising one of the largest pre-seed rounds in Irish software history, and we'll finish it on the main stage at AWS re:Invent in Las Vegas.

Jentic has been selected as the first Irish company to join Amazon Web Services’ Generative AI Accelerator (GAIA), one of 40 startups selected worldwide in 2025. GAIA provides $1 million in AWS credits, close collaboration with AWS engineering and customer teams, access to develop strategic alignment with senior AWS leadership, and a showcase at re:Invent (the main stage of enterprise technology).

AgentCore + Jentic + GAIA

GAIA helps the best AI startups move from prototype to production at enterprise scale. AWS already delivers a secure and scalable agent runtime through Bedrock and AgentCore. Jentic complements this by bridging those agents into the complex world of enterprise API and data systems - building upon existing infrastructure to make agents dependable, governable and ready for production.

With Jentic, enterprises can make their existing APIs and data catalogs discoverable to agents, and with access managed through our secure gateway to ensure full auth, observability and governance. Jentic identifies reusable business logic in agent execution traces, automatically turning them into reliable reusable workflows, driving a reinforcement loop that makes agents fast and dependable in production.

All of this also runs in a sandbox mode, allowing teams to safely evaluate agents, validate workflows, and establish a clear, low-risk path from experimentation to production ROI.

The Enterprise Problem We Solve

Most enterprise agent projects never make it beyond the demo. According to MIT’s “State of AI in Business 2025”, only about five percent of enterprise AI pilots are reaching production. The report cites “brittle workflows, lack of contextual learning, and misalignment with day-to-day operations” as the primary causes.

The reasons are familiar. APIs are messy but high-stakes. Each has its own authentication rituals, rate limits, and compliance constraints. The MIT study observed that enterprise systems “stall due to integration complexity and lack of fit with existing workflows,” and that most tools “don’t integrate well into workflows” at all. In other words, the problem isn’t the AI, it’s the plumbing.

Enterprises also need safe places to experiment without breaking production. MIT found that successful adopters start small, “embedding in non-critical or adjacent processes … then scaling into core workflows.”

And for CIOs and CISOs, governance isn’t optional. This is about customer data, financial data, patient data, employee data - good stewardship is non-negotiable. Executives emphasized “clear data boundaries,” “a vendor we trust,” and “the ability to improve over time” as the decisive factors in choosing AI partners.

That’s exactly the problem we built Jentic to solve.

What Jentic Does Differently

Jentic takes a different approach to bringing agents into production. Instead of asking enterprises to rebuild their systems around AI, we meet their existing infrastructure where it is, respecting every authentication rule, every audit policy, and every compliance boundary along the way.

At its core, Jentic makes existing APIs and data sources safely usable by agents. It combines five tightly connected architectural blocks:

• Ingest and Enrich APIs: Jentic ingests enterprise API and data catalogs, enriching them with metadata and authentication context to make them agent-readable.
• Access-Aware Discoverability: Agents search and discover available APIs and datasets, but only those they’re authorized to see.
• Managed Access: Every request passes through Jentic’s secure gateway, enforcing authentication and authorization, logging all actions, and automatically maintaining a live register of all agents and their activity.
• Workflow Generation and Orchestration: Successful execution traces are captured using Arazzo Engine and converted into reusable, composable workflows that can be orchestrated by any agent.
• Sandbox: A simulated environment mirrors internal APIs and data systems, letting teams test agents safely, identify high-value opportunities, and promote only validated workflows to production.

Together, these layers form a unified control plane that lets enterprises unlock the potential of agentic AI without compromising security, compliance, or reliability. This is a critically needed on-ramp for moving from proof-of-concept to production… a strategy to ensure you are in the 5% of enterprises whose AI pilots succeed.

Ireland’s Place in AI

Ireland has a long history of punching above its weight in software infrastructure. From the pioneering days of IONA Technologies, a forerunner of CORBA and one of the biggest pure-software companies of the 1990s, to the generations of distributed systems and middleware companies that followed, Irish engineers have quietly shaped the backbone of the internet. The same DNA that produced world-class middleware is now working at a new frontier: the enterprise infrastructure layer for AI.

Our team’s experience reflects that legacy. We’ve spent decades building and scaling systems from Ireland that are used by hundreds of millions of users worldwide. Ireland is now a concentrated melting pot, a first or second home to countless startups and tech giants, with a close proximity that delivers an edge when it comes to B2B and enterprise infrastructure.

We’re also fortunate to have Dr. Patricia Scanlon, Ireland’s AI Ambassador and chair of the Irish Government’s AI Advisory Council, as an advisor. As she put it:

“Jentic being accepted onto Amazon’s global AI accelerator program is further evidence that Irish AI startups can stand shoulder-to-shoulder with innovators anywhere in the world, and that Ireland’s unique concentration of expertise and global technology make it fertile ground for fast-scaling domestic companies.”

From middleware to the agentic era, Ireland continues to produce the connective tissue of global technology. We are proud to carry that lineage forward, building the bridge between AI and the APIs that power the modern world.

What’s Next

GAIA begins in Seattle next week, leading to a showcase at re:Invent in Las Vegas in December. We will demonstrate how sandbox-based simulation and deterministic workflows provide an immediate path to production-grade AI agents. With AWS AgentCore and Jentic together, enterprises have a clear, governable path to put AI agents into production. AWS GAIA Program 2025

Gratitude

We’re deeply grateful to the AWS teams in Ireland and the UK, and especially to Leon, Liliana, and Marco, for their early support and belief in what we’re building.

---

Press Assets

Read the full press release and access high-res images here.