Kore.ai launched its Artemis platform in May 2026 with a ground-up redesign of how enterprises build and govern AI agents. At the centre of that redesign is Agent Blueprint Language (ABL), a compiled, declarative language that standardises how agents are defined, validated, and controlled before they ever reach production.
The problem it solves
define what an agent should do, how it should behave, or what happens when it hands off to another agent. Every team builds differently. Teams apply governance after the fact, if at all. The result is a fragmented landscape that grows harder to audit, scale, or trust.
ABL closes that gap.
What Agent Blueprint Language actually is
ABL is a compiled, declarative language built on YAML. It sits between a business objective and the production infrastructure where agents run. You describe what you need an agent to do. ABL translates that into something the platform can validate, compile, and govern before any agent goes live.
It ships with its own parser, compiler, and runtime. Six built-in orchestration patterns govern how multiple agents coordinate: supervisor, delegation, handoff, fan-out, escalation, and agent-to-agent federation. The platform logs every AI decision and workflow in real time.
Governance operates at the platform layer, outside the model itself. Agent behaviour does not depend on prompt engineering or model interpretation. ABL defines, compiles, and controls it directly.
Why this matters for CIOs
Enterprises spent decades standardising how they write code, store data, and secure networks. ABL applies that same logic to AI agents.
For a CIO managing AI projects across multiple departments, vendors, and systems, that standardisation matters. Today’s AI sprawl becomes tomorrow’s compliance risk. ABL gives teams a common foundation to build against, a consistent governance framework across the enterprise, and a way to consolidate fragmented AI systems into something auditable and manageable.
Kore.ai’s Arch feature extends this further: it translates plain-language business objectives directly into production-ready ABL, compressing delivery from months to days.
The bigger picture
Artemis runs natively on Microsoft Azure and connects with Microsoft Foundry, Microsoft Agent 365, Entra ID, and the Microsoft Graph API. For enterprises already on the Microsoft stack, ABL slots into a governed, identity-aware AI environment rather than running as a standalone experiment.
Enterprise-grade agentic AI means agents that are defined, validated, and controlled from day one, not just agents that work.
Contact Think Tank | Learn more about Kore.ai
