The battle for enterprise AI infrastructure just got a new contender betting that code generation should replace code writing.
The Summary
- Kore.ai launched Artemis, a platform that uses AI to build, govern, and optimize other AI agents — compressing months of engineering work into days
- At its core sits Agent Blueprint Language (ABL), a YAML-based declarative language that standardizes how enterprises define and govern AI agents
- The platform's thesis: AI, not human developers, should do most of the heavy lifting in agent creation
- Kore.ai is betting on neutrality and an intermediary language while Microsoft, Salesforce, Google, and ServiceNow race to own the enterprise agent stack
The Signal
Kore.ai's Artemis platform represents a ground-up rebuild of the company's core technology, arriving as every major enterprise software vendor fights to become the default infrastructure layer for AI agents. The company's founder and CEO Raj Koneru framed the release around a simple idea: "you do AI with AI — you design with AI, you build with AI, test with AI, deploy with AI, manage with AI, and optimize with AI."
What makes this more than vaporware is Agent Blueprint Language. ABL is a compiled, declarative language built on YAML that acts as an intermediary layer for defining, validating, and governing AI agents. Think of it as infrastructure for infrastructure: a standardized way to describe what an agent should do, how it should behave, and what guardrails it needs to operate within.
"Agent Blueprint Language standardizes how AI agents, workflows, and multi-agent systems are defined, validated, and governed."
The timing matters. Microsoft is embedding agents into every Copilot surface. Salesforce is building Agentforce as a new revenue engine. Google and ServiceNow are making similar plays. The market is fragmenting before it has fully formed. Kore.ai's answer is neutrality: build once, deploy anywhere, let AI handle the translation layer.
Key platform bets:
- AI-assisted design and build cycles instead of manual coding
- YAML as the common language for agent definition
- Governance and optimization built into the platform, not bolted on after
- Vendor neutrality in a market rapidly consolidating into walled gardens
The compress-months-into-days claim is where this gets interesting. If ABL actually works as an abstraction layer, it could mean enterprises stop hiring teams to hand-code agents and start prompting systems to generate them. That is not incremental efficiency. That is a different job market.
The Implication
Watch whether enterprises adopt ABL or treat it as yet another proprietary standard in a field already drowning in them. The real test is not whether Kore.ai can build AI that builds AI. The test is whether ABL becomes something other vendors support, or whether it dies in obscurity while Microsoft and Salesforce simply bundle agent-building into tools people already use.
For anyone building in this space: the race is no longer about who has the best LLM. It is about who controls the layer where agents get defined, deployed, and governed. If you are still hand-coding agents, you are already behind.