Developer Documentation – Kore.ai Agent Platform (Artemis)

Architecture Overview

The Kore.ai Agent Platform provides a unified runtime for both scripted (deterministic) and reasoning (LLM-driven) AI agents on the same infrastructure. The platform exposes APIs for agent definition, orchestration, and management, and abstracts the underlying LLM so agent logic is portable across model providers.

Key Platform Components

• AI Agent Builder: Low-code/no-code tooling for building, testing, and deploying agents.
• Intelligent Orchestrator: Coordinates multiple purpose-built agents in parallel with independent fault recovery.
• Agentic RAG Search: Connects agents to enterprise data sources for contextual, action-oriented retrieval.
• Admin Controls: Centralized governance with runtime-layer constraint enforcement (not prompt-level), complete audit logs, and compliance reporting.
• Kore.ai Marketplace: Pre-built agents, templates, and third-party integrations for rapid assembly.

Integration & API

• Connect to enterprise systems (CRM, ITSM, HRMS, ERP) via pre-built integrations from the Marketplace.
• APIs available for programmatic agent deployment, monitoring, and interaction logging.
• Supports LLM-agnostic deployment — swap or upgrade underlying models without rewriting agent definitions.

Compliance & Governance

• Constraints enforced at the runtime layer, not via AI instruction, ensuring zero compliance flags.
• 100% of interactions are logged with full audit trails suitable for regulators and board reporting.
• Industry-specific compliance: HIPAA (Healthcare), financial regulation support (Banking).