HAIEC

HAIEC is an evidence-bound AI assurance platform that lets organizations scan, attack-test, and prove the safety of consequential AI systems before they act. It connects source, identity, access, policy, permissions, evidence, and observed behavior to produce an AI Action & Access Map. This map shows what an AI system or agent can reach, what credentials it uses, what actions are technically possible, and what controls apply. HAIEC is applicable to AI applications, AI systems, AI agents, model/API integrations, automated AI workflows, and consequential agentic systems. HAIEC addresses the structural difference between permission and delegation. It asks five evidence questions: intention, policy, credentials, capability, and observed. Missing evidence remains UNKNOWN and is never silently treated as a pass. The platform maps evidence to recognized frameworks and regulations including SOC 2 AI controls, ISO 42001, EU AI Act, NIST AI RMF, NYC Local Law 144, Colorado AI Act, and OWASP Top 10 for LLM, though framework mapping is not the same as assurance. The platform follows a Define, Connect, Collect, Assure, Verify, and Monitor lifecycle. It supports developers, teams, assurance firms, and enterprises, with founder-led POC engagement for enterprise. Monitoring is bounded to supported runtime evidence sources, and evaluated scope is currently structured at the contract level, with structured scope binding under development.

Category: Chatbots
Pricing: Free

Problem solved

Security controls establish whether an identity is permitted to perform an action but not whether the AI was delegated the consequential choice. HAIEC helps engineering, security, and assurance teams understand, document, and validate what AI systems can actually reach and do, using evidence rather than assumptions so dangerous capabilities are caught before deployment.

Use cases

Scan an AI application repository to discover the AI action surface and access findings
Map AI agent reach, credentials, technical actions, and potential consequences
Evaluate consequential AI systems across five evidence questions: intention, policy, credentials, capability, and observed
Generate audit-ready evidence mapping to frameworks like SOC 2 AI, ISO 42001, EU AI Act, NIST AI RMF, and OWASP Top 10 for LLM
Attack-test AI systems and agents before release
Add technical AI evidence to assurance firm workflows
Monitor AI runtime behavior where supported runtime sources are connected
Support enterprise evidence-bound assurance engagements for consequential AI

Audience

Developers
ML Engineers
Data Scientists
Product Managers
Enterprise

Pros

Clear AI Action & Access Map showing reach, credentials, actions, and consequences
Distinguishes permission from delegation to reveal structural gaps
Missing evidence is explicitly marked UNKNOWN rather than silently treated as a pass
Maps evidence to multiple recognized AI governance frameworks and regulations
Offers a free tier

Cons

Not all five evidence questions have native evidence producers today
Monitoring is bounded to supported runtime evidence sources
Evaluated scope is currently at the contract level; structured scope binding is under development
Framework mapping is not the same as formal assurance or a safety certification
Platform is relatively new and enterprise adoption may require founder-led POC engagement

Official website: https://haiec.com/

Canonical LambdaBase page: https://www.lambdabase.com/tools/haiec

← Directory

HAIEC

Free

★ 12 • Updated Sep 7, 2026, 03:19 AM

HAIEC is an evidence-bound AI assurance platform that lets organizations scan, attack-test, and prove the safety of consequential AI systems before they act.

Use Cases ▼

Scan an AI application repository to discover the AI action surface and access findings

Map AI agent reach, credentials, technical actions, and potential consequences

Evaluate consequential AI systems across five evidence questions: intention, policy, credentials, capability, and observed

Generate audit-ready evidence mapping to frameworks like SOC 2 AI, ISO 42001, EU AI Act, NIST AI RMF, and OWASP Top 10 for LLM

Attack-test AI systems and agents before release

Add technical AI evidence to assurance firm workflows

Monitor AI runtime behavior where supported runtime sources are connected

Support enterprise evidence-bound assurance engagements for consequential AI

What Problem It Solves ▼

Key Features ▼

✓

Clear AI Action & Access Map showing reach, credentials, actions, and consequences

✓

Distinguishes permission from delegation to reveal structural gaps

✓

Missing evidence is explicitly marked UNKNOWN rather than silently treated as a pass

✓

Maps evidence to multiple recognized AI governance frameworks and regulations

✓

Offers a free tier

Overview

Full Description ▼

HAIEC addresses the structural difference between permission and delegation. It asks five evidence questions: intention, policy, credentials, capability, and observed. Missing evidence remains UNKNOWN and is never silently treated as a pass. The platform maps evidence to recognized frameworks and regulations including SOC 2 AI controls, ISO 42001, EU AI Act, NIST AI RMF, NYC Local Law 144, Colorado AI Act, and OWASP Top 10 for LLM, though framework mapping is not the same as assurance.

The platform follows a Define, Connect, Collect, Assure, Verify, and Monitor lifecycle. It supports developers, teams, assurance firms, and enterprises, with founder-led POC engagement for enterprise. Monitoring is bounded to supported runtime evidence sources, and evaluated scope is currently structured at the contract level, with structured scope binding under development.