Agentic compliance for AI Companies

Build your AI compliance, with every agent on the record.

Teddy builds your AI compliance step by step: a register of every AI system and agent, the EU AI Act classification for each one, ISO 42001 set up from scratch or on an ISMS you already run, and monitoring of your agents in production.

One registerAll AI systems and agents, classified
ISO 42001Prepared step by step
Per systemRisk level and duties
The AI company’s problem

You ship agents every week.
Your AI compliance is still a spreadsheet.

The EU AI Act asks what each AI system does and how risky it is. ISO 42001 asks for a management system around all of it. Meanwhile engineering ships new agents faster than anyone can document them.

01Risk depends on the use

Teddy classifies each AI system by how it is used.

02An AI management system

Teddy sets up policy, roles and impact assessments for ISO 42001.

03Agents need an owner

Each agent gets an owner, limits and a kill switch.

04Engineering ships, Legal asks, nobody knows what is running.

The inventory is a spreadsheet from May, new agents go live without an owner, and a chatbot can only explain the law in general. With Teddy, you build AI compliance step by step: every system registered and classified, ISO 42001 set up from scratch or on an existing ISMS, agents monitored in production.

How it works

Build AI compliance in five steps

Teddy keeps your AI register current, applies the right obligations to each system and builds ISO 42001 on the ISMS you already have.

01 Evidence Agent

Discover and register every AI system

Teddy finds models, agents and AI features in your cloud and code, and keeps the register current as engineering ships.

AI registerUpdated today
Support chatbot · customer-facingRegistered
refund-bot · autonomous agentNew
CV screening model · HRNew
02 Gap Audit Agent

Classified under the AI Act, with reasoning

Each system is assessed for its risk class and duties: prohibited practices, high-risk use cases, transparency and general-purpose AI obligations.

AI Act classification4 systems
CV screening · Annex IIIHigh risk
Support chatbot · Art. 50Transparency
Code assistantMinimal
03 Policy Agent

ISO 42001 on the ISMS you already run

AI policies, roles, impact assessments and reviews are added to your existing control set. Everything ISO 27001 already covers counts.

ISO 42001Built on ISO 27001
Covered by existing controls61%
AI policy and AI impact assessmentDrafted
Roles for AI system ownersReview
04 Teddy

Agents monitored in production

Every production agent has an owner, permissions and a kill switch. Teddy flags actions outside policy and logs every intervention.

Agent monitoringLive
refund-bot · 2 refunds above limitPaused
Sales email agent · human approval 98%Review
Kill switches tested this monthAll
05 Questionnaire Agent

Proof for customers and auditors

AI governance questions in customer reviews are answered from the register, and the evidence is ready for your ISO 42001 audit.

Customer reviewAI section
Do you use customer data to train models?Answered
How do you oversee autonomous agents?Answered
Teddy: The answers link to your AI policy, the register and last month’s monitoring report.
Never stuck

Ask Teddy. Agents do it. People back you up.

Ask what applies to a new feature before you ship it. Teddy answers from your register and the law. For conformity assessments, our compliance engineers are at your side.

1

Ask TeddyIs this feature high risk? What do we need before launch? Answers before the release, not after.

2

Agents do the workRegister, classifications, policies and monitoring reports kept current automatically.

3

Our team backs you upCompliance engineers help with conformity assessments and your ISO 42001 audit.

TeddyAcme Inc. · AI governance
We want to let the support agent issue refunds. What changes?
The risk class stays the same, but the agent gets access to payments. You need:
  • A named owner and a refund limitMissing
  • Human approval above the limitMissing
  • Logging and a kill switchIn place
Shall I add the controls and update the register before release?
Add controlsAsk our team
CEYour compliance engineer supports conformity assessments and audits.
FAQ

Questions AI companies ask

Do we fall under the EU AI Act?

Almost every company that builds or uses AI in the EU has some duties. Which ones depends on each system’s use case. Teddy assesses each system and explains the result.

How does ISO 42001 relate to ISO 27001?

ISO 42001 adds an AI management system on top of your ISMS. Teddy maps it onto your existing controls, so you only build what is missing.

How does Teddy monitor agents?

Teddy connects read-only to your cloud and logs, checks agent actions against the policies and limits you define and flags anything outside them.

Where do we start if we have nothing yet?

With the register. Teddy discovers what you run, classifies it and builds the rest step by step, starting with what the AI Act requires first.

Ship agents. Keep the record straight.

See your AI register built in the demo.

What changes with Teddy

From a spreadsheet inventory to a live AI register

TypicalMonths behind

The AI inventory is a spreadsheet, and new agents go live before anyone documents them.

With TeddyWithin weeks

Every AI system and agent registered and classified under the AI Act, with ISO 42001 built on your existing ISMS.

Why Teddy

Teddy was built by former CISOs and GRC managers, and one lesson carries straight over to AI: you cannot govern what you have not listed. The register comes first, and it has to keep up with what engineering ships.

Meet the founders →