EU AI Act: a guide for Juro customers

If you’re operating in the EEA and planning to use AI in your contract workflow, you need to understand how the EU AI Act (Regulation 2024/1689) impacts your technology choices. The Act creates obligations for any organisation deploying AI systems to end users in the EEA, regardless of where you are headquartered. This page explains what the Act means for Juro customers and how we have designed our services to enable you to comply with the Act.

Juro’s classification under the Act

The EU AI Act applies different requirements depending on  the system, and your role in the AI ecosystem. Juro is a provider of an AI system under the Act. Juro does not produce its own general purpose AI models, and the platform is not intended for any high risk use case (see more below). This means that Juro’s obligations under the Act are limited to those transparency obligations applicable to providers of low risk AI systems, as required by Article 50 of the Act.

What Juro does not do

Juro does not engage in any prohibited AI practice, as defined in the Act.

Juro does not produce its own general purpose AI models.

Juro is not intended to be used for any safety purpose, or for any of the high risk purposes described in Annex III to the Act.

How Juro meets its transparency obligations

Juro’s primary obligation under the Act is transparency (Article 50). That means that we are required to ensure that our users are informed when they are interacting with an AI system (or it must be obvious that this is happening). Here’s how we do that:

1. Our AI features are conspicuously labelled, and identifiable through distinctive colours and labelling (e.g. “Extract Data with AI” for AI Extract).
2. Any content generated by our platform is marked in a machine-readable format behind the scenes so it can be identified as artificially generated.

How we choose the models that power Juro

We do not build foundational models. Instead, we test, and integrate best-in-class models provided by market-leading model developers. Currently, we use:

• Claude (Anthropic, via AWS Bedrock);
• Gemini (Google, via GCP); and
• GPT (OpenAI, via AWS Bedrock and - for AI Assistant and AI Extract - via Microsoft Azure).

Juro continuously reviews the foundational models used in its services to ensure best-in-class performance.

The shared responsibility model

Juro sources models produced and infrastructure provided by market-leading, reputable providers. We deploy those models in a way that is optimised for our legal use case, and we constantly iterate to improve performance, accuracy and usability.

As the user of the system, you have responsibilities too. You are accountable for the accuracy and reliability of the AI outputs you generate in Juro, and you must check them and ensure they are suitable for your use before you rely on them.

You are also responsible for complying with any relevant regulatory or professional conduct rules applicable to you or your business. You can also improve the performance of our AI features for you and your business over time by optimising how you use them (e.g. by updating underlying prompts) and by providing feedback to Juro.

Our commitment

Juro is committed to being a transparent and accountable AI partner. As the regulatory landscape evolves, we will update this page and communicate any material changes that affect your use of the platform. Our goal is simple: to make sure that adopting AI-powered contracting with Juro is something you can do with confidence.