ReplyFabric Is Ready for the Next EU AI Act Milestone on August 2

ReplyFabric is ready. So are our customers.

On August 2, the European AI Act reaches its next important milestone. As more parts of the regulation become applicable, many organisations are asking themselves the same question:

Do we need to change the way we use AI?

If you're using ReplyFabric, the answer is simple:

No.

We've been following the AI Act since the first proposals were published and have continuously aligned our platform with its principles. Responsible AI has never been an afterthought for us. It has been part of ReplyFabric's design from day one.

Our customers shouldn't have to become experts in AI regulation. That's our job.

AI should help people, not replace them

From the beginning, ReplyFabric has been built around one core idea: AI should remove repetitive work while people remain in control.

We call this Human Premium.

When an email arrives in a shared inbox, ReplyFabric can categorise it, extract relevant information, prepare a reply draft or create a summary. The AI does the repetitive work, but the employee remains responsible for the communication and the business decision.

That philosophy fits naturally with one of the AI Act's central principles: meaningful human oversight.

AI assistance is not the same as AI decision-making

The AI Act distinguishes between systems that assist people and systems that make decisions affecting people.

ReplyFabric belongs firmly in the first category.

Our AI can:

• categorise incoming emails;
• draft replies;
• summarise long conversations;
• extract structured information from emails and documents;
• verify whether required information is present;
• route work to the appropriate colleague.

What it does not do is decide whether a customer should receive a refund, whether a supplier should be approved or whether a job applicant should be hired.

Those decisions always remain with your organisation.

Information extraction instead of evaluation

A good example is CV processing.

ReplyFabric can extract objective information such as education, languages, certifications and work experience from a CV. It can convert unstructured documents into structured information, making it easier for recruiters to review applications.

What it deliberately avoids is evaluating candidates or deciding whether someone is a good fit for a role. That distinction is important, both from a business perspective and under the AI Act.

The same principle applies throughout the platform: we extract information, we don't replace human judgement.

Smarter workflows without autonomous decisions

One feature our customers particularly like is the checklist function.

Suppose a customer submits a warranty claim but forgets to include the invoice number or photos of the damaged product. ReplyFabric compares the email against your predefined checklist, identifies which information appears to be missing and can automatically send a polite request using your approved template.

The AI isn't deciding whether the claim is valid. It's simply helping complete the file so your team can process it more efficiently.

That's exactly the type of workflow automation ReplyFabric was built for.

Transparency where it matters

Some people assume the AI Act means every AI-assisted email will need a disclaimer saying it was written by AI.

In practice, that's not how modern AI assistants work.

When someone uses Gmail's "Help me write" or Microsoft Copilot to draft an email, the sender remains the author of the message. ReplyFabric follows the same principle. Employees review AI-generated drafts before they are sent, so the communication remains theirs.

Internally, our platform clearly indicates when content has been generated by AI. Externally, your customers simply receive a response from your team.

Compliance is more than AI

The AI Act doesn't exist in isolation. Responsible AI also depends on strong governance, security and operational controls.

That's why ReplyFabric combines AI capabilities with an enterprise-grade compliance foundation, including:

• ISO 27001 certified information security;
• SOC 2 audited controls;
• GDPR compliance;
• Human-in-the-Loop workflows;
• audit logging;
• prompt injection protection;
• controlled knowledge sources;
• role-based permissions.

These measures not only support AI Act readiness but also help organisations deploy AI responsibly across their operations.

We keep watching, so you don't have to

The AI Act will continue to evolve over the coming years, and so will ReplyFabric.

We'll keep monitoring new guidance, update the platform where needed and ensure our customers stay aligned with changing requirements. That means you can continue focusing on your business instead of interpreting legislation.

Compliance shouldn't become another project on your to-do list.

It should simply be part of the product you use every day.

At ReplyFabric, that's exactly how we see it.