The EU AI Act makes the use of Artificial Intelligence more binding. From August 2, 2026, large parts of the regulation will become broadly relevant. For many organizations, this means: AI governance will no longer be just an internal IT issue but a topic for procurement, data protection, compliance, and management.
Swiss organizations should also take notice.
Not every Swiss company automatically falls directly under the EU AI Act. But those who work with EU customers, offer AI systems in the EU market, or align with international compliance standards will increasingly have to answer questions about AI governance.
These questions are especially important for AI meeting software. Because meetings often contain confidential information: customer data, personnel data, internal decisions, strategies, contracts, tasks, and assessments.
The central question is therefore not only:
Can the AI create a good protocol?
But also:
Is it clearly regulated how the AI is used, where data is processed, and who has access?
What is the EU AI Act about?
The EU AI Act is the European legal framework for Artificial Intelligence. It follows a risk-based approach. The higher the risk of an AI system to people, safety, or fundamental rights, the higher the requirements.
For many organizations, the following become particularly important:
AI systems must become more transparent
Risks must be better assessed
Responsibilities must be clearer
Human control remains central
Data quality and documentation become more important
Providers and users must better explain how AI is used
For procurement, this means: It will no longer be enough to compare only functions and price. Organizations will ask more precisely how an AI solution is operated, documented, and controlled.
Why does this affect AI meeting minutes?
Meeting minutes may seem harmless at first glance. In practice, however, they often contain very sensitive information.
Depending on usage, AI meeting software processes:
Audio recordings
Transcripts
Speaker recognition
Summaries
Tasks
Decisions
Personal statements
Confidential meeting content
This raises questions that are central to data protection and AI governance:
Where are audio and transcripts stored?
Where does the AI processing take place?
Are external AI providers used?
Are data used for AI training?
Who is allowed to see meeting content?
Are there roles and permissions?
Can data be deleted or exported?
Are there security documents and data processing agreements?
These points become more important in procurement. Especially for municipalities, schools, administrations, healthcare organizations, and companies with confidential customer data.
AI governance becomes part of procurement
Many providers today advertise with terms like GDPR, SOC 2, or HIPAA. This can be relevant but does not automatically answer all questions of a Swiss organization.
For Swiss customers, the following is often decisive:
Is there Swiss hosting?
Does processing take place in Switzerland?
Is there a solution without external AI providers?
How is Swiss German processed?
Is there a data processing agreement?
Are deletion, export, and access clearly regulated?
Can an enterprise or on-premise setup be audited?
Are there contact persons in Switzerland?
This is exactly where the Swiss context differs. For a municipality, school, or Swiss SME, it is not only important whether a provider mentions international certificates. It is also important whether the solution fits the own legal, organizational, and cultural environment.
How Meeting Metrics approaches the topic
Meeting Metrics was developed for organizations that want to use AI for meeting documentation without losing control over their data.
The approach is simple:
AI should document meetings more efficiently, but data protection, access, and processing must remain clearly regulated.
Meeting Metrics offers various levels for this.
1. Clear data processing
Meeting Metrics processes meeting content to create transcripts, summaries, tasks, decisions, and exports.
Important: Customer data is not used for training, retraining, fine-tuning, or improving own or third-party AI models.
This is especially relevant because meeting content is often confidential. Companies, schools, and municipalities use Meeting Metrics to better document their own conversations, not to provide their content for AI training.
2. Enterprise processing in Switzerland
For organizations with increased requirements, Meeting Metrics Enterprise offers processing entirely in Switzerland.
This includes:
Hosting in Switzerland
Data storage in Switzerland
Encrypted backups in Switzerland
Transcription in Switzerland
Speaker recognition in Switzerland
AI processing with local models in Switzerland
In the enterprise setup, meeting content does not have to be sent to external AI providers abroad for AI processing.
External AI providers such as OpenAI, Google Gemini, or Anthropic are not activated in the enterprise setup unless separately audited, contractually regulated, and explicitly approved.
3. Roles, rights, and controlled releases
AI governance is not just about technology. It is also about who is allowed to do what.
Meeting Metrics works with role- and team-based permissions. Users see only content they are authorized for. Meetings can be assigned to specific persons, teams, or explicit releases.
This is important because not every meeting is equally sensitive.
An internal project meeting has different requirements than:
a municipal council meeting
a school discussion
a management meeting
an HR conversation
a confidential customer appointment
a board meeting
Therefore, AI meeting software needs clear access options and controlled exports.
4. Export, deletion, and traceability
Organizations must be able to control meeting data throughout its lifecycle.
This includes:
Creation
Processing
Review
Approval
Export
Retention
Deletion
Meeting Metrics supports exports, deletion options, and defined operational and security processes. This makes AI meeting documentation not just a black box but a controllable process.
5. On-premise for particularly high requirements
For most organizations, Meeting Metrics Enterprise with full processing in Switzerland is the suitable solution.
However, there are cases where additional requirements exist. For example, when internal guidelines do not allow external SaaS processing or data flows must remain entirely within own infrastructure.
For such cases, Meeting Metrics can consider an on-premise or dedicated solution.
This is especially relevant for:
Larger public organizations
Regulated companies
Organizations with very strict IT requirements
Areas with particularly sensitive data
Customers who need their own operating model
Meeting Metrics thus offers both: an enterprise solution with Swiss processing and the possibility to consider an on-premise setup if needed.
Why the Swiss context matters
Many international AI tools are strong in functionality. But for Swiss organizations, other questions often decide on use:
Does the solution understand Swiss German?
Is there Swiss hosting?
Do the data remain in Switzerland?
Are there local contact persons?
Are data processing agreements, security documents, and subprocessors transparent?
Can the solution be pragmatically introduced for municipalities, schools, and SMEs?
Meeting Metrics can credibly occupy this context because the solution comes from Switzerland and is aligned with Swiss requirements.
This is especially relevant in the context of the EU AI Act. Because the more important AI governance becomes, the more it counts not only the best feature but the trustworthy overall package.
What organizations should check in AI meeting software
Those procuring AI meeting software should not only ask about price and functions.
Important procurement questions are:
Where are meeting data stored?
Where does AI processing take place?
Which AI providers are used?
Is there an enterprise configuration without external AI providers?
Are customer data used for AI training?
Are there role and rights concepts?
Can data be deleted and exported?
Is there a data processing agreement?
Are there security and data protection documents?
Is there an on-premise option?
Is the provider familiar with Swiss requirements?
These questions will become even more important due to the EU AI Act and increasing governance requirements.
Conclusion
The EU AI Act shows: AI is becoming more professional, regulated, and more accountable.
For companies, schools, municipalities, and administrations, this means: AI must not only be practical. It must also be comprehensible, controllable, and used responsibly.
This is especially crucial for meeting minutes because sensitive information is often processed there.
Meeting Metrics addresses this topic with a clear Swiss approach:
AI meeting documentation for everyday use
Enterprise processing entirely in Switzerland
Local AI models in the enterprise setup
No use of customer data for AI training
Roles, rights, exports, and deletion options
Security documents and data processing agreements on request
Optional on-premise or dedicated solution for increased requirements
This makes AI not only efficient but also procurement-ready.
And this is exactly what will become increasingly important in the coming years.
FAQ
What does the EU AI Act mean for AI meeting software?
The EU AI Act makes AI governance, transparency, risk assessment, and documentation more important. For AI meeting software, this means: Organizations should know where data is processed, which providers are involved, who has access, and how results are controlled.
Does the EU AI Act also affect Swiss companies?
Not every Swiss company is automatically directly affected. However, the EU AI Act can become relevant if a company offers AI systems in the EU market, serves EU customers, or aligns with European procurement and compliance requirements.
How does Meeting Metrics solve AI governance?
Meeting Metrics relies on clear data processing, role-based access, export and deletion options, no use of customer data for AI training, and an enterprise option with full processing in Switzerland.
Where is data processed in Meeting Metrics Enterprise?
In Meeting Metrics Enterprise, processing takes place entirely in Switzerland. This includes hosting, storage, backups, transcription, speaker recognition, and AI processing with local models.
Does Meeting Metrics use external AI providers?
In standard plans, external AI providers may be used depending on configuration. In the enterprise setup, external AI providers are not activated unless separately audited, contractually regulated, and explicitly approved.
Is there a Meeting Metrics on-premise option?
Yes. For organizations with particularly high requirements, Meeting Metrics can consider an on-premise or dedicated solution.
