Univerx Trust Center
Responsible AI
Transparency, human oversight, fairness, privacy and governance for every AI module in Univerx: AI agents, Univerx Assist, conversation analytics, knowledge and Semantic². Principles that work as settings.
Building blocks
Responsible AI in detail.
The building blocks at a glance – all in the same workspace and under the same administration.
Transparency
AI identifies itself in dialogue; generated answers show their source.
Human oversight
Handover to a person at any time; proposals rather than decisions where it matters.
Logging
AI agent actions and accepted suggestions are recorded with time and outcome.
Approval and versions
Models, prompts and agents are tested, approved and versioned before they go live.
Checking fairness
Outcomes can be analysed by group to spot unequal treatment.
#Principles you can configure
Almost every vendor has AI principles. What matters is whether they reach the product. At Univerx every principle becomes a control you can see, set and check. The Trust Center describes the principles in detail under Responsible AI.
The controls apply to every AI module: AI agents, Univerx Assist, conversation analytics, knowledge and the Semantic² add-on.
#Five principles, five controls
| Principle | Control in Univerx |
|---|---|
| Transparency | Disclosure as AI in dialogue; sources shown for generated answers |
| Human oversight | Handover at any time; proposals rather than automatic decisions |
| Fairness | Outcomes analysed by group to spot unequal treatment |
| Privacy | Approval per function and queue; region and retention as for all data |
| Governance | Testing, approval and versioning of agents, prompts and models; logs |
#Disclosure and oversight day to day
An AI agent says at the start that it is an AI and repeats it when asked. The way to a person is always open. That reflects the transparency obligations in Art. 50 of the AI Act, which Univerx is designed for.
Where a decision has consequences for people, AI makes proposals, not decisions. Semantic² delivers findings that an officer accepts or rejects. Assist suggests replies that agents check.
Generated answers rest on your knowledge base and cite their source. If there is no source, the AI says so instead of inventing something. How that works is explained in the glossary under grounding.
#Approval before use
New agents, changed prompts and new model versions do not go live directly. In Agent Studio they are tested, approved by an authorised person and versioned. If a change does not prove itself, the previous version is restored.
Univerx Trust Center
Security and regulation as a design principle, not an appendix.
Security, privacy, resilience and responsible AI as a family of its own, plus the Europe layer from GDPR to the Accessibility Act. With settings you make yourself and documents on request.
Privacy
Your data stays your data: region choice within the EU, a sub-processor list, deletion and export, role-based rights and a data processing agreement under Art. 28 GDPR on request.
- DSGVO
- EU
Security
Encryption in transit and at rest, identity and access control, network and API security and secure operations. What the platform brings, and what you set yourself in administration.
- DSGVO
- BSI C5
- NIS2
Resilience
Multi-zone availability, independent services, continuity plans and a planned status page: Univerx is designed to keep your customer service running when part of the infrastructure fails.
- NIS2
- DORA
- EU
Europe & regulation
The Univerx Europe layer: GDPR, EU data residency, the EU AI Act, BSI C5, eIDAS, NIS2, DORA and the Accessibility Act. For each framework the honest status, what “designed for” means and which documents exist.
- DSGVO
- EU
- AI Act
- BSI C5
- eIDAS
- NIS2
- DORA
- BFSG
Frequently asked questions
Does Univerx meet the requirements of the EU AI Act?
Univerx is designed for the requirements of the EU AI Act: disclosure, oversight, logging, approval. Whether your specific use falls into a risk class we clarify with you; the product does not replace a legal assessment.
Can we switch off individual AI functions?
Yes. Every AI function can be enabled or disabled per queue, team or channel.
Who is responsible if the AI makes a mistake?
Responsibility stays where it was: you decide what AI may do, Univerx supplies controls and logs. That is why approvals and limits are not a side issue but part of every rollout.
Related
ProductsAI Agents
AI agents understand the request, act in connected systems and hand over to people with context. On the phone as a voicebot, in chat as a text agent, always disclosed as AI.
ProductsAgent Studio
The workbench for AI agents and prompts: build without code, test with real examples, approve through roles and version every change. Built once, usable in every channel.
ProductsUniverx Assist
AI assistance in the conversation: suggested replies, relevant knowledge, next steps and automatic summaries for agents. For team leads, an overview of topics and coaching needs.
Use casesSelf-service around the clock
AI agents resolve routine requests on the phone and in chat, at any hour. They are disclosed as AI, act only in approved systems and hand over with context as soon as a person is needed.
Use casesBack-office automation
Steer back-office work the way you steer customer contacts: task routing with SLAs, workbins and tracking. Semantic² checks the documents and returns findings in euros with their source. A person always makes the decision.
IndustriesPublic sector
The complete platform for citizen service: every channel in one queue, public service numbers and hotlines with AI agents, appointment booking, video, secure messaging and translation. Plus the digital citizen office for a visit to the authority in one session. Designed for social data protection and accessibility.
IndustriesInsurance
Claims service with calm and pace: notifications on every channel, checks of invoices and policies with Semantic², agreements signed in the conversation and routine requests resolved in self-service.
IndustriesHealthcare
Hospitals, practices, care services and pharmacies on one platform: appointment service with reminders and callbacks, video consultations in the browser, pre-sorting of requests and patient service on every channel. Designed for the protection of special categories of data.
Guides2 November 2026 · 11 min readIntroducing AI agents responsibly
Disclosure, handover, permissions, logging, approval and measurement: what should be settled before an AI agent goes live in customer service – step by step, with a checklist for the project team.
GlossaryEU AI Act
The EU's AI regulation classifies AI systems by risk and attaches duties to each class: prohibitions, strict requirements for high-risk AI, disclosure in dialogue. What that means for AI in customer service.
GlossaryHallucination
A hallucination is an AI answer that sounds convincing but is wrong or invented. It follows from how language models work and can be limited, but never ruled out completely.
GlossaryGrounding (RAG)
Grounding ties AI answers to verified sources. The common method is retrieval-augmented generation (RAG): first find the relevant content, then answer only from it and name the source.
GlossaryAgentic AI
Agentic AI receives a goal instead of a single instruction. It breaks the goal into steps, uses approved tools and checks interim results. Its value grows, and so does the need for limits and oversight.
See Univerx on your own example.
In a demo we show the platform along your channels and requests, from AI agents to video, workforce scheduling and analytics.
Online, in German or English, with no preparation on your side.