Responsible by design: Zadarma’s approach to AI in communication

Article by Grazina Savickaite – Operations and Content Specialist at Zadarma, published in AI Horizon Journal for AI Ethics & Integrity International Association.

What are the key ethical principles underlying your work with AI?

At Zadarma we base our AI ethical principles on recommendations from various organizations, such as UNESCO, the EU AI Act, and others.

The key ethical principles underlying our work with AI are: 

Privacy and data protection: We prioritize user privacy and comply with regulations such as GDPR.

• Transparency and explainability: At Zadarma, we provide a clear guide to setting up the AI agent. All answers and decisions made by the agent are preconfigured by the user. 
• Human agency and oversight: We ensure that AI systems improve, rather than replace, human participation and decision making. 
• Safety, reliability, and security: Our AI voice agent is reliable, safe, and secure with different security layers such as 2FA and automatic security breach detections. Our infrastructure and workflows have been independently verified to meet the international information security standard ISO/IEC 27001.
• Inclusiveness: At Zadarma, we make AI technologies accessible to all users, independent of their background.

How do you ensure transparency and keep users informed about the functioning of your AI agent?

The AI agent comes with a step-by-step configuration guide. The guide clearly explains each setup step and its meaning. Users can build the agent themselves by selecting from the provided options and adding their own knowledge base. Our 24/7 multilingual customer and technical support is also ready to assist with any questions.

How does your AI voice agent change the way users interact with the company’s services?

Our AI agent helps companies reduce employee workload. This allows them to focus on complex tasks and resolve customer issues faster and more efficiently. The AI agent can also process calls outside working hours, which helps businesses grow their customer base. By automating routine tasks like answering basic questions, it improves customer communication and employee satisfaction.

What new AI development directions does your company plan to pursue in the coming years?

Our future development will focus on real-time AI to reduce response latency and improve conversational quality. Our goal is to improve agent understanding and performance, and expand support for multiple LLM models, such as Gemini.

However, there is no magic button that configures the AI agent automatically. For the agent to work reliably and avoid guessing answers when information is missing from the knowledge base, users should provide accurate information and clear rules to achieve the best performance.

How does your company combine innovation with ethical responsibility in its daily AI work?

Our company combines innovation with ethical responsibility by integrating ethical principles into the development of our AI systems. We protect user privacy and regularly check our models to ensure they perform reliably. We maintain fast innovation, but at the same time we stay transparent, accountable, and trustworthy for our users.