FAQs

Frag-Maria is a digital employee for SMEs. It combines AI chatbot and voice agent and helps companies advise customers, handle service requests, answer calls, and also relieve colleagues internally as a knowledgeable assistant — around the clock, brand-consistent, and in almost all international languages.

Maria is not intended as a simple chat function but as a digital employee. According to the website, she was designed as a role rather than a widget: she informs, guides, qualifies, structures, relieves, and hands off with context, brand persona, and expertise. In contrast, classic rule-based chats usually only react to keywords, follow rigid flows, and often end with a redirection to customer service for more complex questions.

Frag-Maria is aimed at SMEs in the DACH region who want to make their digital customer dialogue more professional. The site particularly appeals to companies that stand for quality, brand personality, and good customer relationships, but whose teams are not available indefinitely. This is exactly where Maria is intended to bridge the gap between aspiration and digital reality.

The term means that Maria not only provides answers but acts as an active, brand-oriented support within the company. She accompanies conversations, helps with purchasing decisions, structures service cases, supports internal knowledge access, and can, if necessary, hand over to real employees via Human-in-the-Loop.

Maria is available to work in German, French, Italian, and English. These four languages are offered as standard in both the service description and the proposal models. However, many other international foreign languages are also possible.

Yes. Maria has been "Live in action since 2025" and has several references, including kitchen-more.ch, hurom.swiss, rotel.swiss and kochenamtisch.ch. and more.

Among other things, Maria can handle product advice, 24/7 FAQ & support, appointment scheduling, callback management, order status and delivery information, returns and complaint processing, brand and certification questions, onboarding and initial consultation, and multilingual customer dialogue. In addition, inbound and outbound telephony as well as human fallback and internal employee training are possible.

Maria explains that rule-based chats often only cover simple FAQs and reach their limits with complex product questions, purchasing decisions, or structured service cases. The central weaknesses are rigid flows, a missing brand persona, no learning process after launch, missed purchasing moments, increased effort due to incorrect advice, and insufficient security protection. Maria closes precisely these gaps.

Through a multi-layered system: (1) Retrieval-Augmented Design – Maria responds exclusively based on the integrated knowledge corpus, not from generic LLM knowledge. (2) Defined fallbacks in case of uncertainty – Maria escalates to a human instead of hallucinating. (3) Continuous quality assurance – conversation logs are regularly analyzed, and incorrect answers are corrected in the next sprint. The result after 12 months of live operation: 98% error-free sessions – and a decreasing error rate with each iteration.

Specifically: (1) Opt-in management is integrated. (2) Data minimization as a design principle – Maria only asks for what she needs. (3) Conversation data with a clearly defined retention period. (4) No training on customer data without explicit consent. (5) Role/rights system, audit trails for all data protection-relevant operations. (6) Processing compliant with GDPR Art. 6. (7) Operation according to Swiss DPA, based on GDPR. EU AI Act: Maria falls under Limited-Risk and fulfills all transparency requirements.

ROI is generated on several levels: (1) Support relief – standard inquiries are resolved without team effort. (2) Conversion increase – Maria supports customers at the moment of purchase decision, increasing dwell time and conversion rate (proven +1.2% conversion at Kitchen-More.ch). (3) 24/7 availability without personnel costs. (4) Error reduction through proactive advice – fewer returns, fewer complaints. The KPIs: First Contact Resolution (FCR), escalation rate, CSAT, Conversion Rate. We provide a performance dashboard.

We take care of 1st-level support, technical maintenance, and security and compatibility updates. Your team will be fully empowered in phase 6 (handover): content management in the system, escalation management. Regular optimization sprints ensure that Maria grows with your company. Frag-Maria is integrated via an embedding code – without access to your system infrastructure.

Maria handles far more than simple FAQ answers in the chat. She assists with product advice, answers service questions, supports purchasing decisions, helps with scheduling appointments, provides information on order status and delivery, processes returns and complaint inquiries, and conducts multilingual customer dialogues.

Maria is strong exactly where many classic chats fall short: at the decisive moment before a purchase. She answers specific product questions, understands industry and product context, maintains the brand tone, and guides the user structurally to the next meaningful decision. This is precisely what differentiates her from purely rule-based FAQ or widget solutions.

Yes. With Voice AI, Maria can answer, intelligently qualify, accompany, and document incoming calls. The larger model also includes real-time speech recognition, natural language-based responses, seamless handovers for complex cases, and integration into existing telephone systems.

Chat AI is text-based customer dialogue on websites and in shops. Voice AI expands Maria with AI telephony, featuring real-time speech recognition and natural language playback. The avatar is the next step in development and is planned on the roadmap as a visual brand persona with facial expressions and gestures.

Yes. Maria is not only designed for external customer dialogue but also supports internally as an informative colleague. Teams are relieved of recurring questions, and knowledge is made available in a structured way.

Maria does not work in isolation, but with clear escalation logic. If a case becomes complex or human support is useful, it can be handed over to a team – with full context retention. This way, no previous conversation information is lost, and the transition remains comprehensible for customers.

Yes. Exactly these use cases are among Maria's strengths. These include order status, delivery information, returns, complaints, product questions, support cases, and other structured service requests. Maria helps to properly record and process these inquiries.

Maria was conceived as a role, not as a widget. She works with brand persona, tonality profile, and expertise, maintains the brand tone, and is intended to make a company's quality promise visible in digital customer communication. Precisely because of this, she does not appear generic, but rather like a recognizable digital employee.

Yes. Maria is designed for use on websites and in shops. The telephony model also includes integration into existing telephone systems and compatibility with cloud telephony systems. In the Enterprise model, deeper CRM, ERP, and shop integrations are also possible.

Maria is familiar not only with the concept but also with live data from real operations. Currently, she achieves an average of 98% error-free sessions per platform deployment, an average of 3,490 monthly conversations, over 257 escalations routed to support, and a measured conversion increase of +1.2% in shop deployments.

The onboarding starts with a free use case scan. In about 30 minutes, we jointly analyze which initial lever in your company will make the biggest difference. Afterwards, we decide which model and which first use case is best suited for the start.

The Use Case Scan is a free initial consultation where we assess how Frag-Maria can provide the greatest benefit to you. The goal is not a theoretical pitch, but a realistic first application case with a clear impact on your daily work.

For the first productive simple use case, the setup takes 2 to 3 weeks. After that, Maria is ready for operation, measurable, and integrated into the business with real KPIs.

The Fair-Start model allows for an initial commitment of 3 months. The one-time setup fee will be credited if the contract is extended. The idea behind it is: experience first, then decide – transparent and with no hidden agenda.

Enterprise & Custom is designed for businesses that need more than a standard setup. This includes unlimited chats and calls, deeper CRM, ERP, and shop integrations, outbound telephony and campaigns, multi-brand or multi-location operations, avatar development, dedicated security audits, and SLA models with defined response times.

Standard models have a minimum term of 12 months. The Fair Start model initially has a Fair Start phase of 3 months; if continued, the standard term of 12 months then applies.

The "upgrade models" have a clear offer logic from entry-level to growth and enterprise. Frag-Maria is designed for growth and expansion, enabling larger setups.

Maria offers transparent fixed prices, no hidden hourly billing, and a free use case scan before contract signing. Additionally, "No lock-in without added value."

This is exactly why we created the Use Case Scan. Together, we assess which initial lever, which channel, and which model will bring the greatest benefit to your company. The decision should not be based on a pitch deck, but on a realistic entry point with measurable added value.