1. What AI Systems We Use

Sprijin360 uses three AI systems:

Important: No AI system on Sprijin360 makes diagnoses, provides clinical advice, or replaces consultation with a professional.

2. The AI Chatbot — How It Works

2.1. What It Is

The chatbot is a virtual assistant that answers questions about the services and providers listed on the platform. Visitors can inquire about available services, prices, working hours, specializations, and can initiate an appointment — 24/7.

The chatbot is NOT a professional (therapist, doctor, lawyer). It is an automated system providing information based on data entered by service providers.

2.2. Information Sources

The chatbot responds based on two sources of information:

Source 1 — Data entered by providers (direct data):

These are the details providers have entered and updated in their admin panel:

• Provider name and description
• Offered services (title, description, price, duration)
• Team (members, specializations, certifications, experience)
• Working hours and availability
• Contact information
• Active discounts (via the affiliation system)

This data is the most current and comes directly from the platform’s database. The chatbot consults it in real time for each query.

Source 2 — Verified knowledge base (frequently asked questions):

The platform builds a knowledge base from visitors’ frequently asked questions:

• Questions and answers are reviewed and approved by an administrator before inclusion in the knowledge base
• Only material meeting quality standards is included
• The knowledge base complements direct data — providing answers to general questions not covered by provider data

Priority rule: When information exists from both sources, the provider’s direct data always takes precedence (being more current and accurate).

2.3. How a Question Is Processed

When a visitor asks a question, the chatbot follows these steps:

1. Understands the question — identifies what information is needed (about a provider, a service, availability, etc.)
2. Consults relevant sources — searches provider data and the verified knowledge base
3. Composes the answer — formulates a response in the visitor’s language based on the information found
4. Responds — displays the answer to the visitor

The chatbot always responds in the language of the page the visitor is on.

2.4. What the Chatbot Can and Cannot Do

2.5. Important Limitations

• Accuracy depends on the provider: The chatbot responds based on information entered by providers. If a provider has not updated their data, the response may be incomplete or outdated.
• Does not understand clinical context: The chatbot cannot interpret symptoms, recommend treatments, or assess the urgency of a situation.
• May err: Like any AI system, the chatbot can generate inaccurate or incomplete responses. Information provided should be verified directly with the service provider.
• Does not automatically retain previous conversations: Each new conversation starts without context from prior interactions unless the visitor voluntarily identifies themselves.

3. Automatic Translation — How It Works

3.1. What It Is

The translation system allows providers to automatically translate their content (service descriptions, profile texts, administrative labels) from the primary language into other languages.

3.2. How It Works

1. The provider writes content in the primary language
2. The provider requests translation into another language
3. The AI system translates the text, preserving meaning and professional terminology
4. The provider reviews and corrects the translation before publishing

3.3. Limitations

• Automatic translations may contain inaccuracies, especially for specialized professional terminology
• The provider is responsible for verifying and correcting translations before publishing
• The platform does not guarantee the accuracy of automatic translations

4. AI Text Suggestions — How It Works

4.1. What It Is

The AI suggestion system assists providers in drafting texts (service descriptions, articles). The provider can request suggestions, which they edit and approve before publishing.

4.2. How It Works

1. The provider requests a suggestion (e.g., "describe this service")
2. The AI system generates a text proposal
3. The provider edits, corrects, and approves the text before publishing

4.3. Limitations

• Suggestions are generated automatically and may contain generic or inaccurate information
• The provider is solely responsible for the final published content
• Suggested text must always be verified and adapted before use

5. Data Protection in the AI Context

5.1. What Data the Chatbot Processes

• Conversation text — visitor questions and chatbot responses
• Language — automatically detected from the visitor’s message
• Affiliation code — if the visitor accessed the platform via an affiliate link

5.2. What the Chatbot Does NOT Process

• Medical or clinical data
• Biometric data
• Financial information (cards, bank accounts)
• Visitor’s exact location

5.3. Conversation Storage

• Conversations are stored in the platform’s database to improve the service
• Conversations may be used (after anonymization and administrator review) to enhance the knowledge base
• The visitor can request deletion of conversations at any time, according to the right to erasure (GDPR, art. 17)

5.4. Data Transfers

• Visitor messages are transmitted to an external AI provider (OpenAI) to generate responses
• This transmission is based on standard contractual clauses (GDPR, art. 46)
• The AI provider does not store data for its own purposes

6. Human Oversight

AI systems on Sprijin360 operate under human supervision:

• Chatbot: Conversations are periodically monitored. Questions and answers are reviewed by administrators before inclusion in the knowledge base.
• Translation: The provider reviews and approves each translation before publishing.
• Text suggestions: The provider edits and approves each suggested text before publishing.

No automated decision is made without the possibility of human intervention.

7. Your Rights

As a user of the AI systems on Sprijin360, you have the right to:

• Know you are interacting with an AI system — the chatbot identifies itself as a virtual assistant
• Request deletion of conversations — at any time, in accordance with GDPR
• Not be subject to a decision based solely on automated processing that produces legal effects (GDPR, art. 22) — the chatbot does not make such decisions
• Contact a human — you can always request contact information to speak directly with the service provider
• Request additional information about how AI works — at contact@sprijin360.ro

8. Updates

This document is updated whenever a significant change occurs in the functioning of AI systems. The current version and date of the last update are indicated in the document header.

Document drafted in accordance with Regulation (EU) 2024/1689 on artificial intelligence, art. 4 (AI Literacy) and art. 50 (Transparency).