CLAUDIA
Security Center
Security Center
Here you will find an overview of the security measures implemented to protect and ensure compliance of our customers' data. Any additional request can be made via the email legal@cloudhumans.com.
legal@cloudhumans.com
The fastest growing companies scale customer service with Cloud Humans:
Overview
Infrastructure
Internal policies
Stack
Controls
Models
Advanced language models
ClaudIA’s core reasoning, decision making, and multi agent routing run on state of the art foundation models, selected through rigorous internal benchmarks that evaluate response quality, reliability, and real world production performance.
Today, the primary foundation model in use is GPT 5.2 (reasoning), responsible for high level decisions, planning, task decomposition, agent coordination, and final response generation.
Beyond the foundation model, ClaudIA operates with an ecosystem of specialized models, triggered based on task type and context, including:
• fine tuned models optimized for low latency and cost in specific tasks
• speech recognition and voice processing models
• non LLM AI models used for classification, routing, scoring, and deterministic automation
This hybrid architecture ensures every action is handled by the most appropriate and reliable model for the job, delivering accurate, consistent, and efficient responses with full predictability and scalability.
Use
AI built for customer support
ClaudIA is an AI powered customer support agent that talks like a human and resolves the most common support tickets. It finds the right information within each company’s customized content base to deliver accurate answers.
It can also connect to any database or internal system via API, allowing information to be queried, updated, and customized as needed.
Searches publicly available support policies
Searches private, company specific support documentation
Optional: queries and updates data in custom company databases
Training
We only collect data that users voluntarily provide when interacting with our AI agent. Common examples of questions and data processed include:
Question about products and features
Pricing questions and clarifications
Sales related inquiries
Contact information such as email and phone
Sensitive or protected data such as health information or political opinions
Protected personal characteristics such as race or gender
Biometric data
Inference
During inference, Cloud Humans uses the following data to generate recommendations and insights, while ensuring privacy and compliance:
User search queries
Profile data without personally identifiable information
Input commands used to generate personalized messages
Personally identifiable information
Sensitive personal data
AI Data Governance
Data source for AI models
Compliance with data sharing for LLM-based models
Exclusion of customer data from model training
Testing and Monitoring of AI Models
Model effectiveness and quality assurance
Continuous monitoring of model performance
Event log collection for models
Improvement of model explainability
AI Cybersecurity
Protection mechanisms against malicious prompts
Protection against training data poisoning
Strict access control and authentication enforcement
Ongoing supervision of AI system operations
Third-party AI tools and audits
Model effectiveness and quality assurance
Continuous monitoring of model performance
Event log collection for models
Overview
Infrastructure
Internal policies
Stack
Controls
Models
Advanced language models
ClaudIA’s core reasoning, decision making, and multi agent routing run on state of the art foundation models, selected through rigorous internal benchmarks that evaluate response quality, reliability, and real world production performance.
Today, the primary foundation model in use is GPT 5.2 (reasoning), responsible for high level decisions, planning, task decomposition, agent coordination, and final response generation.
Beyond the foundation model, ClaudIA operates with an ecosystem of specialized models, triggered based on task type and context, including:
• fine tuned models optimized for low latency and cost in specific tasks
• speech recognition and voice processing models
• non LLM AI models used for classification, routing, scoring, and deterministic automation
This hybrid architecture ensures every action is handled by the most appropriate and reliable model for the job, delivering accurate, consistent, and efficient responses with full predictability and scalability.
Use
AI built for customer support
ClaudIA is an AI powered customer support agent that talks like a human and resolves the most common support tickets. It finds the right information within each company’s customized content base to deliver accurate answers.
It can also connect to any database or internal system via API, allowing information to be queried, updated, and customized as needed.
Searches publicly available support policies
Searches private, company specific support documentation
Optional: queries and updates data in custom company databases
Training
We only collect data that users voluntarily provide when interacting with our AI agent. Common examples of questions and data processed include:
Question about products and features
Pricing questions and clarifications
Sales related inquiries
Contact information such as email and phone
Sensitive or protected data such as health information or political opinions
Protected personal characteristics such as race or gender
Biometric data
Inference
During inference, Cloud Humans uses the following data to generate recommendations and insights, while ensuring privacy and compliance:
User search queries
Profile data without personally identifiable information
Input commands used to generate personalized messages
Personally identifiable information
Sensitive personal data
AI Data Governance
Data source for AI models
Compliance with data sharing for LLM-based models
Exclusion of customer data from model training
Testing and Monitoring of AI Models
Model effectiveness and quality assurance
Continuous monitoring of model performance
Event log collection for models
Improvement of model explainability
AI Cybersecurity
Protection mechanisms against malicious prompts
Protection against training data poisoning
Strict access control and authentication enforcement
Ongoing supervision of AI system operations
Third-party AI tools and audits
Model effectiveness and quality assurance
Continuous monitoring of model performance
Event log collection for models
Overview
Infrastructure
Internal policies
Stack
Controls
Models
Advanced language models
ClaudIA’s core reasoning, decision making, and multi agent routing run on state of the art foundation models, selected through rigorous internal benchmarks that evaluate response quality, reliability, and real world production performance.
Today, the primary foundation model in use is GPT 5.2 (reasoning), responsible for high level decisions, planning, task decomposition, agent coordination, and final response generation.
Beyond the foundation model, ClaudIA operates with an ecosystem of specialized models, triggered based on task type and context, including:
• fine tuned models optimized for low latency and cost in specific tasks
• speech recognition and voice processing models
• non LLM AI models used for classification, routing, scoring, and deterministic automation
This hybrid architecture ensures every action is handled by the most appropriate and reliable model for the job, delivering accurate, consistent, and efficient responses with full predictability and scalability.
Use
AI built for customer support
ClaudIA is an AI powered customer support agent that talks like a human and resolves the most common support tickets. It finds the right information within each company’s customized content base to deliver accurate answers.
It can also connect to any database or internal system via API, allowing information to be queried, updated, and customized as needed.
Searches publicly available support policies
Searches private, company specific support documentation
Optional: queries and updates data in custom company databases
Training
We only collect data that users voluntarily provide when interacting with our AI agent. Common examples of questions and data processed include:
Question about products and features
Pricing questions and clarifications
Sales related inquiries
Contact information such as email and phone
Sensitive or protected data such as health information or political opinions
Protected personal characteristics such as race or gender
Biometric data
Inference
During inference, Cloud Humans uses the following data to generate recommendations and insights, while ensuring privacy and compliance:
User search queries
Profile data without personally identifiable information
Input commands used to generate personalized messages
Personally identifiable information
Sensitive personal data
AI Data Governance
Data source for AI models
Compliance with data sharing for LLM-based models
Exclusion of customer data from model training
Testing and Monitoring of AI Models
Model effectiveness and quality assurance
Continuous monitoring of model performance
Event log collection for models
Improvement of model explainability
AI Cybersecurity
Protection mechanisms against malicious prompts
Protection against training data poisoning
Strict access control and authentication enforcement
Ongoing supervision of AI system operations
Third-party AI tools and audits
Model effectiveness and quality assurance
Continuous monitoring of model performance
Event log collection for models