AffiniAI (Beta)

1) You submit a query through the Affinitext platform.
2) AffiniAI receives and analyses your query, identifying the relevant legal concepts and information needed.
3) AffiniAI gathers relevant data from your project’s documents (such as clauses, definitions, or schedules).
4) AffiniAI sends the data to ChatGPT-4.1, the advanced language model powering AffiniAI.
5) ChatGPT-4.1 processes the information and may request additional context if needed.
• If more information is required, AffiniAI retrieves it and resubmits to ChatGPT-4.1.
• If not, ChatGPT-4.1 generates a draft response.
6) AffiniAI reviews and formats the response, ensuring clarity, accuracy, and traceability to source material.
7) The final response is delivered to you within the Affinitext Intelligent Document platform

‍

AffiniAI is powered by ChatGPT-4.1, but it is fundamentally different from simply using ChatGPT-4.1 or other large language models (LLMs) on their own. When you use ChatGPT-4.1 or another LLM directly, it generates responses quickly based on your prompt and its pre-trained knowledge. However, it does not deeply consult or rigorously analyse your specific documents, nor does it provide the quality, evidence-backed answers that AffiniAI delivers through its orchestrated, multi-step workflow.

AffiniAI prioritizes accuracy, traceability, and legal defensibility over raw speed. As a result, response times are typically measured in minutes rather than seconds.

Affinitext offers you a choice in where your queries are processed by ChatGPT-4.1:
• Azure Option: Hosted on Microsoft’s secure servers in Sweden.
• OpenAI Option: Hosted on OpenAI’s servers, generally in the United States.
Regardless of your choice, your data is transmitted securely and processed with the highest level of protection. Data residency may affect your organisation's compliance requirements.

No. Your data (e.g., full documents, contracts, and related files) remains within the Affinitext system and is not shared in full with Azure, OpenAI, or any third parties. Only the minimum necessary portions of a query context (such as specific clauses or definitions) are temporarily shared with ChatGPT-4.1 to generate a response for your particular query and session. After the session ends, this data is discarded and not stored or reused.

No. Affinitext or AffiniAI do not use your data for training purposes. All documents, queries, and interactions are processed solely to provide you with accurate responses and insights. We do not store, reuse, or share your data for AI model training, ensuring your sensitive contract information remains confidential and secure.

AffiniAI stores each user's chat history, which includes conversation details. This chat history is available only to the user. You can rename or delete your chats at any time—once deleted, they are gone for good!

We take data privacy very seriously. Only the minimal required data (such as specific clauses) is shared with ChatGPT-4.1, and complete documents or sensitive information are never transmitted. All data is handled through secure, encrypted channels, and we regularly review and update our security protocols to address emerging threats.

AffiniAI is built on the secure Affinitext platform, which adheres to industry best practices, such as:
• Data encryption in transit and at rest
• Role-based access controls
• Encrypted connections
• Regular security audits and penetration testing
Only authorised users can access AffiniAI, with stringent control of permissions.

Affinitext upholds and complies with vital security certifications, including ISO 27001:2022, Cyber Essentials, Cyber Essentials Plus, and UK Ministry of Defence Accreditation (OFFICIAL SENSITIVE).

Azure’s security accreditations include:
• ISO/IEC 27001, for information security management systems.
• ISO/IEC 27017, for cloud-specific security.
• ISO/IEC 27018, for privacy protection of personally identifiable information (PII).
• SOC 1, SOC 2, SOC 3, independent audit reports that verify controls related to security, availability, processing integrity, confidentiality, and privacy.
• GDPR Compliance, adhering to the EU’s General Data Protection Regulation (GDPR).
• FedRAMP, for U.S. federal agencies, with stringent security standards.
• HIPAA Compliance, for safeguarding protected health information (PHI).

OpenAI certifications include:
• SOC 2 Type 1 Certification, ensuring compliance with security, availability, and confidentiality standards.
• Encryption both in transit and at rest, safeguarding your data during processing.
• Data privacy and confidentiality measures, with a commitment to ensuring that user data is not stored or used for training without explicit permission.
• Privacy-focused updates to align with global data protection standards.

We chose ChatGPT-4.1 because it offers advanced natural language understanding and processing capabilities. These enable AffiniAI to interpret complex queries and provide accurate, contextual responses—ideal for sophisticated contract-related questions.
AffiniAI is designed with flexibility in mind. If a more advanced or suitable Large Language Model (LLM) becomes available, we can seamlessly integrate it into the system. AffiniAI will continue leveraging the best available AI technology to provide the most accurate and efficient responses.

We welcome your feedback! Please contact our support team through Chat on the Affinitext platform or email us at support@affinitextgroup.com with any questions, concerns or suggestions.