How We Use AI

This page is for clients and potential clients who want to understand exactly how we use AI in our work — and what that means for their data. No legal language, just a straight answer.

1. The Tools We Build With

We use a small, deliberate set of AI tools, and we pick them based on what each project actually needs:

• OpenAI GPT models — for LLM integration, content generation pipelines, and agentic reasoning.
• Google Gemini — for multimodal tasks and document processing.
• Anthropic Claude — for code review, complex reasoning, and structured output work.
• n8n — for workflow automation and agent orchestration.

If a project calls for a different or additional tool, we tell you at the scoping stage and get your agreement before using it. No surprises.

2. What Happens to Your Data When We Use These Tools

• We pass only the minimum data required to complete the task at hand.
• Data shared with LLM APIs during a project is processed within the session. We do not store client data in any AI platform account.
• We do not use your data to fine-tune or train any model.
• Credentials, API keys, and proprietary business information are handled in isolated environments and are never shared across client engagements.

3. What We Do Not Do

• We do not sell client data.
• We do not use one client's business logic, data, or systems to inform or train anything for another client.
• We do not retain access to client systems after project handover unless a separate maintenance agreement is in place and explicitly signed.

4. Staying Current

The AI landscape changes fast. We track changes to the tools we use — including API policy updates, deprecations, and pricing shifts — and proactively flag anything that could affect a delivered product. If something changes after handover that you should know about, we'll reach out.

5. Questions

If you have questions about how we use AI in a specific project or in general, email us at hello@sahajta.com. We respond within one business day.