Your chatbot answers customers in seconds. Do you know which country it does that work in?
Until recently, most small businesses never asked. On 12 November 2026, OpenAI gave them a reason to start. The company announced that supported requests to its API, and the AI work that generates the answers, will now be processed on servers in India, for businesses that need their AI work done in the country.
What happened
OpenAI announced local processing and inference for its API in India. For supported workloads, both the API request and the model computation that produces the response now take place in India, and supported customer content is stored there too.
The offering starts with one model, GPT-6 Luna, through the Responses and Chat Completions APIs, for direct API use. It does not automatically cover ChatGPT or other models.
It builds on data residency OpenAI started in India in May 2025, which covered storage for ChatGPT Enterprise, ChatGPT Edu and API Platform users. This step adds the processing itself.
India joins Australia and Japan as the latest markets where OpenAI offers API inference residency.
The change matters most where rules require it. In India, the Reserve Bank of India has mandated local processing and storage of data for the financial sector. Nitin Bawankule, OpenAI’s Head of Enterprise in India, said local inference gives businesses more control over where supported workloads run, and helps remove a practical barrier to using AI at work.
There are limits worth knowing. Access requires an organisation to have active Zero Data Retention approval from OpenAI, and the India region supports standard pay-as-you-go processing. OpenAI’s hosted tools are not available for regional processing there.
What it means for a small business
Where your AI works is becoming a buying question, not a technical detail. A customer, a partner or a regulator can now reasonably ask: does my message leave the country?
This is not only an India story. Data rules differ by country, and the direction is the same everywhere. The United Kingdom has data protection rules, Pakistan is building its own data framework, and India already requires local handling for some sectors. The list of countries where AI work must stay local is growing.
For a small business, the risk is practical. If you serve customers in a country with data rules, and your chatbot or AI tool processes their messages somewhere else, you may be blocked from that market, or from a contract that requires local handling. Finding out after you sign a customer is the expensive way to learn.
The opportunity runs the other way too. Being able to say where your customers’ data is processed is a trust signal. It is the kind of plain answer that wins a cautious customer.
What to do about it
- Ask your AI and chatbot vendors one question: in which country is my customers’ data processed and stored? Get the answer in writing. If the vendor cannot answer, that is an answer.
- Check the rules for the customers you actually serve. If you sell to customers in India, check whether their sector requires local processing. The same check applies to the United Kingdom, the United States and Pakistan: know the rules where your customers are, not only where you are.
- Add data location to your checklist when choosing AI tools. Price and features matter, but so does where the work happens. A cheaper tool that processes data in the wrong country can cost more than it saves.
- If you build with OpenAI’s API and serve Indian customers, look at the India option. It needs Zero Data Retention approval, and it covers GPT-6 Luna through the Responses and Chat Completions APIs, so check the eligibility before you plan around it.
- Tell your customers, briefly, where their data goes. One plain sentence on your website or in your chatbot’s first message builds more trust than a long policy page nobody reads.
- Keep your own records straight. Note which tools process which customer data and where. When a customer or a regulator asks, you should be able to answer in a minute.
Digilistan builds chatbots and AI automation for small businesses, and choosing where customer data is processed is part of getting that setup right. If your chatbot or AI setup needs a review, that is a conversation worth having.
OpenAI’s India launch is one company’s one-country change. The pattern behind it is the story: AI is getting a postcode. The businesses that ask where their AI works, and can answer plainly, will be ready for the next country on the list. The ones that never ask will find out from a customer, or a contract, instead.
Sources
- Business Standard, “OpenAI to process API data, run AI inference locally for Indian companies”, 12 November 2026: https://www.business-standard.com/amp/technology/artificial-intelligence/openai-to-process-api-data-run-ai-inference-locally-for-indian-companies-126111201464_1.html
- The Hindu BusinessLine, “OpenAI announces local AI processing in India with API inference residency”, 12 November 2026: https://www.thehindubusinessline.com/info-tech/openaiannounceslocalaiprocessinginindiawithapi-inferenceresidency/article71692832.ece
- The Economic Times, “OpenAI brings local inference to India, starting with GPT-6 Luna”, 12 November 2026: https://economictimes.indiatimes.com/ai/ai-insights/openai-brings-local-inference-to-india-starting-with-gpt-6-luna/articleshow/135760461.cms
- Express Computer, “OpenAI extends API inference residency to India for enterprise AI workloads”, 12 November 2026: https://www.expresscomputer.in/news/openai-extends-api-inference-residency-to-india-for-enterprise-ai-workloads/141213/
