An AI Assistant That Knows Your Documents

01 What is RAG?

RAG, retrieval-augmented generation, makes a language model answer from your documents instead of from memory. The assistant first finds the passages that relate to the question, then the model writes the answer from those passages only. When nothing relevant is found, a well-built assistant says so instead of guessing.

02 How is a RAG chatbot different from ChatGPT?

A RAG chatbot answers from your own documents. A general chatbot answers from what its model learned in training, which knows nothing about your contracts or policies. The RAG chatbot looks things up in your files each time, so its answers change when your documents do.

03 Which documents can it read?

PDFs, Word files, emails, help articles and the records in your database, as long as the text can be read. We agree with you which folders and systems are in, and which stay out, before we make anything searchable.

04 Can it show where each answer came from?

It can. Under each answer we can show the document and the passage it came from, so your team can check it. Whether to show sources, and how, is a choice we make with you when we design the assistant.

05 How do you stop it making things up?

The model gets only the passages retrieval found and is told to answer from them alone. A second check compares the answer with those passages, and when the documents do not hold the answer, the assistant says so. A test set of real questions is rerun after every change.

06 Is our data kept private?

Where your data goes is your choice. With hosted models from OpenAI or Anthropic, it is processed by that provider. With open models we run on your own servers or cloud through Ollama, documents and prompts stay inside your network. We work only in the accounts you invite us into.

07 Can different people see different documents?

Yes. Each person signs in, and the assistant searches only the documents their role may see, so an answer never draws on a file that person could not open themselves.

08 What happens when our documents change?

The document pipeline picks up new and changed files and updates the search index, so answers follow your latest version. We agree with you how often it runs, and which folders it watches.

09 RAG or fine-tuning: which do we need?

Usually RAG. It answers from your documents as they are today, changes when they change, and you can see where an answer came from. Fine-tuning trains a model on examples to change how it writes or behaves; it does not keep facts current. Some projects use both.

10 Is this the same as a customer support chatbot?

Not quite. This service covers assistants that answer from your internal knowledge and documents, mostly for your own team. Answering customer tickets across email, chat and WhatsApp is support automation, a separate service, though it often uses the same retrieval underneath.

11 How much does RAG development cost?

Three things drive the cost: how many documents and systems the assistant reads, who uses it and with what access, and how much testing the answers need. Tell us what your team asks, and within two working days we come back with an approach, a rough timeline and a rough number.

12 How long does it take to build?

It depends on the documents and the systems they live in. We start with a pilot on a sample of your documents, so your team asks real questions before we add the rest. We agree the timeline at the discovery call.

13 Who owns the code and the index?

You do. The code goes straight into your own code repository, and the cloud account that holds the search index is in your company's name. Our team joins as collaborators, and you can remove our access at any time.

14 Do you work with UK and US clients from India?

Yes. We sign an NDA before we start, on request, and work only in the accounts you invite us into. The code repository, cloud account and domain stay in your company's name, and you can remove our access at any time.