Conversational AI
Conversational AI allows you to have natural-language discussions about your documents. Even though it has a deceptively simple user interface, it is a very powerful tool that can help you answer questions and elicit insights about one or multiple documents at once.
Contrary to general-purpose AI chat bots, ATLAS.ti Conversational AI returns the citations its answers are based on, with easy access to read and code them for later retrieval or further analysis. There is no functional limit to the length of documents.
Please note that ATLAS.ti’s Conversational AI is a beta feature. You are welcome to use it freely, and we welcome your feedback on all aspects of it.
You can access Conversational AI via the Search & Code ribbon, in the Analyze section of the document and document manager toolbars, or the Analysis section of the document or document group context menus.
After selecting the document or documents you need to analyze, Conversational AI shows a window where you will be able to “chat” with your documents.
Clicking on the reference links will open the segment in question so you can easily read it in context and code it. You can also code multiple references at once by clicking “Code all references”. This will add your coding to all references immediately preceding the link, allowing you to quickly code each identified topic.
Below each reply, you will find buttons to create a memo with the question/answer pair, or copy it for use elsewhere. You can also do the same for the whole conversation with the buttons in the very top-right corner of the conversation window.
Conversational AI remembers your current conversation and knows the contents of the documents you add, but it has no context for these documents such as document number, name, group membership, or extant coding. To get the most out of Conversational AI, start the chat with documents that belong together. If you want metadata such as interviewee age to be available to the AI, add it to the documents prior to starting the conversation.
Feel free to ask for clarification, re-word your questions, or clarify your instructions to the AI. Like humans, AIs get “lazy” when reading through large amounts of text, so you can ask to concentrate on specific aspects or even scold it a little. Be careful not to put words in the AI’s mouth, and review the references it gives you. Most important of all, have fun and experiment!
Conversational AI will upload your your questions and selected document content to ATLAS.ti and OpenAI servers. We will never upload your data without your explicit consent. If you wish to proceed, toggle the checkbox where you acknowledge that you agree to our EULA and Privacy Policy. OpenAI will not use ATLAS.ti user data to train OpenAI’s models.
Conversational AI is currently only available on US servers. It is open to our users around the globe, but will always use US servers, regardless of your AI Region setting.