How does ChatBot use AI?
ChatBot leverages AI to help customers build smarter and more dynamic bots using large language models (LLMs). AI Assist, an optional feature, allows the ChatBot to respond based on AI Knowledge uploaded by the user. This includes websites, PDFs, terms and conditions, help articles, and other documents.
If no manual responses are defined in the ChatBot scenario, the AI Assist draws on this AI Knowledge to generate answers.
If manual responses exist, they take precedence over AI responses.
For a more detailed overview of generative AI for chats, please visit the ChatBot AI Assist page.
What is AI Assist?
AI Assist is a chatbot feature that automatically answers customer questions by referencing your website content, help center articles, and other support materials. Unlike traditional bots that rely on predefined flows or scripts, AI Assist uses generative AI to interpret and respond to queries dynamically. It uses generative AI to come up with helpful responses on the spot — like a knowledgeable team member who’s always available.
How accurate is ChatBot AI Assist?
It provides answers based on the content it was previously trained on, such as website pages or text articles. There’s always a risk that an AI model can misinterpret the provided information. Make sure that the bot is trained with the proper scope of data and that it’s thoroughly tested before publishing.
Do you use a proprietary AI model?
Yes. We use our own instance of the open-source model. We’ve further tuned it using internal methods and custom instruction datasets tailored for ChatBot use cases. The model is hosted on our infrastructure and fully managed by us.
Is customer data used to train the model?
We do not use your data to train large language models (LLMs), and we never share or mix your data with other customers.
However, we may use your data — such as uploaded content or bot interactions — to improve the functionality and performance of our services. This includes training small, specialized in-house AI models for features like tag suggestions or canned responses.
All data used in this way remains within our secure infrastructure and is handled with the highest standards of privacy and integrity.
Do you send data to third-party AI providers like OpenAI?
In general, no. All AI processing happens within our infrastructure. However, there is one exception: If you’re using a multilingual bot and a user communicates in a different language, we may use OpenAI’s model only for final message translation. In this case, the final ChatBot response is sent to OpenAI to translate it into the target language.
How does ChatBot ensure GDPR compliance when using AI?
When we select an AI-authorized partner as a sub-processor with a primary focus on communication and customer service enrichment, we finalize a data processing agreement to secure the exchange of information. We authorize these partners to process your data in accordance with and only for the purpose specified in our Privacy Policy and Data Processing Addendum (DPA).
Our DPA binds these partners, and any international transfer of personal data is carried out under the terms of the DPA, using Standard Contractual Clauses (SCC) as a transfer mechanism. The details of data transfers can be found here.
What security measures are in place to protect my data?
ChatBot employs several robust security measures to protect your data. These include data encryption at rest and in transit, regular security audits, and compliance checks to ensure that all security practices meet or exceed industry standards. For comprehensive details on our technical and organizational security measures, please refer to our Security page and Exhibits A and B to our Data Processing Addendum, where you can find more information about how we protect client data.
Who owns the outcomes from the AI features?
When you use ChatBot’s AI features, you own the results, including all responses and insights generated during your interactions. We ensure that these benefits directly enhance your business. While you own your data and the immediate outputs, ChatBot retains ownership of the underlying AI technology and software. The capabilities of our AI, including any improvements, are the intellectual property of ChatBot or its licensors. You’re responsible for how you use the AI-generated outputs.
It’s important to understand that while you own the immediate outputs from our AI, this ownership does not extend to any third party intellectual property that might inadvertently be included in these outputs. Due to the generative nature of AI, outputs can occasionally contain elements that are similar to material created by or belonging to others. You are responsible for ensuring that your use of these AI-generated outputs does not infringe on third-party rights and complies with all relevant laws. This means reviewing the outputs carefully, especially if they are to be used publicly or commercially, to ensure they do not contain or mimic proprietary content without permission.