FAQ

Question: What model is this chatbot based off of?
Answer: This chatbot is based on the GPT-4 model which is the most powerful and most expensive language model in the world. We are doing a lot in the backend to make this work for you!

Question: Does this chatbot hallucinate or provide inaccurate information?
Answer: AI models like ChatGPT can sometimes 'hallucinate', However, the potential for hallucination significantly diminishes when you carefully construct your system prompt. GPT-4 is engineered to reduce such instances, and when paired with well-structured prompts, the system should avoid hallucinating about 99% of the time. That said, it's important to remember that no system is perfect and there's always the potential for unexpected outputs. Nevertheless, continuous improvements and updates are being made to minimize this occurrence.

Question: How does automatic.chat understand the context of the user's query?
Answer: automatic.chat utilizes advanced machine learning techniques to convert text into mathematical vectors, which can capture context and semantics. When a user inputs a query, it's converted to a vector and compared with vectors in a vector database created from website text. The best matches provide contextually relevant information to formulate a response.

Question: How do you ensure the privacy of our data?
Answer: We take privacy very seriously. The data processed by our chatbot is only used to deliver relevant responses and improve the user experience. Conversations are securely stored for you to see on your bot portal and we have strict policies in place to prevent unauthorized access.

Question: Can this chatbot understand different languages?
Answer: automatic.chat supports over 24 languages, including, French, Spanish, German, Italian, Afrikaans, Indonesian, Russian, Polish, Ukranian, Greek, Latvian, Mandarin, Arabic, Turkish, Japanese, Swahili, Welsh, Korean, Icelandic, Bengali, Urdu, Nepali, Punjabi, Thai, Marathi, Telugu

Question: How will the chatbot integration impact the speed and performance of our website?
Answer: automatic.chat is designed to work seamlessly without noticeably affecting your website's performance. It operates via server-side processes, so it won't impact your site's loading speed or overall user experience.

Question: Can we customize the chatbot to fit the style and theme of our website?
Answer: Absolutely! automatic.chat's visual elements can be customized to align with your brand's identity, including color, typography, and more.

Question: How does automatic.chat handle questions that it doesn't have the information to answer?
Answer: If automatic.chat doesn't find relevant text from your website for a certain query, it uses a pre-defined set of responses that YOU set in your system prompt. By default it will say something like "I'm sorry, I can't seem to find an answer for that on this website"

Question: How often does automatic.chat update its knowledge base from our website?
Answer: When initially deployed, automatic.chat creates its knowledge base from your website's existing text. Currently, you may re-create the knowledge base manually by going to the Sources tab, however we are working on automating this process.

Question: Can we track the chatbot's user interactions?
Answer: Yes, in the bot portal, you can access all conversations that the chatbot has had with users. This includes the user's query, the chatbot's response.

Question: Can automatic.chat handle complex inquiries, such as multi-part questions?
Answer: Yes, automatic.chat is designed to understand and respond to complex queries, thanks to the advanced natural language processing capabilities of the OpenAI API. However, its effectiveness will depend on the richness of the information on your website and the clarity of the user's question.

Question: How much does it cost to integrate your chatbot into our website?
Answer: See the pricing page here

Question: Can automatic.chat be trained to understand industry-specific jargon?
Answer: Yes, automatic.chat can understand and use industry-specific terms as long as they are present in the text of your website. It learns from the information on your site and can adapt to different industries and their specific terminologies.

Question: How long does it take to integrate the chatbot into our website?
Answer: The integration process typically can be done within a few seconds to a few minutes. The duration depends on the complexity of your website and the volume of data.

Question: Can your chatbot handle high traffic volumes without disruption?
Answer: Absolutely. automatic.chat is designed to handle high volumes of user queries without any disruption, ensuring smooth and fast user interactions even during peak times.

Question: How does your chatbot handle sensitive inquiries?
Answer: The chatbot's intended use is to be an informational tool for your website, however if you choose to use it for sensitive inquiries, please contact us and we can discuss how to best handle it.

Question: Can automatic.chat facilitate transactions or take orders?
Answer: Not at the moment. automatic.chat is primarily designed for information retrieval and user interaction. However, with additional programming and depending on your plan, custom feature requests can be made. See pricing here

Question: How does automatic.chat deal with spam or malicious inputs?
Answer: Determining how automatic.chat deals with spam inputs depends on YOU! You can set up a custom system prompt that will respond to spam inputs in a way that you see fit.

Question: Is your chatbot able to learn and improve over time?
Answer: Since automatic.chat allows its users to customize the system prompt themselves, there is currently no functionality to learn over time. However a feature that self-optimizes for system messages is something that is in our roadmap.

Question: Can your chatbot support voice-based queries?
Answer: Currently, our chatbot is text-based. However, it could be integrated with a voice recognition system to handle voice-based queries. This would require additional development and resources and can be created upon request if you are a high priority user.

Question: How do users feel about interacting with a chatbot?
Answer: User perception varies, but many find chatbots helpful for getting quick answers and 24/7 support. We also ensure our chatbot is designed to provide a user-friendly and engaging experience to enhance user satisfaction. This is definitely better than the current chatbot alternative, which is to not understand the context of your question, and provide general resources.

Question: Can we get a demo of your chatbot before integration?
Answer: Absolutely! Go on automatic.chat and enter your website URL for the 10 second demo

Question: Does your chatbot support multiple websites?
Answer: Yes, our chatbot can be integrated with multiple websites within one bot! It can maintain separate knowledge bases for each site, ensuring contextually relevant responses for each one.

Question: Can the chatbot provide feedback or reviews to us?
Answer: While automatic.chat itself doesn't provide or collect feedback explicitly, the user may provide feedback, and you as the organization can look at the conversation history to analyze user interactions, you can gain insights into user satisfaction, popular topics, and areas for improvement.

Question: Can the chatbot be used for marketing or promotional purposes?
Answer: Yes, automatic.chat can be programmed to deliver promotional messages or provide information about sales and offers, making it a valuable tool for marketing efforts. This all depends on how you craft the system prompt.

Question: How does your chatbot compare to others on the market?
Answer: automatic.chat's distinguishing features is the insanely fast chatbot creation process, the ability to provide multiple URLS within one chatbot, the ability to customize the system prompt, the extremely high accuracy and reliable responses, and very good customer service. We really care about solving your problem, and we would love to provide instant support to you.

Question: Can your chatbot work offline?
Answer: As our chatbot relies on server-side processes and the OpenAI API, it requires an internet connection to function.

Question: Can we customize the name of the chatbot?
Answer: Absolutely. The name of the chatbot can be customized to align with your branding or preferences.

Question: What kind of support will we get post-integration?
Answer: The type of support depends on your pricing plan. We provide priority support to those on the Business Plan, post-integration. This includes updates, maintenance, troubleshooting, and customization as per your needs.

Question: How secure is your chatbot?
Answer: Our chatbot is designed with strong security measures. User interactions are processed with high levels of encryption and we follow industry-standard practices to ensure data privacy and security.

Question: Can your chatbot interact with other applications or systems?
Answer: At the moment, the chatbot's only functionality is to scan the contents of your website and provide answers based off of what it reads in the website. It does not currently have the capability of interacting with other applications or systems.

Question: Can I make multiple accounts to create multiple bots?
Answer: No, creating multiple accounts to create multiple bots is strictly prohibited and against our terms. If you would like to create multiple bots, please contact us and we can discuss a plan that works for you.

Question: Do you train bots over sitemap xml?
Answer: Yes, we can train bots over sitemap xml. All you have to do is provide us with the sitemap xml. We support a range of other data ingestion methods such as PDFs, Notion docs, Google Docs, Q&A pairs, and raw text.

Question: Can I see the data that the bot is trained on?
Answer: Yes, the individual pieces of your data can be seen when you click on the 'Sources' tab in the bot portal. Just locate a source, and then click on it to see what text has been trained

Question: Can the chatbot handle sarcastic or ambiguous queries?
Answer: Handling sarcasm or ambiguity can be challenging for any AI. However, our chatbot leverages OpenAI's advanced language model, which is designed to interpret and respond to a wide range of linguistic styles and nuances, including some forms of sarcasm and ambiguity. Furthermore, all of this can be explicitly stated in your customizable system prompt, which will tell the AI how to handle sarcasm or ambiguity.