> For the complete documentation index, see [llms.txt](https://help.sipgate.de/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://help.sipgate.de/cloud-telefonanlage/en/using-sipgate/sipgate-ai-agents/wie-funktioniert-das-wissen-der-sipgate-ai-agents.md). # How does the knowledge of sipgate AI Agents work? The section **Frequently asked customer questions** serves as an internal knowledge base for sipgate AI Agents. Here, admins store information about the company, products, or services that the agent accesses in real time during the conversation to answer questions directly. In the section **Playbooks** structured conversation flows can be created, in which customer data is collected in a targeted way — for example, for a callback or further processing. In addition to the general standard playbook “Other”, individual playbooks such as e.g. **Appointment booking** or \*\*Support request\*\* can be created to cluster calls by topic and evaluate them in a targeted way. sipgate AI Agents uses the information from the website entered during the signup process to obtain the most important data for answering questions. The data captured here can be changed manually at any time afterward. In addition, authorized users can enter recurring customer inquiries together with the corresponding answers into the sipgate AI Agents interface. sipgate AI Agents recognizes the customer's request through keywords and outputs the corresponding answer based on the stored information. #### How do I add new knowledge? sipgate AI Agents can be supplied with new information at any time. To do this, a new playbook is created. Each newly created playbook is linked to the database and stored as a possible answer to a customer inquiry. {% hint style="warning" %} Important: When creating the frequent customer inquiries, please make sure they are not too similar. This could lead to incorrect statements by the sipgate Agent. Also try to formulate the questions asked as precisely as possible. {% endhint %} **Negative example:** * How do I reach the hotline? * How do I reach customer service? **Positive example:** * How do I reach the hotline? * What times is the hotline staffed? If sipgate AI Agents outputs incorrect data, please check whether the information in the relevant answer overlaps in whole or in part with the information in another answer. You can also improve the capture of questions by sipgate AI Agents by using targeted keywords in the respective question. #### Provide knowledge base via upload You can add PDF documents with FAQs and information about your company or your products to the agent's knowledge base. The AI uses this content to answer customer questions specifically during the conversation. For formatting and provision, we have compiled some recommendations that have worked well for us. * Structure: Clear organization with headings * Formatting: Body text, no scans, avoid tables * Images in PDFs are not analyzed * With a lot of content: split across several documents by topic * do not change automatically generated document titles * avoid duplicates with other documents and “Frequent Customer Questions” #### What do I do if sipgate AI Agents creates incorrect or incomplete statements? The AI Agents tries, based on the stored information and scenarios as well as the caller's captured data, to identify sensible customer inquiries and thus provide the corresponding answers. If you notice that incorrect or incomplete information is being given on a specific topic, check the answers in the overview. Try to formulate the answers as precisely as possible. Also make sure that the answers to different inquiries do not become too similar. You can enter various preconfigured tasks in the scenarios or create your own tasks. If you discover an error in the application while doing so, delete the corresponding task and try to improve the workflow with a new formulation. We are happy to receive suggestions as well as improvement and feature requests. --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter: ``` GET https://help.sipgate.de/cloud-telefonanlage/en/using-sipgate/sipgate-ai-agents/wie-funktioniert-das-wissen-der-sipgate-ai-agents.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.