The moin widget offers maximum compatibility and is optimally displayed in every browser and on every device.

PDFs can be fully captured automatically during import using OCR (Mistral Document AI). Content from images, tables and scanned pages is also included.

Different agents can be equipped with different roles and personas - e.g. as a friendly, advisory sales agent or a factual, reserved customer service agent.

The AI recognizes based on conversations which topics might be additionally relevant. New topic suggestions are created automatically and gradually expand the knowledge scope of the chatbot.

If the chatbot was unable to provide a suitable response, the problem is identified using an analysis tool.

Enables evaluation of all chat histories. In the detail view, questions, formulations and user reactions can be traced to identify optimization potential.

Notifies agents immediately of new chats for quick response.

Content, design and functions of the teaser can be freely configured, including text, buttons, avatar and colors - matching the intended use and CI. The teaser clarifies the availability of the chatbot and facilitates entry into a conversation.

An overview of all changes to content and channels is available prior to publication.

When no live agent is available, the chatbot takes over the conversation.

Texts can be customized so that the chatbot matches your brand identity both linguistically and visually. You have full editorial control over all chatbot content.

Colors, position and other design elements of the chat widget can be flexibly adapted to the existing corporate design.

When a query is made via the chatbot, certain sources are updated in real time and the latest version is accessed.

To pre-qualify leads, the chatbot can automatically ask questions and forward potential customers to the sales team in a targeted manner.

moinAI understands and processes 98 languages including numerous dialects. This makes the chatbot flexible for international use.

With the response template editor in the Knowledge Base, you can create custom templates such as info cards or interactive buttons, which are dynamically populated with content in GenAI responses.

Enable structured and GDPR-compliant collection of information directly in the chat process.

By adjusting the context size, the chatbot’s performance can be optimized: respond faster with a limited context or provide more thorough answers with a larger context.

The chat widget can be controlled via a JavaScript API. This allows external events such as opening the widget or sending a message to be triggered.

Shows the current availability status of agents for targeted routing and efficient handovers.

Der Zugriff auf moinAI lässt sich gezielt auf bestimmte IP-Adressen oder Domains beschränken. Die Idee hinter einer Allowlist ist simpel: Alles, was nicht auf der Allowlist steht, ist auch nicht erlaubt – erhält also keinen Zugriff.

Access rights to the moinAI Hub can be controlled individually for each team or user. Access to sensitive data is logged and can be traced at any time.

The follow-up question logic analyzes the context in advance to avoid unnecessary follow-up questions and improve user guidance.

Manually maintained content that can be played out independently of generated answers. They ensure maximum content control and avoid inappropriate AI responses.

Chat histories are permanently available for analysis and quality assurance.

Live chat agents can see whether customers are online in the chat.

The AI generates a summary for each conversation, including sentiment analysis, topic tags, and key content. It also allows you to instantly see whether an issue has been resolved or not.

The chatbot's greeting teaser can be adapted according to the page to address users in a context-specific way.

Supervisors can actively assign chats to individual agents and thus steer the team in a targeted manner.

Chatbot changes can be tested before publication in the integrated preview tool - without affecting live operations.

Various tools can be used with moinAI to automate processes with the chatbot. For inspiration, we collect useful AI workflows on our site.

The moinAI MCP server connects external LLMs, such as Claude, with the moinAI Knowledge Base. This allows content in the Knowledge Base to be managed via external AI platforms: knowledge can be searched, created, retrieved, updated, and deleted.

The expert team continuously analyzes chatbot performance and provides targeted recommendations for further optimization.

Offer detailed control options for natural language processing. Configurable include the recognition logic and individual debugging steps. Developed for individual customizations and complex use cases.

The chat widget offers a JavaScript API for integration into existing web applications. Messages can be pre-transmitted or the widget can be opened specifically. Further details can be found in the technical documentation.

The integrated OAuth interface enables direct connection to common CRM systems like HubSpot. Form contents are automatically assigned to the respective contact records and are centrally available to sales and service.

Depending on the channel - e.g. website or app - language, approach and tonality can be individually adapted.

The switch to moinAI succeeds without interruption: existing content, processes and data can be transferred seamlessly.

Bilder im Chat können mit Alt-Texten ausgestattet werden, sodass sie z. B. von Screenreadern erfasst und vorgelesen werden.

The underlying language model continuously improves through independent learning. Additionally, it can be manually trained and optimized - for example with the AI training tool in the hub.

moinAI can collect all data anonymously. If personal information is required, consent is obtained from the users.

Enables individual configuration of privacy-relevant requirements - e.g. for notices, consents and storage rules.

Automatically translates incoming and outgoing messages for global support in 98 languages.

Forms can be individually configured and integrated into the chatbot flow. Captured data can be forwarded to internal systems like HubSpot or used for handovers to live chat. This facilitates data collection and lead generation.

An active conversation context is saved and restored - even during longer breaks. The chatbot can store, validate and pass on information to external systems.

Enables independent selection of the underlying large language model. Depending on the use case, state-of-the-art models from the GPT family or open-weight models can be used. Models can be configured for each bot or project to optimally balance cost, performance, and compliance.

The chatbot can access current information from stored websites and incorporate it into its responses.

Answers can be output in simple language if required - for better comprehensibility and greater accessibility.

In the Playground, AI-generated responses can be regenerated using your own instructions, allowing you to test different scenarios.

System messages at the beginning of the Live Chat, during handover to another agent, and at the end of the chat can be individually customized.

Limits the number of simultaneous chats per agent to avoid overload and ensure quality.

Conversations can be handed over to human colleagues at any time - via live chat (e.g. Zendesk, Userlike) or alternatively by email. The handover takes place automatically and seamlessly in the ongoing dialogue.

Enables the visual design of an individual chatbot avatar. Colors, shapes and logos can be customized to align the design with brand identity.

Enables viewing of completed live chats for follow-up, quality assurance or training.

Standardized text modules can be triggered via shortcuts directly in the input field.

Enables targeted assignment of resources to specific AI agents for clearly defined knowledge areas.

The moinAI widget meets the requirements of the Accessibility Enhancement Act (BFSG) and thus supports barrier-free digital service operations.

moinAI can be implemented and used productively within a few weeks.

Enables the connection of various formats - including PDFs, DOCs, links and question-answer pairs - as knowledge sources for the knowledge base.

Facilitate targeted analysis of large amounts of data. Conversations can be filtered and structured according to topics, ratings or misunderstood requests, for example.

Show in real time which topics are requested how often. The data is constantly updated and enables early detection of trends or recurring concerns.

If users provide unclear information, AI agents will ask for clarification in order to provide the best possible answer.

moinAI can be used across channels. Channels can be, for example, the German-language website and an app. Subdomains are also understood as separate channels if desired.

Access to the hub can be additionally secured by two-factor authentication - for increased protection against unauthorized access.

The technical documentation provides comprehensive information on integrating moinAI and facilitates the work of developers.

Ready-to-use templates are available for many industries—as a quick start guide or inspiration for your own implementation.

The emails sent when users fill in bot forms can easily be created using automatic templates.

The moinAI widget meets the requirements of the WCAG 2.1 guidelines.

Users can click a button to reset the conversation with the chatbot and start again directly within the widget.

The analysis area shows the most important metrics at a glance. Contains data on topic distribution, automation rate and conversion development over freely selectable time periods.

Context variables capture relevant information from the conversation and forward it to external systems via webhooks - in the desired format and to the right place.

Provides a REST API to keep connected resources automatically synchronized.

Topics and answers can be easily created, edited and managed in the editor - individually or together as a team.

Responses from AI agents can be displayed in complex formats such as tables or code blocks.

In the Knowledge Base, entire websites, including their subpages, can be added with a single click.

Users can send files such as PDFs, images or videos directly in the chat to specify concerns or provide additional information.

AI Agents can search CSV resources for specific numbers and solve complex table queries.

Enables the upload of multiple files directly in the chat for better processing.

AI agents incorporate the current time into their responses, ensuring context-relevant communication, e.g. through adapted greetings or time-dependent information.

Enables the setup and customization of the chatbot without programming knowledge. Content, forms and processes can be configured intuitively via drag & drop or visual editors.

Content that should not be openly displayed can be marked as private and remains available exclusively for internal response processes.

Each company is supported by a dedicated contact person in the customer success team - for continuous support and individual assistance.

The help center provides 24/7 access to answers to frequently asked questions - for quick help even outside support hours.

In the hub playground, direct feedback can be given on responses from an AI agent – this feedback is then taken into account in the next response.

Competitors can be specifically excluded so that they are neither mentioned nor recommended by the chatbot.

At the end of a form, information from the chat can be sent to HubSpot – creating a new ticket or updating existing contacts.

Enables login via an existing Google account. Reduces the effort for password management and simplifies user management through central authentication.

Combines RAG technology with generative AI and uses verified content for precise, context-related answers.

Your chatbot's responses are automatically translated into the user's language, enabling direct and clear communication.

Chats can be tagged to facilitate evaluation and categorization.

With webhooks, external systems can be connected and real-time data processed directly from the chatbot. Information can be sent, retrieved or further processed - e.g. for validation or to start external processes.

Parallel conversations are color-coded, enabling better overview in everyday work.

An AI assistant within the hub that automatically generates response suggestions for the chatbot. Supports content creation and reduces editorial effort.

Automatically trigger topic-specific follow-up actions - such as displaying a suitable topic or form - ensuring seamless transitions and higher conversions.

Central overview of all relevant performance indicators. Shows conversation volume, automation level, CSAT scores and conversion rates in real time. The best basis for data-driven optimizations.

Individual onboarding supports a successful project start. This accelerates implementation and enables rapid benefit.

Regular webinars present and explain new features - directly from the product or customer success team.

Enables simultaneous handling of multiple chats, increasing efficiency and capacity.

Set the days of the week, times and periods during which certain teasers are activated.

Enable company-compliant communication by storing specific compliance guidelines that the chatbot follows when generating responses.