Service:

Internal AI chat for your company knowledge

Short description:

Processing time:

5-8 working days

Price:

CHF 149 / month¹

Description:

Our AI assistant makes your corporate knowledge searchable using natural language – across cloud drives, intranets, and documents. Thanks to semantic search, it delivers precise answers instead of mere keyword hits.

What is included:

• Connection to cloud services
• Integration of web data sources
• Processing of documents
• Semantic AI search across all content
• Interactive natural language chatbot
• Interfaces to Slack, Teams or web chat
• Access and authorization management
• GDPR-compliant data processing
• Adaptation to your workflows
• Scalable architecture (cloud or on-premises)

Creation process:

• Joint needs analysis & goal definition
• Data source audit & connection concept
• Workflow design for AI integration
• Implementation of semantic search
• Integration into your system environment
• Tests & Quality Assurance
• Training & Go-Live
• Handover with documentation

This is what you get:

• Configured AI assistant in your environment
• Connection to all agreed data sources
• Semantic search across your content
• Integrated chatbot interface for users
• Project documentation with architecture & data flow
• User Manual for Admins & Users

Add-ons:

🌎 Multilingual assistant

Surcharge: 960 CHF

Expanding the assistant to include additional languages (e.g., English, French). Including content translation or the use of a multilingual AI model to ensure queries and responses work in multiple languages.

Extra time: +3 days

🔌 Additional system integration

Surcharge: 1960 CHF

Connecting another specialized data source or application. For example, integrating a CRM system, a database, or other tools so that the assistant can also obtain information from them.

Extra time: +3 days

🤖 Advanced AI option

Surcharge: 960 CHF

Use a more powerful AI model (e.g., upgrade from GPT-4 to GPT-5) or fine-tune the language model to your company data for even more precise answers. Recommended for very specific subject matter.

Extra time: +0 days

📊 Analytics Dashboard

Surcharge: 960 CHF

Integration of a dashboard to analyze assistant usage. You gain insight into queries, frequently searched topics, and user feedback, allowing you to continuously optimize your knowledge pool.

Extra time: +4 days

Other/Hosting:

🇪🇺 EU-Hosted AI Integration (GDPR compliant)

Use of EU-hosted AI models (e.g. B. OpenAI or Azure) via GDPR-compliant API connection
Data security & compliance with European data protection standards

🇨🇭 Swiss hosting (GDPR/FADP compliant)

Hosting and operation of the LLM server (AI) in Switzerland
For applications that have the highest data protection requirements or are legally obliged to process data exclusively in Switzerland.

FAQ:

What data sources can an internal AI chat use?

Our solution supports a variety of sources: cloud storage like Google Drive, OneDrive, and Dropbox; websites/intranets—such as your company wiki, knowledge base, or SharePoint—as well as files (PDF, Word, Excel, PowerPoint, and many more) that you provide us. We configure dedicated connectors for each agreed data source so that the assistant can index and search the content. Additional systems (e.g., databases, CRM) can be integrated via add-ons.

How is data security and data protection ensured?

Data security is our top priority. All data remains exclusively within your infrastructure – no data is transferred to unauthorized third parties. An internal AI chat is operated in a closed environment, ensuring no confidential information is leaked. Our solution is fully GDPR-compliant and can be hosted on servers in Switzerland/the EU if required. Access is encrypted and role-based. Upon request, we offer an add-on for on-premises installation or enhanced security measures.

Through which channels can you use the assistant?

You can integrate the knowledge assistant via chat platforms like Slack or Microsoft Teams—for example, as a chatbot that can be @mentioned in channels. Web chat integration is also possible (for example, on your intranet page or helpdesk portal). This ensures that your users access the service exactly where they need it.

Does the assistant support multiple languages?

By default, an internal AI chat is configured for one language (e.g., German). However, the assistant can be configured for multiple languages upon request – we offer the *Multilingual Assistant* add-on for this purpose. This allows requests and documents to be processed in multiple languages. The AI recognizes the language of the request and provides answers in the same language. This allows you to cover German and English content simultaneously, for example.

Are there any ongoing costs after implementation?

The monthly price includes everything – including setup, maintenance, and support. Any usage costs for connected AI services (e.g., API requests or storage) are also covered, as long as they remain within normal limits. You don't have to worry about anything else – we take care of reliable operation for you.

Can the knowledge assistant run on our own servers?

Yes, if you don't want to use the cloud for compliance reasons, an on-premises deployment is possible. In this case, the solution is installed in your data center or private cloud, and all components (database, search index, AI models) run under your control. This scenario requires additional effort for installation and configuration. With on-premises, you retain full data sovereignty, but you must provide the appropriate server infrastructure.

What is the difference between semantic search and traditional search?

A conventional search usually compares precisely entered keywords with the database. Semantic search, on the other hand, understands the context of the query. This means it also recognizes synonyms, connections, and the intent behind a question. This allows it to deliver relevant information even if the user doesn't use the exact keywords. In short: Semantic search delivers more precise and contextual results, while keyword-based search only filters for word occurrences. This leads to a significantly better search experience for users.