techport HR Suite

    Knowledge Management Software with AI: Answers from Your Own Documents

    Your company's knowledge exists. It is contained in handbooks, policies, process descriptions, and works agreements. Only nobody finds it when they need it. So employees ask the HR department, and HR answers the same question for the tenth time.

    The Company Knowledge module of the techport HR Suite starts right there. An AI assistant answers questions from your own documents, cites the source for every answer, and states openly when your documents do not provide a reliable answer.

    Sound familiar?

    • The travel expense policy is somewhere on the drive, in three different versions. Nobody knows exactly which one applies.
    • New employees ask the exact same questions in their first few weeks as their predecessors. The answers arrive by email, and with the next new starter, the cycle begins all over again.
    • The intranet wiki was launched with great commitment. Today hardly anyone maintains it, and the search returns twenty results, but not a single answer.
    • Those who have been with the company for a long time know how things work. When this person retires, the knowledge leaves with them. How you can secure it beforehand is described in our advisory service on knowledge transfer before retirement.

    Company Knowledge: The Knowledge Base with AI

    Company Knowledge is a knowledge base that answers. You upload your documents via the interface as PDF, DOCX, Markdown, or text files. Employees ask their question in their own words, just as they would ask a colleague. The assistant searches for the relevant passages in your documents and formulates an answer from them.

    Well suited are, for example:

    • Handbooks and policies
    • Process descriptions and work instructions
    • Works agreements
    • Onboarding documents

    Questions can be asked in German and English, as well as in French and Italian. The answer is provided in the language of the question.

    Every answer with source citation

    Beneath every answer are the sources on which it is based. A single click opens the specific passage in detail. This allows employees to verify for themselves whether the answer is correct and read on if necessary. For HR, this means: the document remains authoritative. The assistant simply guides them there more quickly.

    Notice instead of an invented answer

    Language models tend to fill gaps with plausible-sounding sentences. Company Knowledge therefore displays a confidence level for every answer. If the basis in your documents is too thin, employees receive a notice instead of an invented answer. This is crucial for HR topics. Incorrect information regarding annual leave or notice periods creates more work than no answer at all.

    Permissions by role and individual

    Not every document is intended for everyone. You define who is permitted to use Company Knowledge: access can be switched on or off for each individual, and areas can be assigned to specific roles. The system verifies permissions server-side. Anyone not permitted to see an area will not receive any answers from it.

    Report knowledge gaps with 'Topic missing'

    If the assistant cannot find an answer, often a document is simply missing. Using 'Topic missing', employees can report this gap directly to those responsible. In this way, you close gaps in a targeted manner rather than drafting texts on speculation.

    Labelled as an AI response

    Every answer is clearly identifiable as an AI response. This is required by Art. 50 of the EU AI Act (KI-VO), which applies from 2 August 2026.

    Onboarding: New Employees Find Answers Themselves

    The first few weeks in a new job consist to a large extent of questions. New colleagues want to know where templates are stored, who to contact for IT issues, and how travel expense claims work. Company Knowledge answers such questions at any time from your onboarding documents. In addition, there are manageable onboarding checklists for new employees so that no step is forgotten.

    The tasks for HR, IT, facilities, and managers are managed within the personnel administration onboarding module. Find out more on the Onboarding Software page.

    Wiki Software or AI Assistant?

    Many companies already have a wiki. This is not a contradiction. A wiki is powerful when multiple people write and structure texts together. An AI assistant is powerful when people quickly need an answer from existing documents. The difference in everyday use:

    Wiki softwareAI assistant (Company Knowledge)
    SearchKeywords, result is a list of hitsQuestion in own words, result is an answer with sources
    MaintenanceContent is rewritten as wiki pagesExisting documents are uploaded
    Typical riskPages become outdated because no one maintains themAn answer summarises incorrectly, which is why sources and confidence levels are provided
    StrengthCollaborative writing and structuringFast access to knowledge contained in documents

    You can use a well-maintained company wiki as a source: export content, provided your wiki supports this, and upload it into Company Knowledge.

    Building a Knowledge Base in Five Steps

    Building a knowledge base with AI is primarily a matter of selection, rather than technology.

    1. Collect questions. For two weeks, note down what HR, IT, and managers are asked. This reveals which topics belong in there first.
    2. Select documents. Include only the current, valid version of each policy. Outdated versions confuse humans and AI alike.
    3. Designate owners. Every topic needs a person who keeps the document up to date and deals with reported gaps.
    4. Clarify permissions. Determine which areas everyone is permitted to see and which are restricted to specific roles, such as managers or HR.
    5. Start small and evaluate. Begin with one department or with onboarding. Reports via 'Topic missing' will show what is needed next.

    Involve the works council early on. An AI assistant used by employees may constitute technical equipment within the meaning of § 87 Abs. 1 Nr. 6 BetrVG (Works Constitution Act). What needs to be considered in this context can be found in our guide on the Works Council and AI.

    AI Knowledge Management: Data Protection and Hosting

    With AI knowledge management, internal documents are passed to a language model. It therefore matters where it runs and what happens to the content. With Company Knowledge, the AI operates in data centres in Germany or the EU. Your content is not used to train third-party models. For the techport HR Suite as a whole: operated on servers in Germany, on request as an on-premises installation at your company.

    You will receive the data processing agreement pursuant to Art. 28 GDPR (DSGVO) from us. What you should generally consider regarding AI applications is described in our guide on GDPR-compliant AI in Business.

    Part of the techport HR Suite

    Company Knowledge is one of five modules in the techport HR Suite. Core HR, recruiting, AI applicant evaluation, and Company Knowledge operate within a single platform with a shared database: one login, one interface, one search. Using the global search (Ctrl+K), you can ask a question to Company Knowledge directly from the search bar without switching modules.

    For the works council, there is a dedicated assistant that conducts research across legislation and your works agreements. Find out more on the Software for the Works Council page.

    See Company Knowledge in the Demo

    In the demo, we show you how Company Knowledge answers with sample documents, how sources and permissions work, and how the module interacts with core HR and recruiting. We provide pricing upon request, depending on headcount and modules.

    Frequently asked questions

    What is knowledge management software?

    Knowledge management software, often referred to as a knowledge management tool, consolidates a company's knowledge in one place and makes it discoverable. Traditional solutions are wikis or document repositories with search functionality. Newer solutions such as Company Knowledge in the techport HR Suite answer questions in natural language using AI and cite the sources from the uploaded documents.

    How can I build a knowledge base with AI?

    You build a knowledge base with AI by selecting valid documents, uploading them, and defining permissions for each role. More important than the technology is the curation: only up-to-date versions, a designated owner for each topic, and a way to report missing topics. Start with one area, such as onboarding, and expand gradually.

    Does the AI invent answers when it finds nothing?

    Company Knowledge is designed to provide a disclaimer rather than an invented answer when the factual basis is too thin. Every answer includes a confidence level and cites the sources on which it is based. Errors can never be entirely ruled out with language models. That is why employees can open each source passage and verify it themselves.

    Which documents can I upload?

    You can upload PDF, DOCX, Markdown, and text files via the user interface. Typical content includes handbooks, policies, process descriptions, works agreements, and onboarding documents. You determine which role can view which area. Access to Company Knowledge can also be toggled on or off on an individual basis.

    What distinguishes a wiki from an AI assistant?

    A wiki provides a list of pages in response to keywords, whereas an AI assistant answers a question in its own words. In a wiki, you write content from scratch; with an AI assistant, you upload existing documents. Both can be combined: a well-maintained wiki can serve as a source for the assistant if you export its content.

    Are our documents used to train AI models?

    No, content in Company Knowledge is not used to train third-party models. The AI operates in data centres in Germany or the EU. The techport HR Suite is hosted on servers in Germany, or optionally as an on-premises installation at your company. You will receive a data processing agreement pursuant to Art. 28 GDPR (DSGVO) from us.

    Further reading