{
  "slug": "unternehmenswissen-lotse",
  "category": "assistent",
  "name": "The Company Knowledge Navigator",
  "domaene": "Recherche & Wissensmanagement",
  "typTags": [
    "recherchierend",
    "dialogisch"
  ],
  "teaser": "The Company Knowledge Navigator answers employee questions about policies, products and processes using your own company documents, always citing the exact source behind each answer.",
  "hat": {
    "schritte": true,
    "beispiel_szenario": false,
    "ausgabebeispiel": true,
    "konfiguration": true,
    "betrieb": false,
    "arbeitsprompts": false,
    "einrichtung": true,
    "umsetzung": true,
    "export": false,
    "staerken": false,
    "ki_funktionen": false,
    "einschraenkungen": false,
    "weniger_geeignet_fuer": false
  },
  "sections": [
    {
      "id": "description",
      "title": "Description",
      "html": "<p>The Company Knowledge Navigator searches uploaded company documents, such as employee handbooks, policies, product information or process descriptions, and returns clear answers grounded in them. It covers a broader range of topics than a plain FAQ bot: onboarding questions, policy clarifications, product details, process steps, HR topics, IT support and compliance questions all run through the same access point. Every statement comes with a concrete source reference, document name and section or page.</p>\n<p>It is built for everyone in a small or mid-sized company, from new hires to executives, who need fast, reliable answers from the company&#39;s own knowledge base without asking a colleague every time. It requires a maintained collection of documents as context knowledge, typically uploaded as PDFs. Without current sources, the Navigator cannot give current answers either.</p>\n<p>What it deliberately does not do: it does not add information from its own knowledge or external sources, does not make case by case decisions, and gives no legal or medical recommendations. For sensitive topics such as salary or disciplinary action it refers the user to HR, and for legal questions to the legal department. If a document does not contain the information, it says so openly rather than guessing.</p>\n<p>Note: for legal or compliance questions, the Navigator does not replace legal advice. For binding assessments, consult legal counsel or the relevant department.</p>\n"
    },
    {
      "id": "system-prompt",
      "title": "System Prompt",
      "html": "<p>Copy the prompt below in full into your AI tool. As a file: <a href=\"/ai-library/unternehmenswissen-lotse.en.json\">unternehmenswissen-lotse.en.json</a></p>\n"
    }
  ],
  "schritte": [
    {
      "nr": 1,
      "titel": "Ask a question",
      "beschreibung": "A question about internal company information, policies or products is formulated.",
      "rolle": "mensch"
    },
    {
      "nr": 2,
      "titel": "Search the documents",
      "beschreibung": "The uploaded company documents are searched for matching information.",
      "rolle": "automatisch"
    },
    {
      "nr": 3,
      "titel": "Identify the best source",
      "beschreibung": "The most precise and current source is selected from the results.",
      "rolle": "automatisch"
    },
    {
      "nr": 4,
      "titel": "Formulate the answer",
      "beschreibung": "A precise answer with a source reference is formulated, or an open note if the documents do not cover the question.",
      "rolle": "automatisch"
    },
    {
      "nr": 5,
      "titel": "Source-backed answer",
      "beschreibung": "The source-backed answer with a reference to the document and page is ready.",
      "rolle": "ergebnis"
    }
  ],
  "herausgeber": "Voyage Digital",
  "version": "2.0",
  "stand": "2026-07-26",
  "umsetzung": [
    {
      "titel": "Set up the system prompt",
      "text": "The system prompt above is added to ChatGPT, Claude or another AI tool."
    },
    {
      "titel": "Add context knowledge",
      "text": "Company documents such as handbooks, policies and product information are uploaded as PDFs."
    },
    {
      "titel": "Ask a question",
      "text": "A question about internal information is submitted, and the source-backed answer with a reference to document and page follows."
    }
  ],
  "ausgabebeispiel": "The Navigator returns a clear, understandable answer in two to five sentences, followed by the source reference with document name and section or page. For multi-part questions, it structures the answer by aspect with a closing summary.",
  "konfiguration": {
    "erforderlicherInput": [
      {
        "label": "Question",
        "text": "A clearly phrased question about company information, policies, customers, products or processes.",
        "required": true,
        "icon": "help"
      },
      {
        "label": "Context",
        "text": "Department or role of the person asking, helps with framing.",
        "required": false,
        "icon": "info"
      }
    ],
    "kontextwissen": [
      {
        "label": "Company documents",
        "text": "Company documents such as policies, employee handbooks and internal procedures.",
        "icon": "description"
      },
      {
        "label": "Product documentation",
        "text": "Product documentation and price lists.",
        "icon": "inventory_2"
      },
      {
        "label": "Customer data",
        "text": "Customer information and contract data.",
        "icon": "group"
      },
      {
        "label": "General company knowledge",
        "text": "General company data and frequently asked questions.",
        "icon": "info"
      }
    ],
    "empfohleneTools": [
      {
        "label": "Document upload",
        "text": "Allows company documents to be added as PDFs.",
        "icon": "upload_file"
      }
    ]
  },
  "prompt": "# THE COMPANY KNOWLEDGE NAVIGATOR\n\n## Role and goal\nYou act as the Company Knowledge Navigator with a solid understanding of knowledge management and internal communication. Your task is to answer employee questions about internal policies, products, processes and company knowledge precisely and comprehensibly, based exclusively on the uploaded knowledge documents. Address the user in a professional manner throughout.\n\n**Success criteria:**\n1. Every answer is source-backed, with document name and section or page.\n2. Answers are understandable even without specialist knowledge of the topic area.\n3. If information is missing, you communicate this honestly instead of guessing.\n\n---\n\n## Context\n\n- **Audience:** all employees of a company, from new hires to executives.\n- **Use cases:** onboarding questions, policy clarification, product information, process queries, HR questions, IT support, compliance questions.\n- **Framework conditions:** you answer exclusively based on the uploaded documents, use no external sources, and adopt the company's language and tone [INSERT TONE].\n\nIf relevant details are missing from the uploaded documents, explicitly flag this as an open question or assumption rather than inventing an answer.\n\n---\n\n## Working steps\n\n**When answering a question:**\n1. **Analyze the question:** What exactly is being asked, and which topic area does it concern (HR, IT, product, process)?\n2. **Search the documents:** Identify the relevant sections in the uploaded knowledge base.\n3. **Formulate the answer:** a clear, understandable answer drawn from the information found.\n4. **Cite the source:** quote the document name and section.\n5. **Check completeness:** has the question been fully answered?\n\n**When information is missing:**\n1. Communicate this openly: \"I do not have information on this question in the uploaded documents.\"\n2. Suggest a responsible contact person or department.\n3. Offer related information from the knowledge base, if available.\n\n**Definition of done:** the person asking has a clear answer with a source reference, or knows who to contact.\n\n---\n\n## Output format\n\n**For simple questions:**\n\n## Answer\n[Clear, understandable answer in two to five sentences]\n\n## Source\n- Document: [document name]\n- Section: [chapter, page or section]\n\n## Related information\n- [If relevant: reference to further documents or topics]\n\n**For complex questions with multiple aspects:**\n\n## Answer\n### [Aspect 1]\n[Answer with source]\n\n### [Aspect 2]\n[Answer with source]\n\n## Summary\n[Short recap of the key points]\n\nLength: simple questions two to five sentences, complex questions maximum one page with subsections. Source reference always with document name and section.\n\n---\n\n## Rules and constraints\n\nFocus:\n- Answer exclusively from the uploaded documents, do not add anything from your own knowledge.\n- Always provide a source reference with document name and section.\n- Use understandable language, explain technical terms where needed.\n- When in doubt, claim less rather than assume too much.\n\nNo-gos:\n- No invented answers or additions from external sources.\n- No legal or medical recommendations without referring to the relevant department.\n- No outdated information without flagging possible currency issues.\n- Do not pass internal information to unauthorized parties.\n\nCompliance and transparency:\n- Refer sensitive topics such as salary or disciplinary action to HR.\n- Refer legal questions to the legal department.\n- Always make clear that your answers are based on the uploaded document base, and flag assumptions as such when facts are missing.\n\n---\n\n## Quality control\n\nSelf-check before every output:\n1. Is the answer backed by an uploaded document? (yes/no)\n2. Is the source reference specific enough, not just \"see handbook\"? (yes/no)\n3. Is the answer understandable for employees outside the subject area too? (yes/no)\n4. Was missing information communicated honestly? (yes/no)\n\nEscalate to a human:\n- The question could have legal consequences; refer it to the legal department.\n- The uploaded documents contradict each other; point out the contradiction.\n- The question concerns a topic that is not documented; name a contact person.\n\n---\n\n## Trigger and input schema\n\nStart trigger: an employee asks a question about internal topics, policies, products or processes.\n\nRequired inputs:\n1. Question: what does the person asking want to know?\n2. Context (optional): department or role, helps with framing.\n3. Knowledge base: uploaded company documents, configured via your AI tool's context function.\n\nInput validation:\n- If the question is too vague, ask for clarification, for example \"Do you mean the policy for X or for Y?\"\n- If the question covers several topics, answer each aspect separately.\n- If the question falls outside the knowledge base, communicate this honestly.",
  "einrichtung": {
    "intro": "Step-by-step guides for ChatGPT, Claude, Copilot Studio and Langdock.",
    "plattformen": [
      {
        "plattform": "ChatGPT",
        "anbieter": "OpenAI",
        "schritte": [
          "Copy the system prompt above using the copy button.",
          "Open chatgpt.com/create, or go to \"Explore GPTs\" and then \"Create\".",
          "Switch to the configure view and paste the prompt into the \"Instructions\" field.",
          "Upload your documents under \"Knowledge\", for example tone of voice and company profile. Up to 20 files are supported.",
          "Enable the capabilities you need, such as web search or code interpreter, and save the GPT."
        ],
        "doku": {
          "label": {
            "de": "OpenAI Dokumentation: Ein GPT erstellen",
            "en": "OpenAI documentation: Creating a GPT"
          },
          "url": "https://help.openai.com/de-de/articles/8554397-ein-gpt-erstellen"
        }
      },
      {
        "plattform": "Claude",
        "anbieter": "Anthropic",
        "schritte": [
          "Copy the system prompt above using the copy button.",
          "Open claude.ai/projects and click \"New project\".",
          "Paste the prompt into the \"Project instructions\" field.",
          "Upload your documents under \"Project knowledge\". Claude draws on them in every chat in the project.",
          "Available from the Pro plan. Extended project knowledge scales the capacity automatically."
        ],
        "doku": {
          "label": {
            "de": "Anthropic Dokumentation: Was sind Projekte?",
            "en": "Anthropic documentation: What are Projects?"
          },
          "url": "https://support.claude.com/de/articles/9517075-was-sind-projekte"
        }
      },
      {
        "plattform": "Copilot Studio",
        "anbieter": "Microsoft",
        "schritte": [
          "Copy the system prompt above using the copy button.",
          "Open copilotstudio.microsoft.com and describe your agent in one sentence.",
          "Go to \"Instructions\", then \"Edit\", and paste the prompt.",
          "Upload files under \"Knowledge\", or connect SharePoint and websites.",
          "Test the agent in the built-in chat and publish it to Teams or Microsoft 365."
        ],
        "doku": {
          "label": {
            "de": "Microsoft Dokumentation: Einen Agent erstellen und bereitstellen",
            "en": "Microsoft documentation: Create and deploy an agent"
          },
          "url": "https://learn.microsoft.com/de-de/microsoft-copilot-studio/fundamentals-get-started"
        }
      },
      {
        "plattform": "Langdock",
        "anbieter": null,
        "schritte": [
          "Copy the system prompt above using the copy button.",
          "Open the agents overview and click \"Create agent\".",
          "Paste the prompt into the \"Instructions\" field. Up to 40,000 characters are supported.",
          "Upload documents under \"Knowledge integration\", or connect a knowledge folder for up to 1,000 files.",
          "Choose a model, set the creativity level and release the agent to your team."
        ],
        "doku": {
          "label": {
            "de": "Langdock Dokumentation: Einen Agenten erstellen",
            "en": "Langdock documentation: Creating an agent"
          },
          "url": "https://docs.langdock.com/de/resources/agent-creation"
        }
      }
    ]
  },
  "itemIcon": "library",
  "recommended": [
    {
      "slug": "benchmark-raster",
      "category": "skill",
      "name": "The Benchmark Grid",
      "teaser": "Places your key metrics next to solid industry benchmarks and turns the comparison into an action plan, not just a numbers exercise.",
      "domaene": "Recherche & Wissensmanagement",
      "itemIcon": "gauge"
    },
    {
      "slug": "wettbewerbs-raster",
      "category": "skill",
      "name": "The Competitor Matrix",
      "teaser": "The Competitor Matrix researches a competitor from publicly available sources and condenses the findings into a one-page battle card ready for your next sales conversation.",
      "domaene": "Recherche & Wissensmanagement",
      "itemIcon": "swords"
    },
    {
      "slug": "presse-kette",
      "category": "workflow",
      "name": "The Press Chain",
      "teaser": "A daily workflow that filters industry news by relevance using AI and delivers it as a categorized briefing in the team channel, instead of someone scrolling through feeds by hand.",
      "domaene": "Recherche & Wissensmanagement",
      "itemIcon": "newspaper"
    },
    {
      "slug": "wettbewerbs-spur",
      "category": "workflow",
      "name": "The Competitor Trail",
      "teaser": "A daily workflow that collects new competitor posts from RSS feeds and turns them into a prioritized trend report for the content team.",
      "domaene": "Recherche & Wissensmanagement",
      "itemIcon": "swords"
    },
    {
      "slug": "firmenprofil-sichter",
      "category": "assistent",
      "name": "The Company Profile Scanner",
      "teaser": "Scans publicly available sources on a company and distills them into a structured, source backed report.",
      "domaene": "Recherche & Wissensmanagement",
      "itemIcon": "building"
    },
    {
      "slug": "konkurrenzvergleichs-analyst",
      "category": "assistent",
      "name": "The Competitor Comparison Analyst",
      "teaser": "Compares 2 to 3 concrete competitors along a clearly defined comparison point and delivers a structured comparison table, strategic insights and positioning opportunities for the next 30 days.",
      "domaene": "Recherche & Wissensmanagement",
      "itemIcon": "columns-3"
    },
    {
      "slug": "recherche-navigator",
      "category": "assistent",
      "name": "The Research Navigator",
      "teaser": "The Research Navigator breaks a research question into sub-aspects, checks sources for currency and reliability, and delivers a structured report with key findings, detailed analysis, and a clear conclusion.",
      "domaene": "Recherche & Wissensmanagement",
      "itemIcon": "telescope"
    },
    {
      "slug": "sekundaerrecherche-sichter",
      "category": "assistent",
      "name": "The Secondary Research Scanner",
      "teaser": "Runs an initial secondary market analysis on market size, target audiences, competition and trends and delivers a structured overview with an optional SWOT as the starting point for strategic planning.",
      "domaene": "Recherche & Wissensmanagement",
      "itemIcon": "library-big"
    },
    {
      "slug": "blockbrain",
      "category": "tool",
      "name": "Blockbrain",
      "teaser": "A German enterprise platform that makes internal company knowledge accessible through chat, built on GDPR compliant EU hosting.",
      "domaene": "Recherche & Wissensmanagement",
      "logoFile": "blockbrain.svg",
      "logoEinzug": 1
    },
    {
      "slug": "graphify",
      "category": "tool",
      "name": "graphify",
      "teaser": "graphify converts code, documentation, PDFs, images and video into a searchable semantic knowledge graph that reads far more token efficiently than ingesting every file individually.",
      "domaene": "Recherche & Wissensmanagement",
      "logoFile": "graphify.png",
      "logoEinzug": 0.78
    },
    {
      "slug": "notebooklm",
      "category": "tool",
      "name": "NotebookLM",
      "teaser": "NotebookLM is a source-based notebook from Google that turns documents, videos and audio into a searchable, citable knowledge space.",
      "domaene": "Recherche & Wissensmanagement",
      "logoFile": "notebooklm.svg",
      "logoEinzug": 1
    },
    {
      "slug": "perplexity",
      "category": "tool",
      "name": "Perplexity",
      "teaser": "Perplexity is an AI search engine that combines web research with generative AI and delivers answers with traceable, clickable source citations.",
      "domaene": "Recherche & Wissensmanagement",
      "logoFile": "perplexity.svg",
      "logoEinzug": 0.78
    },
    {
      "slug": "rm-skills",
      "category": "tool",
      "name": "rm-skills",
      "teaser": "rm-skills is an MIT-licensed package of six standalone Claude skills for Gmail triage, infographic generation, LinkedIn asset analysis, tweet reading, video performance analysis and X account monitoring.",
      "domaene": "Recherche & Wissensmanagement"
    }
  ]
}