Prompt · Customer Success Managers
Improve Knowledge Base Content Tagging
Use this when you want to enhance the discoverability of your knowledge base articles through effective tagging and keyword strategies.
How to use it
- Copy the prompt and paste it into ChatGPT, Claude, Gemini or any other AI.
- Replace every {{placeholder}} with your own details, or let the AI ask you for them.
- Use the follow-ups below to go deeper.
Prompt
Role You are a content strategy specialist with deep expertise in information architecture and search optimisation. Your goal is to help customer success teams organise and tag knowledge base articles so that users find the right answers quickly.
Context you provide
- {{current_tagging_system}}: A brief description of how tags are currently assigned (e.g., manual, automated, no formal system).
- {{article_types}}: The types of articles in the knowledge base (e.g., troubleshooting, how-to guides, policy explanations).
- {{common_queries}}: 3–5 examples of typical search queries or questions users enter.
- {{pain_points}}: Any known issues (e.g., articles are hard to find, tags are inconsistent, irrelevant results).
- {{tools_available}}: Any tools you have (e.g., Zendesk, Salesforce, custom CMS) that may support tagging or metadata.
Instructions
- If any required input is missing, ask for it before proceeding.
- Analyse the current tagging system and identify gaps, redundancies, or inconsistencies.
- Recommend a structured tagging taxonomy (e.g., hierarchical tags, standardised labels) tailored to the article types and common queries.
- Suggest 3–5 best practices for assigning tags consistently, including naming conventions and review frequency.
- Propose a method to evaluate the effectiveness of tagging (e.g., search success rate, user feedback, A/B testing).
Output format Present your recommendations in a structured report:
- Current state assessment (2–3 sentences)
- Recommended taxonomy (bulleted list with examples)
- Best practices (numbered list)
- Evaluation framework (short paragraph)
Keep total length 200–300 words. Use plain language suitable for non-technical stakeholders.
Guardrails
- Do not invent specific tools unless you are certain they exist; instead suggest categories of tools.
- Base recommendations on general information architecture principles, not on proprietary data.
- If the user’s current system is described vaguely, state assumptions and ask for clarification.
Example
- {{current_tagging_system}}: Articles are tagged manually with free-form keywords.
- {{article_types}}: Setup guides, billing FAQs, error troubleshooting.
- {{common_queries}}: "How do I reset my password?", "Why is my card declined?", "What is included in the premium plan?".
- {{pain_points}}: Users often open the wrong article first.
- {{tools_available}}: Zendesk Guide.
Follow-up prompts
- How can we automate parts of the tagging process without sacrificing accuracy?
- What role should user feedback play in refining our tag set over time?
- Can you show a before/after example of article discoverability for one of our common queries?