Flow description
Purpose and benefits
The Topic Clustering Tool is designed to automate the organization of keyword lists into structured, easy-to-analyze tables based on topic clusters. This flow is especially useful for content strategists, SEO specialists, and marketers who frequently work with large sets of keywords and need to categorize them efficiently for content planning, site architecture, or campaign development.
How the Workflow Operates
1. User Engagement & Welcome
- When a user opens the chat interface, the workflow is automatically triggered.
- The user is greeted with a welcoming message that explains the tool’s purpose:
“Welcome to the Keyword Categorization Tool! I’m here to help you organize your keyword list into a structured table. Simply provide your list of keywords, and I’ll categorize them for you…”
- The user provides a list of keywords via the chat input.
- The tool can also access chat history, allowing it to consider previous interactions and user feedback, which is especially helpful for iterative refinement if the user wants further adjustments.
3. Prompt Creation for Clustering
- The workflow takes the user’s keywords and, together with any relevant chat history, uses a prompt template to formulate a clear instruction for an AI language model.
- The prompt explicitly asks the AI to:
- Assign topic clusters to the provided keywords.
- Return the result as a markdown table, with keywords in the first column and clusters in subsequent columns.
- Ensure the table is center-aligned for readability.
Prompt Example:
You are tasked with assigning topic cluster to keywords {input}. Result should be in a table where first column is named Keywords (with keywords in rows under), clusters should be in the second and further columns. Everything needs to be aligned to center. OUTPUT THE TABLE IN MARKDOWN FORMAT
4. Automated Table Generation
- The prompt is sent to a large language model (LLM) generator node.
- The AI processes the request and returns a markdown-formatted table categorizing the keywords.
5. Output Delivery
- The generated table is displayed in the chat as a response to the user.
- If the user is not satisfied, the process can iterate, utilizing chat history to refine the clustering based on feedback.
Workflow Structure Summary
Stage | Node/Component | Function |
---|
Chat Opened | ChatOpenedTrigger | Starts the workflow and triggers the welcome message |
Welcome Message | MessageWidget | Informs and guides the user |
User Input | ChatInput | Accepts keyword lists from the user |
Chat Memory | ChatHistory | Stores and provides previous messages for context/refinement |
Prompt Preparation | PromptTemplate | Formats the instruction for the AI, inserting keywords and context |
AI Generation | Generator | Uses LLM to cluster keywords and generate a markdown table |
Output Display | ChatOutput | Presents the resulting table to the user in the chat interface |
- Scalability: Automates the clustering of large keyword lists, saving significant manual effort.
- Consistency: Ensures a uniform approach to categorization, reducing human bias and error.
- Iteration: Incorporates feedback loops, allowing users to refine results based on previous attempts or clarifications.
- Presentation: Delivers results in markdown table format, making it easy to copy, share, or integrate into other workflows or documentation.
- User-Friendly: Guided onboarding and intuitive interaction make it accessible even for non-technical users.
Example Use Cases
- SEO Planning: Quickly organize hundreds of keywords into topic clusters for pillar content and supporting articles.
- Content Audits: Re-categorize existing keywords or topics to identify gaps and opportunities.
- Campaign Organization: Structure ad or content campaigns around logically grouped keyword sets.
By automating the clustering of keywords and presenting them in a structured, visually clear way, this workflow streamlines content organization and significantly boosts efficiency for anyone managing keyword-heavy projects.