← Research, Analysis & Decisions

Classification · Categorization · Organization

How to Use AI to Classify and Categorize Information

Classification -- sorting items into categories based on defined criteria -- is one of the most tedious and time-consuming tasks in knowledge work. AI handles it at a scale and speed that would be impossible manually. People use it to categorize customer feedback by theme, sort survey responses into buckets, classify expenses by category, tag content for a CMS, label data for machine learning, sort leads by likelihood to convert, and organize any large body of unstructured information into a structured, analyzable form. The key to getting good results is defining your categories clearly upfront and giving AI a few examples of what belongs in each -- ambiguous categories produce inconsistent results, and explicit examples solve most of the ambiguity.

5 Best Prompts for Classifying Information to Ask Claude or ChatGPT

Copy any prompt below and paste it directly into your AI of choice.

  1. Prompt 01 · Categorize a list

    "Here is a list of [items: customer feedback / expenses / tasks / content pieces]: [paste list]. I want to categorize each one into these categories: [list categories with brief descriptions]. Can you go through each item, assign it to the most appropriate category, and flag any items that are ambiguous or do not fit cleanly?"

    Best for: turning an unorganized list into a structured, analyzable dataset.

  2. Prompt 02 · Define categories from the data

    "Here is a body of unstructured information: [paste data -- feedback, responses, notes]. I do not have predefined categories. Can you analyze it and suggest the most useful category structure -- the groupings that would capture meaningful patterns -- and then classify each item into the categories you define?"

    Best for: exploratory classification when you do not know your categories in advance.

  3. Prompt 03 · Tag content

    "Here are [articles / posts / documents]: [paste or describe]. I need each one tagged with relevant labels from this taxonomy: [list tags]. Please assign up to [number] tags per item, ordered by relevance, and explain your reasoning for any tags that are not immediately obvious."

    Best for: content tagging at scale for search, filtering, or recommendation systems.

  4. Prompt 04 · Sort and prioritize

    "Here are [leads / tasks / opportunities / applications]: [list them with relevant details]. Using these criteria: [list criteria and their relative weight], can you sort them from highest to lowest priority and briefly explain why each one landed where it did?"

    Best for: any situation where you need to prioritize a large number of items against defined criteria.

  5. Prompt 05 · Identify themes in qualitative data

    "Here are [customer reviews / interview responses / open-ended survey answers]: [paste]. Can you identify the main themes that appear across multiple responses, quote specific examples for each theme, and tell me how frequently each theme appears? I want to understand what people are actually saying at a pattern level."

    Best for: making sense of large amounts of qualitative feedback without reading every response individually.