# What is keyword extraction from a transcript? > Keyword extraction pulls the words and phrases that come up most in a transcript or document, with counts. What it shows, what it misses, how to check it. Source: https://learn.ynteractive.com/content/glossary/keyword-extraction · Updated 2026-10-19 · Free from Stepthrough (https://learn.ynteractive.com) [Data, Sheets and Gmail](https://learn.ynteractive.com/content/glossary#topic-data-sheets-and-gmail) · [AI glossary](https://learn.ynteractive.com/content/glossary) Quick answer Keyword extraction is when software scans a transcript or document and lists the important words and phrases in it, usually with a count of how often each appears. Some tools also pull out the people, companies and places named, so you can see what a conversation was about at a glance. Last updated October 19, 2026 [Take the free module →](https://learn.ynteractive.com/courses/ai-apps-worth-knowing/31)Module 31 of the free [AI Apps Worth Knowing](https://learn.ynteractive.com/courses/ai-apps-worth-knowing) course, about 10 minutes. No account needed. Updated October 19, 2026 · By [Robert Breen](https://learn.ynteractive.com/content/about) ## Why it matters for a small business Reading every transcript in full does not scale. A team with twenty recorded client calls a month cannot read all of them, and skimming misses the quiet patterns. A keyword list with counts is a thirty-second scan: if "bank statements" and "first invoice" sit near the top of several calls, you know where to look without reading a page. The catch is that keywords are words, not meaning. A count tells you a topic came up, not whether the client was happy or angry about it, and extraction tools guess when they label names and brands. Treat the list as a map that tells you which parts of the transcript to read, then read those parts. ## In a real lesson: Speak AI: Turn Customer Interviews into Themes and Quotes In the [Speak AI lesson](https://learn.ynteractive.com/content/speak-ai-turn-customer-interviews-into-themes-and-quotes), you open Insights on Harbor Lane Accounting's 2:58 interview and click the Keywords tab. It shows **Keywords (20)**, each with a count: First week (3), Onboarding (2), Paperwork, Welcome email, Document portal, Bank statements, First invoice and more. Below that, **People (3)** lists Mike, Dana and Priya, the bookkeeper. **Brands** lists Harbor Lane Accounting, and also Fair, picked up from the phrase "It was fair". That one mistake is the lesson's reason to check what the AI finds. Across five interviews, the keywords that repeat point you to the [themes](https://learn.ynteractive.com/content/glossary/thematic-analysis) worth reading about in the [transcript](https://learn.ynteractive.com/content/glossary/meeting-transcript). [Try this lesson free](https://learn.ynteractive.com/modules/speakai-customer-interview-themes) or [read the step-by-step guide](https://learn.ynteractive.com/content/speak-ai-turn-customer-interviews-into-themes-and-quotes). ## Common confusions ### Keyword extraction vs thematic analysis Keyword extraction counts words. [Thematic analysis](https://learn.ynteractive.com/content/glossary/thematic-analysis) groups ideas, even when people use different words for the same problem, such as "paperwork" and "too many attachments". ### Keyword extraction vs SEO keywords SEO keywords are the searches you want a web page to rank for. Transcript keywords are simply the phrases that came up most in a conversation. ## Tips - Compare keyword lists across several calls; a word that tops one call may be noise. - Check the people and brand lists before you rely on them, since the tool guesses. - Click into the transcript around a keyword to see how it was said. ## Related terms [Thematic analysis](https://learn.ynteractive.com/content/glossary/thematic-analysis) · [Sentiment analysis](https://learn.ynteractive.com/content/glossary/sentiment-analysis) · [Speech-to-text](https://learn.ynteractive.com/content/glossary/speech-to-text) · [Meeting transcript](https://learn.ynteractive.com/content/glossary/meeting-transcript) · [Voice of the customer (VoC)](https://learn.ynteractive.com/content/glossary/voice-of-the-customer) ## More Data, Sheets and Gmail terms [Merge field](https://learn.ynteractive.com/content/glossary/merge-field) · [Parsing rules](https://learn.ynteractive.com/content/glossary/parsing-rules) · [Pivot table](https://learn.ynteractive.com/content/glossary/pivot-table) · [Plain-text email](https://learn.ynteractive.com/content/glossary/plain-text-email) ## Where you use it: free lessons - [Speak AI: Turn Customer Interviews into Themes and Quotes](https://learn.ynteractive.com/content/speak-ai-turn-customer-interviews-into-themes-and-quotes) (Speak AI, 10 min) ## AI tools that use it - [Speak AI](https://learn.ynteractive.com/content/ai-tools/meetings-notes#tool-speak-ai) [More terms by tool](https://learn.ynteractive.com/content/glossary/ai-tool-terms) ## Frequently asked questions **How does keyword extraction work?** The software looks for words and phrases that appear often or stand out in the text, ranks them, and usually shows a count for each. **Can keyword extraction find themes?** It points to them. Repeated keywords across calls are a good clue, but grouping them into themes still takes reading or an AI question about the transcript. **Why did the tool tag a normal word as a brand?** Name and brand detection is a guess. In the Speak AI lesson, "fair" from "It was fair" was listed as a brand, so a quick check is part of the job. [All AI glossary terms, A to Z](https://learn.ynteractive.com/content/glossary) · [Free prompt templates](https://learn.ynteractive.com/content/prompts)