{
  "patent_number": "US 11521601",
  "country": "US",
  "title": "How AI Cleans Up Irrelevant Topics in Recorded Phone Calls",
  "original_title": "Detecting extraneous topic information using artificial intelligence models",
  "summary": "A system that automatically identifies and removes 'noisy' or irrelevant topics from call center transcripts by analyzing how consistently and broadly those words appear.",
  "what_it_does": "This patent describes a way to make AI-powered call analysis more accurate by filtering out 'extraneous' topics that don't provide useful business insights. It works by looking at call transcripts and calculating two specific metrics for each identified topic: a vocabulary score (based on the entropy of word distributions) and a consistency score (how often that topic appears across different calls). If a topic is too broad or lacks consistency, the system flags it as noise. It then automatically updates the AI model to either remove that topic from the results or visually separate it, ensuring that human analysts only see the most relevant conversation data.",
  "what_it_does_not_cover": [
    "Does not cover general speech-to-text transcription methods.",
    "Does not cover manual topic tagging performed by human agents.",
    "Does not cover sentiment analysis or emotion detection in calls.",
    "Does not cover real-time audio processing (it requires pre-stored transcripts)."
  ],
  "filed": "2020-08-18",
  "granted": "2022-12-06",
  "expires": null,
  "status": "active",
  "holder": "Invoca Inc",
  "holder_url": "https://patentbrief.org/company/invoca-inc",
  "inventors": [
    {
      "name": "Michael McCourt",
      "url": "https://patentbrief.org/inventor/michael-mccourt"
    },
    {
      "name": "Michael Lawrence",
      "url": "https://patentbrief.org/inventor/michael-lawrence"
    }
  ],
  "times_cited": 1,
  "tags": [
    "ai_ml",
    "software",
    "telecommunications"
  ],
  "abstract": "Systems and methods for improving machine learning systems used to model topics on a plurality of calls are described herein. In an embodiment, a server computer receives plurality of digitally stored call transcripts that have been prepared from digitally recorded voice calls. The server computer uses a topic model of an artificial intelligence machine learning system, the topic model modeling words of a call as a function of one or more word distributions for each topic of a plurality of topics, to generate an output of the topic model which identifies the plurality of topics represented in the plurality of call transcripts. The server computer computes, for a particular topic of the plurality of topics a first value representing a vocabulary of the particular topic and a second value representing a consistency of the particular topic in two more call transcripts of the plurality of call transcripts which include the particular topic. Based, at least in part, on one or more of the first value or the second value, the server computer determines that the particular topic meets a particular criterion and, in response, updates the output of the topic model to remove the particular topic or distinguish the particular topic from other topics of the plurality of topics which do not meet the particular criterion.",
  "url": "https://patentbrief.org/patent/us/11521601/stylegan",
  "markdown_url": "https://patentbrief.org/patent/us/11521601/stylegan/md",
  "google_patents_url": "https://patents.google.com/patent/US11521601",
  "relatedPatents": []
}