MEETINGINSIGHTNLPMLExtract Insights from meetings using Exemplary AI

John Jacob

John Jacob

November 7, 2022·6 min read

In today's remote friendly world, communication is key and online meetings have become central to enabling remote teams to collaborate, share ideas and make important decisions regardless of location. It's also great to record meetings because team members may not be able to attend all their meetings or may need a recap of past meetings to surface context. Further more, writing a short description or a summary improves context for all involved.

Here at Exemplary, we have automated the flow to process meetings by helping you store and process recordings, view and edit transcripts, and also extract valuable insights from them through use of Automated Speech Recognition, Natural Language Processing, Large Language Models. We also enable you to share them with ease. All you have to do is, send our bot to a meeting or upload meetings onto our platform using our API, and access them via our dashboard where insights would be automatically generated. Exemplary AI also provides a searchable index for historical context mining, if you choose to enable it.

5 Simple Steps to Access Valuable Insights via Exemplary AI

1. Access our Dashboard via our API.
2. Upload your recordings onto our Exemplary AI platform, either manually, or connect with your favorite Video Conferencing tool’s integration (Zoom, Webex etc.), and pull in all your meeting recordings.
3. Transcribe meetings using top ASR providers (like Deepgram).
4. Apply NLP and NLU processes like summary, action items, sentiment, meeting notes, keywords, and generate highlight reels.
5. Share insights via email or integrate with your preferred app (eg. Slack).

Types of Insights

1. Conversational Insights
2. Customer Insights
3. Speaker, or Participant Insights
4. Actionable Insights
5. Custom Insights

• Conversational Insights

  Generate AI-driven insights automatically, as well as using workflows involving humans.

  • Summary Generation
    Meeting summaries are important points to remember in the transcript, neglecting chatter and focusing on relevant information.

  • Create Key Highlights
    Highlights are important points discussed in a meeting. Using the tagging feature in the editor tool kit, you can select and create key highlights, and store them as insights.

  • Automated Notetaking
    Taking manual notes during a meeting makes it hard for the participant to concentrate on the meeting at hand. Use our summaries and extracted insights to ease the process and compile better notes.
    For Example: In sales, notes help you understand the customer’s goals better, previously stated pain points etc. streamlining your sales process.

  • Topics discussed
    Extracting the main topics discussed in a transcription, processed by an NLP model, from an audio or video recording.

    For Example: Consider a recorded meeting between Jessica and James. The Topics discussed were predominately about Strategy, Sales, Growth, and Project costs. The NLP model can not only detect topics, but also show you who the segment of the conversation where the topic was discussed

  • Entity Detection
    Entity Detection and Named Entity Extraction (NER), is a data extraction technique that identifies key entities such as company names, product names, dates, events, locations, etc. from the transcription. Identify commonly recurring entities for further analysis.
    For Example: Splunk is an entity, and is classified as a organization.

  • PII Redaction
    Automatically detect and remove Personally Identifiable Information (PII) content such as name, address, bank account details, SSN, etc. from your transcription.

• Customer Insights

  Extract Customer Insights from your meeting recordings, where customer share their expectations, experiences, and preferences and provide feedback. Knowledge collected from these meetings enables you to provide individually catered, and timely interactions with customers. Improper assessment of customer satisfaction may lead to customer churn.

  • Customer Sentiment
    Sentiment Analysis can help detect emotional value in conversations, by retrieving positive, negative, and neutral statements in the transcript, which can be considered as valuable feedback.
    For Example: Calculate the average call sentiment from the list of a customer's meeting recordings.
  • Customer vs Sales Rep
    Calculate the talk-time ratio between the customer and the sales representative. For example, for sales calls, you might want your reps to strive to achieve an 80/20 listen-to-talk ratio. We can help you make such decisions, reflect on performance and tune your conversations to achieve the best outcomes.
  • Time Keeping
    Not starting or ending the meeting at the scheduled time will result in a waste of the customer’s time. Keep track of overall meeting timing, and get gentle reminders on how far the meeting has progressed or time left.
  • Keywords
    Keywords can be the most commonly used words, terms, and names of products or companies. Using these keywords, a Word cloud is created to help you visualize their frequency and enables you to make strategic decisions with your meeting data.
  • Customer trends and Churn analysis
    Business insights such as customer trends, business risks, revenue generation opportunities, user anomalies, etc. are discussed regularly in customer meetings and can be extracted from the noise by using our NLP and NLU solutions. Be several steps ahead of your customers by leveraging customer empathy and being able to easily pull up feedback and insights from customer conversations.

When extracting insights from conversations, we build a searchable knowledge database of customer information. Using this search feature, you can easily sort through preferred topics discussed, customer sentiments, keyword occurrences, highlighted moments etc.

• Speaker Insights

  Speaker insights are derived from the knowledge collected from speaker related data.

  • Track Speaker Activity
  • Observe Speaker Engagement
  • Identify Speaker Trends

  

• Actionable Insights

  Use Machine Learning (ML) Techniques to extract actionable information from a transcript. This data can go into your CRM or your calendar and can be used to create reminders and action items so follow-up tasks don't slip by you during those busy weeks.

  • Action items
    Detect action items in the transcript of a recorded meeting. The tasks that come up during the meetings are automatically noted, making for easy follow-up and tracking of tasks mentioned during calls. Detect commitments and set reminders automatically, freeing your mind to focus on the conversations you have. For example: "Please review this report by the end of the day" gets noted as an action item, and can be set up as a reminder.
  • Follow-ups
    Detect mentions of follow-up activities and meetings in the transcript of a recorded meeting, allowing you to keep track of follow-ups required and ensuring you never miss talking to all the key stakeholders or necessary follow-ups.
  • Decisions
    Decisions are made by observing the data derived from a generated insights report. The system analyzes your conversation and extracts key decisions made during the call.

• Custom Insights

  Generate custom insights via the transcript editor by creating and tagging snippets from the conversation. You can even compile and share created highlight compilations, for sharing with your teams for easy review and decision-making. Once you've annotated the transcription enough times, we can also automatically detect these highlights for you in the future.

Use Cases

We delve into a couple of use cases:

1. Sales and Marketing Teams
2. Educational platforms

• Sales and Marketing Teams
  Sales and Marketing Teams are part of multiple customer meets throughout the day. Sometimes, the team might lose track of key points discussed in previous meetings. Utilize insights, and make strategic sales decisions using those insights, bringing in more clarity and improving your sales flow and process.

  With Meeting Insights, you can understand your customers better.

  • Track the purpose of the meeting.
  • Identify key takeaways from Customer stories.
  • Easily capture and organize all meeting-generated content, from sales training sessions to customer calls.

  You can even use our editor to create compilations of your best performing teams, highlighting key moments where they successfully pushed the meeting towards achieving their key performance metrics. These compilations can serve as highly useful training material and help onboard newer team members, enabling them to quickly empathize with your customers and understand the domain.

• Educational platforms
  In today’s world, many Instructors teach their students by utilizing educational platforms such as virtual classrooms, Massive Open Online Courses (MOOCs), and remote learning platforms. Instructors manage students across all levels and grades. Using our platform, instructors, and students can easily revisit topics, highlights, and keywords discussed in lectures or tutoring sessions, making it easy to track progress and review topics. Instructors can create structured learning paths and highlight key moments for students, and students can sort through lectures for easy review and create their own study plans based on their weaknesses and strengths. For students who missed a class or want a recap, they can easily access the transcript via Exemplary AI’s dashboard. Reach out to us for white-labeling solutions if you want to incorporate our platform into your classroom or your organization's learning system.

Conclusion

Watching a meeting recording or reviewing the whole transcript of a meeting can be taxing and time-consuming. Taking manual notes during a meeting makes you lose focus on the important discussions at hand.

With Exemplary AI’s approach to Meeting Insights, you can use our cutting-edge ML techniques to seamlessly extract knowledge and generate insight based on your needs.

Check out our other blogs for more interesting content.