Can AI be used to analyze and derive actionable insights from group discussions?
Novigi is an Australia-based leading digital transformation provider focused on the finance sector. We have been providing them a range of services from AI strategy, to solutions, over time.
While working with them, we tackled a tedious task many businesses, research organizations, government agencies and non-government entities are all too familiar with: deriving useful insights from a group discussion. Is there a faster, more efficient way to gain insights from a group discussion, stakeholder meeting, or a conference, without having to go through recordings or transcriptions one by one, analyzing the content for days or even weeks?
Deriving insights with natural language processing
We knew that this was a problem that had the potential to be addressed with a natural language processing (NLP) pipeline. We started building this pipeline for Novigi, and our first step was transcribing audio recordings, which, in this setting, was our source data. We did this with the help of Amazon Transcribe, one of the more mature speech to text services available. Then, we proceeded to develop an NLP pipeline with the help of state of the art NLP libraries to perform analyses. This pipeline identifies the most spoken words or phrases during the discussion, using word frequencies. It also recognizes named entities, and groups them into different categories. When combined, these results uncover valuable insights about the discussion.
Beyond that, sentiment analysis with its ability to gauge whether a particular statement is positive or negative, is at the core of our NLP pipeline. It can determine the overall sentiment of the participants about the discussion, as well as on specific discussion topics. For instance, if the discussion is on tackling climate change, it is possible to understand whether the overall discussion sentiment towards tackling climate change is positive or negative. If there is a specific discussion point on the contribution of the meat industry to climate change, it is again possible to understand the overall sentiment of the participants on this topic -- whether they think the meat industry contributes to climate change or not.
We also used topic modeling in our pipeline, a technique which clusters similar terms together based on certain patterns in words or phrases -- these clusters could then be interpreted as various topics discussed to derive valuable insights from unstructured text.
Scales beyond human capacity
We found that our NLP pipeline was highly effective in providing a fast analysis of the discussions, immediately after their conclusion. For instance,we obtained results which said that the overall sentiment towards the discussion was positive, and that a very high percentage of the participants had a positive sentiment towards the topic in discussion.
Furthermore, our NLP pipeline made it possible for Novigi to strategically position themselves as a company that uses bleeding edge technology. With our AI solution, it is possible to cut back on time spent on manually analyzing data from discussions from days or weeks, to minutes or hours. In fact, the analyses go beyond human capacity, as a human will not be able to perform all the tasks from automated transcription to sentiment analysis, at scale.
If you are looking to find similar solutions for existing challenges in your day to day operations, do reach out to us. Our team of AI experts would be keen to work with you to develop cutting edge AI solutions.
Can AI be used to analyze and derive actionable insights from group discussions?
Novigi is an Australia-based leading digital transformation provider focused on the finance sector. We have been providing them a range of services from AI strategy, to solutions, over time.
While working with them, we tackled a tedious task many businesses, research organizations, government agencies and non-government entities are all too familiar with: deriving useful insights from a group discussion. Is there a faster, more efficient way to gain insights from a group discussion, stakeholder meeting, or a conference, without having to go through recordings or transcriptions one by one, analyzing the content for days or even weeks?
Deriving insights with natural language processing
We knew that this was a problem that had the potential to be addressed with a natural language processing (NLP) pipeline. We started building this pipeline for Novigi, and our first step was transcribing audio recordings, which, in this setting, was our source data. We did this with the help of Amazon Transcribe, one of the more mature speech to text services available. Then, we proceeded to develop an NLP pipeline with the help of state of the art NLP libraries to perform analyses. This pipeline identifies the most spoken words or phrases during the discussion, using word frequencies. It also recognizes named entities, and groups them into different categories. When combined, these results uncover valuable insights about the discussion.
Beyond that, sentiment analysis with its ability to gauge whether a particular statement is positive or negative, is at the core of our NLP pipeline. It can determine the overall sentiment of the participants about the discussion, as well as on specific discussion topics. For instance, if the discussion is on tackling climate change, it is possible to understand whether the overall discussion sentiment towards tackling climate change is positive or negative. If there is a specific discussion point on the contribution of the meat industry to climate change, it is again possible to understand the overall sentiment of the participants on this topic -- whether they think the meat industry contributes to climate change or not.
We also used topic modeling in our pipeline, a technique which clusters similar terms together based on certain patterns in words or phrases -- these clusters could then be interpreted as various topics discussed to derive valuable insights from unstructured text.
Scales beyond human capacity
We found that our NLP pipeline was highly effective in providing a fast analysis of the discussions, immediately after their conclusion. For instance,we obtained results which said that the overall sentiment towards the discussion was positive, and that a very high percentage of the participants had a positive sentiment towards the topic in discussion.
Furthermore, our NLP pipeline made it possible for Novigi to strategically position themselves as a company that uses bleeding edge technology. With our AI solution, it is possible to cut back on time spent on manually analyzing data from discussions from days or weeks, to minutes or hours. In fact, the analyses go beyond human capacity, as a human will not be able to perform all the tasks from automated transcription to sentiment analysis, at scale.
If you are looking to find similar solutions for existing challenges in your day to day operations, do reach out to us. Our team of AI experts would be keen to work with you to develop cutting edge AI solutions.
