scroll

Generative AI Prototype

A web app with a feature to transcribe meetings, analyze them, and answer related questions

Generative AI Prototype utilizes AI and chatbot features to transcribe the meetings, analyze them and answer related questions. The app's primary goal is to create a summary of the meetings' key points, actions taken, and sentiments without a need to participate in them physically.

Expertise

Project Management

Frontend Development

Backend Development

Industry

IT

Timeline

6 weeks

Tech stack

ChatGPT API, LangChain

Requirements
Challenge
Process
Technical Requirements

Our cutting-edge conversational prototype leverages state-of-the-art techniques in natural language processing and generative AI.

The system architecture is powered by a suite of large language models, including GPT-3.5 and GPT-4, enabling natural language understanding and response generation capabilities. The integration with Langchain provides an abstraction layer for seamless large-language model querying.

Leveraging its capabilities, we built customized chains to enable robust transcription processing and analysis. This included chains for multi-step workflows like generating summaries, extracting action items, and identifying key points from transcripts. The integration of question-answering functionality empowers users to extract salient information as needed.

Additionally, we integrated sentiment analysis modules to gain insights into the contextual emotional tone and valence within the conversational transcripts. Prompt engineering techniques were employed to refine the chains for optimal performance on our specific dataset.

We assessed performance via two primary methods: user feedback gathered through interaction icons and automated evaluations. For each transcript, the system executes question generation and then answer prediction using multiple configurations. These predictions are compared to the generated answers by semantic similarity to identify optimal parameters. In summary, the configurable pipeline and analysis toolkit enable data-driven optimization of cost and performance.

Additionally, we conducted evaluations to assess the integration costs, benefits, and potential efficiency gains from incorporating third-party platforms like SageMaker and HuggingFace. This analysis explored pathways for optimizing the technology stack in future iterations.

The Challenge

In developing our Generative AI Prototype, we aimed to provide a platform where users could browse meeting notes created from transcriptions, ask questions about recorded meetings, and receive detailed answers without needing to sift through the notes.

However, this ambitious vision posed several unique challenges:

  • Transcription quality
  • Natural Language Processing
  • User experience
  • Data organization
  • Scalability
  • Data privacy
  • Performance evaluation
  • The rapid evolution of the field
Our Process

With our tool, you can effortlessly garner insights from meetings without the hassle of reviewing extensive notes or long recordings.

It utilizes pioneering natural language processing techniques to transcribe, analyze, summarize, and highlight the core points of meetings. Not just that, it enables users to interact with a chatbot for real-time access to summarized data, streamlining the process of fetching specific details or clarifications.

Say goodbye to laborious manual searches as our Generative AI Prototype brings all necessary information to your fingertips with a simple chatbot query.

what partners say about us?

Just as learning a new language opens doors to new cultures, learning the language of AI opens doors to new dimensions of understanding.
Mariano Cigliano
Tech Lead of the Project
The code and the work were good quality and really what we were looking for. They were able to bring technical design thinking to the project. Project management was tight and I always knew what was happening.
Daniel Kiyoi
CEO & Founder Magic Dust
Apptension was flexible and professional. When I needed to quickly add capacity, it took a week or two at worst - often days. The cooperation enabled me to slowly scale up my own IT team, and the company was very helpful until the last moments of the transition.
Mateusz Oleksiuk
CEO LESS_
The technical creativity delivered by the team at Apptension was invaluable in the formation of our mixed reality start-up Hyper.
Nathan Sparshott
Co-Founder & CEO of Hyper
The SaaS Boilerplate Apptension built was a huge reason we were so successful, because all these little seemingly unrelated tasks and integrations needed to happen. If we had been working with anybody else, It would have probably taken months to do the same work.
Kwame Nyanning
CEO blkbx.
We needed their help to build the backend services and capabilities needed to deliver a production level on demand. Their ability to produce high-quality work with consistency while dealing with a new type of project was impressive.
Kelly O’Conor
Product Lead, Siberia
The project’s success has resulted in a long-term partnership for design and development. Apptension is a fantastic partner, who is willing to go above and beyond in order to deliver what the client needs.
Catarina Rocha
p(r)oud solutions
Apptension fared well in our project, working with our bespoke CMS and complex requirements. The designs were implemented well and the schedule was kept tight.
Christian Marc Schmidt
Partner at Schema
Looking for similar outcomes?

recent case studies

Generative AI Prototype
A web app with a feature to transcribe meetings, analyze them, and answer related questions
About
Generative AI Prototype utilizes AI and chatbot features to transcribe the meetings, analyze them and answer related questions. The app's primary goal is to create a summary of the meetings' key points, actions taken, and sentiments without a need to participate in them physically.

Expertise

Project Management

Frontend Development

Backend Development

Industry

IT

Timeline

6 weeks

django
flask
react native
node.js
next.js
python
gatsby.js
vue.js
react.js
javascript
aws
docker
serverless
figma
photoshop
illustrator
after effects
firebase
blender
graphql
nuxt.js
Scss
typescript
apollo
saas boilerplate
styled components
D3
bigcommerce
bigquery
Wagtail CMS
Django Rest
redux
gsap
i18next
kubernetes
Google Cloud
Platform
rabbitmq
celery
Before
After
Prerequisites
Challenge
Challenge

In developing our Generative AI Prototype, we aimed to provide a platform where users could browse meeting notes created from transcriptions, ask questions about recorded meetings, and receive detailed answers without needing to sift through the notes.

However, this ambitious vision posed several unique challenges:

  • Transcription quality
  • Natural Language Processing
  • User experience
  • Data organization
  • Scalability
  • Data privacy
  • Performance evaluation
  • The rapid evolution of the field
Requirements
Technical process

Our cutting-edge conversational prototype leverages state-of-the-art techniques in natural language processing and generative AI.

The system architecture is powered by a suite of large language models, including GPT-3.5 and GPT-4, enabling natural language understanding and response generation capabilities. The integration with Langchain provides an abstraction layer for seamless large-language model querying.

Leveraging its capabilities, we built customized chains to enable robust transcription processing and analysis. This included chains for multi-step workflows like generating summaries, extracting action items, and identifying key points from transcripts. The integration of question-answering functionality empowers users to extract salient information as needed.

Additionally, we integrated sentiment analysis modules to gain insights into the contextual emotional tone and valence within the conversational transcripts. Prompt engineering techniques were employed to refine the chains for optimal performance on our specific dataset.

We assessed performance via two primary methods: user feedback gathered through interaction icons and automated evaluations. For each transcript, the system executes question generation and then answer prediction using multiple configurations. These predictions are compared to the generated answers by semantic similarity to identify optimal parameters. In summary, the configurable pipeline and analysis toolkit enable data-driven optimization of cost and performance.

Additionally, we conducted evaluations to assess the integration costs, benefits, and potential efficiency gains from incorporating third-party platforms like SageMaker and HuggingFace. This analysis explored pathways for optimizing the technology stack in future iterations.

Roadmap
Solution
Solution

With our tool, you can effortlessly garner insights from meetings without the hassle of reviewing extensive notes or long recordings.

It utilizes pioneering natural language processing techniques to transcribe, analyze, summarize, and highlight the core points of meetings. Not just that, it enables users to interact with a chatbot for real-time access to summarized data, streamlining the process of fetching specific details or clarifications.

Say goodbye to laborious manual searches as our Generative AI Prototype brings all necessary information to your fingertips with a simple chatbot query.

WHAT The CLIENT Said ABOUT this project?

WHAT THE R&D Tech Leader SAYS ABOUT THE PROJECT?

Just as learning a new language opens doors to new cultures, learning the language of AI opens doors to new dimensions of understanding.
Mariano Cigliano
Tech Lead of the Project
Just as learning a new language opens doors to new cultures, learning the language of AI opens doors to new dimensions of understanding.
Tech Lead of the Project

what partners say about us?

The code and the work were good quality and really what we were looking for. They were able to bring technical design thinking to the project. Project management was tight and I always knew what was happening.
Daniel Kiyoi
CEO & Founder Magic Dust
Apptension was flexible and professional. When I needed to quickly add capacity, it took a week or two at worst - often days. The cooperation enabled me to slowly scale up my own IT team, and the company was very helpful until the last moments of the transition.
Mateusz Oleksiuk
CEO LESS_
The technical creativity delivered by the team at Apptension was invaluable in the formation of our mixed reality start-up Hyper.
Nathan Sparshott
Co-Founder & CEO of Hyper
The SaaS Boilerplate Apptension built was a huge reason we were so successful, because all these little seemingly unrelated tasks and integrations needed to happen. If we had been working with anybody else, It would have probably taken months to do the same work.
Kwame Nyanning
CEO blkbx.
We needed their help to build the backend services and capabilities needed to deliver a production level on demand. Their ability to produce high-quality work with consistency while dealing with a new type of project was impressive.
Kelly O’Conor
Product Lead, Siberia
The project’s success has resulted in a long-term partnership for design and development. Apptension is a fantastic partner, who is willing to go above and beyond in order to deliver what the client needs.
Catarina Rocha
p(r)oud solutions
Apptension fared well in our project, working with our bespoke CMS and complex requirements. The designs were implemented well and the schedule was kept tight.
Christian Marc Schmidt
Partner at Schema

other success stories

Conversational AI Avatar
Building a real-time conversational AI avatar
Riyadh Air
Building a microsite with a 3D plane model
Feelit
A SaaS mobile app designed to bridge the gap between event attendees and event makers
Creative agency
Developing an eCommerce website for a luxury brand
Mojo
A website solution for an AI-based fertility tester
mTab
Building a full-stack data analysis, visualization and storytelling platform