Business Problem

• Providing SaaS platform for Store Operations, for Mom n Pop stores, Taascom faced a challenge, as the target users were not very educated or had any Data Acumen
• Due to the limited understanding of analytics/data sets, the platform adoption was very low
• They wanted to provide NLP based, real time insights & recommendations on the complex store data, comprehensible by the naive users, with nil to limited Data Acumen, to help them improve Store Operations & eventually revenue

About Taascom

TAASCOM helps companies transform their business model to a subscription-based service model. They achieve this by providing the technology, experience and expertise to enable this transformation.

Solution

Taascom partnered with GoML to integrate Generative AI-powered in-app analytics, leveraging GPT-4. The solution was built on top of complex Store Data, built into RedShift based DWH, to provide real time, intelligent, actionable insights & recommendations, directly influencing operations, inventory & sales, to help the store managers make better decisions, hence contributing to increased revenue.

The solution was built as a Chat agent, allowing the store managers to have NLP based, open ended conversation, with in - context follow up questions, what if scenarios & sales prediction queries, based on inventory & pricing changes.

Canvas Page Development

The implementation will commence with the development of a Canvas Page within Taascom’s platform. This page will serve as the interface for users to access proactive recommendations and query history seamlessly.

Proactive Recommendations

The heart of the solution lies in providing actionable insights to users based on the latest data trends. Leveraging Generative AI models like GPT-4, the system will analyze real-time data streams and generate proactive recommendations tailored to each user’s needs and preferences.

Real-time NLP-powered Data Analytics

The integration of NLP capabilities into Taascom’s analytics platform will enable users to derive insights from unstructured data in real-time. By leveraging Lyzr SDKs and advanced NLP algorithms, the system will extract meaningful insights from textual data, empowering users with comprehensive analytics capabilities.

API Development

To ensure seamless integration with Taascom’s existing infrastructure, GoML will develop APIs using Lyzr SDKs. These APIs will facilitate communication between the NLP engine and Taascom’s analytics platform, enabling smooth data exchange and interaction.

Amazon Redshift Integration

The implementation will leverage Amazon Redshift for robust data storage and processing capabilities. By integrating with Amazon Redshift, Taascom’s analytics platform will benefit from scalable and efficient data management, enhancing overall performance and reliability.

Architecture

Frontend Interaction managed by Amazon Route 53 for reliable DNS routing.

Elastic Load Balancer distributes traffic across Amazon EC2 instances for high availability.

Backend services are hosted on EC2 with Auto Scaling and are managed by Amazon ECS.

Data is stored on Amazon S3, Amazon RDS, and Amazon DynamoDB and orchestrated by the AWS Data Pipeline.

AWS Lambda is used for serverless functions, AWS CloudWatch is used for monitoring, and AWS IAM is used for secure access.

Amazon Elastic Search Service is used for real-time search and Amazon SageMaker for ML model development.