Request Demo

Enhancing Chrombot.ai’s Customer Support:

 

A 75% Reduction in Support Time with Hybrid.chat’s AI Virtual Assistant for HPLC Chromatography

Client: Rohan Ayyappa, Chrombot.ai

Industry: Chatbot

Team: Hybrid.chat

Tools and technologies: Javascript, React, Python, Node.js, Mongo, Slack, 360dialog, Whatsapp, and AWS, Dialogflow, OpenAI, Langchain, Pinecone

Features

  • AI-powered Whatsapp and web chatbot capable of addressing HPLC chromatography-related queries with over 95% accuracy by referencing trained PDFs.
  • Simplified process for updating the bot’s training content (PDFs or other documents).
  • Shared inbox for Chrombot.ai’s human agents to respond to customers.
  • On-premise deployment on a dedicated server.

About the client

Diahealthcare, a Mumbai-based company, provides support to scientists with chromatography-related queries, dealing with HPLC Columns, GC Columns, HPLC Consumables, etc.

Mission: To empower customers in making the world healthier, cleaner, and safer by accelerating life sciences research, solving analytical challenges, improving patient diagnostics, delivering medicines to market, and increasing laboratory productivity.

 

Challenge

While dealing with scientists, Diahealthcare found a pattern of repetitive queries so they started creating a PDF for queries. Their human agents needed to find answers from the complex table-based PDFs to provide answers. They tried existing document AI chatbot solutions in the market but as PDFs had a lot of tables and linked data, it didn’t work well for them. Because accuracy was the core parameter for them to make decisions on product

Our Approach and solution:

The Hybrid.chat team developed a custom solution as a new product chrombot.ai. This revolutionary chatbot encompasses several key features:

  • Bulk PDF Training Data Handling: The bot is designed to handle large volumes of PDF training data, with multiple books uploaded to train the chatbot.

 

  • Complex Table-Based PDF Reading: A significant challenge was the accuracy of interpreting the linked data in the PDFs. The Hybrid.chat team developed a training model capable of understanding and accurately interpreting this complex data.

 

  • Real-Time Training Data Updates: An on-premise user interface was developed, allowing new PDFs to be uploaded and trained in real time.

 

  • WhatsApp API Support and Scalable On-Premise Hosting: To serve a large community efficiently, WhatsApp support was essential. The Hybrid.chat team utilized Meta APIs with custom middleware, providing WhatsApp API admin privileges and enabling secure hosting of WhatsApp bot solutions on separate, independent servers.

 

  • Shared Inbox for Human Agents: Chrombot.ai’s human agents can respond to scientists and maintain conversation records in the in-house CRM, Hybrid.chat, ensuring a seamless customer support experience

 

Result and impact:

Hybrid.chat‘s solution has revamped the way scientists engage with Chrombot.ai:

  • A remarkable 75% reduction in customer support time.
  • Swift automated responses leading to increased user satisfaction.

 

Tools and technologies used:

💡Tools /Platforms: Hybrid.chat, Slack, Whatsapp, AWS, Rasa, Dialogflow

💡 Tech stack: MERN, Python

💡 Database: Pinecone, Mongo

💡AI: Inhouse NLX pipeline, OpenAI

💡 Orchestration: Langchain

 

Inspired by Chrombot.ai’s transformation? Discover how Hybrid.chat can revolutionize your customer support experience. Request a Demo Today!

Request Demo