bg

Transform Your Supply Chain Planning and Marketing Strategies with Google Cloud and SAP Integration

Cloud Marketplace

Explore Pi Agent pi-logo

Pluto7 and Google Cloud helped the University of New Mexico and Project ECHO empower local communities using intelligent chatbots

succes stories images

“The chatbot POC was also a step in the direction of understanding how effective Google’s machine learning and AI can be in handling more complex questions and how it can be used in other data sets at Project ECHO.  We were won over by Pluto 7’s depth of knowledge in this domain and the technical expertise of their team.

Debi Phillpott -IT Project Manager at ECHO Institute™.

Introduction

Project ECHO has a goal of touching a billion lives by 2025. To achieve this goal they need to explore opportunities to engage and expand the participation of our current users, as well as to ensure information about the ECHO model is readily available for new users. Pluto7 proof-of-concept (POC) was to develop a chatbot for their website, which helped solve an immediate problem of ensuring people visiting our website easily find information about ECHO programs with minimal staff intervention. 

The chatbot POC was also a step in the direction of understanding how effective Google’s machine learning and AI can be in handling more complex questions and how it can be used in other data sets at Project ECHO. 

Why We Chose Pluto7

Pluto7, a Google Cloud Premier Partner, specializes in preparing and deploying accelerated solutions and custom applications in smart analytics, machine learning, and AI. 

The stated desires for a chat agent match the capabilities of Dialogflow and other Google Cloud Platform tools, such as Cloud Firestore. The Pluto7 team leveraged its knowledge and expertise to build a solution that utilizes customer data and natural language processing capabilities to interact with individuals in multiple languages.

Project ECHO team were won over by Pluto 7’s depth of knowledge in this domain and the technical expertise of their team. This POC was also a step in the direction of understanding how effective Google’s machine learning and AI can be in handling more complex questions and how it can be used in other data sets at Project ECHO. Since this was a POC we kept the scope very narrow.  Even with the limited use cases we outlined for Pluto 7, we were able to clearly see the possibilities of the power of Pluto 7’s capabilities and their close partnership with Google. 

Solution

People and organizations from all over the world are leveraging the ECHO Model to scale their own social initiatives in fields like healthcare, education, and social justice. ECHO mission is to connect groups of community providers with specialists at centers of excellence in regular real-time collaborative sessions. As a way to improve their online engagement, Pluto7 designed and created a virtual chat agent for assisting website visitors with information about ECHO,. 

By leveraging Google Cloud technology, Pluto7 data engineers trained the agent to have a friendly, interactive, and profession person supporting both English and Hindi languages. 

Results

A virtual Chat Agent that assists Users in a polite, user-friendly, interactive manner supporting both English and Hindi languages. Built using Google DialogFlow, the Agent can detect the user intention, misspelled words, short names, even partial names. Additional attributes were built to:

  1. Identify all possible questions: the agent was trained using a conversations flow chart  which covers all possible questions a user may ask agent thereby making the solution more precise.
  2. Send Accurate Responses: Trained chat agent to respond in a polite and user-friendly manner. The responses are very specific without any ambiguity
  3. Detect the user intent: Pluto7 leveraged Google’s Dialogflow to detect the user’s intention correctly. Internally, Dialogflow uses Machine Learning on the trained sample question phrases and detects user intent.
  4. Handle Misspellings and partial words: the agent could recognize common words like names of countries, cities, even when  the user misspells them. In addition to this, Pluto7 team have implemented an NLP technique at the backend which can recognize the correct hub, program name even if the user misspells or gives a partial name.

 

Enable Decision Intelligence Into Every Corner Of Your Product And Operations.

industry image

Industry Education

Platform Education-ml

Organization Name: University of New Mexico

Challenges

  • Designing all possible conversations between Agent and User
  • Fine-tuning and processing of data
  • Recognizing user intents
  • User Language
  • Handling complex and diverse data sets

Results

  • Components necessary to solicit customer data
  • An agent that can respond to user queries in natural language
  • Conversations Flow Chart
  • Specific Responses – no ambiguity
  • User Intent Detection
  • Partial words and misspelling recognition.
  • Name, hub and program recognition.

Products Used

  • Google Assistant
  • Dialogflow
  • Google App Engine
  • Google BigQuery
  • Google Cloud Dataprep
  • Google Cloud Storage
  • Google Cloud Natural Language
  • Google Cloud Speech