Deep - learning to mimic human behaviour

A unique communication platform that assists customers to resolve their queries or support requests at point of contact through Web, App or WhatsApp. Uses advanced pre-trained neural network algorithms to process multiple queries simultaneously, and assist in quick closure of tickets by the customer



  • Superior, human-like cognitive capabilities that allow it to read text, images as well as documents, creating a super-efficient system for interacting with customers
  • Patent pending feature that enables resolution of multiple intents simultaneously
  • Auto complete feature simplifies customer experience journey, prompting him with the exact suggestion of what he is looking for
  • Intelligent UI and design that allows the customer to change request type by himself
  • Can function as a virtual sales assistant, upselling and cross-selling products, across the board


  • Can identify customer intent with high degree of accuracy and handle complex business problems with ease
  • Assists the customer in resolving most of his support queries at the point of contact, with minimal human intervention
  • Request soft copy of documents
  • Register a claim
  • Request for claim status
  • Change name / address or other account related information
  • Process Information

Acu- men-tal

  • Its objective is to reduce ticket generation by 80%
  • Intelligent built in analytics with deep learning enables personalized promotions and intelligent cross sell / upsell of products based on customer profile and user behavior / nature of query

Six Advantages

There are 6 ways in which Assisted Closure offers functional
superiority and distinct advantages to its users: 

  • Easy to train
  • one window
  • memory-enabled
  • virtual salesman
  • cost-saver
  • time-saver

Easy to Train

  • Can be trained using email data, hence ample data available for training
  • Long, unstructured texts can be read and hence access to more data
  • Superfast learning curve on limited data requirements  – AI needs just seven days for training on dataset across any industry.
  • Assisted Closure possesses deep learning capabilities that use some of the best available processes in NLP and ML. These inbuilt capabilities automatically assess each request and identify new categories that are potentially training data to constantly improve the clarity and understanding of the system. Each new situation becomes a training ground for learning and improvement, further enhancing the product’s capabilities.


One Window

  • Multi Intents can be broken down and be addressed one at a time
  • Long, unstructured texts can be broken into smaller components and all the issues of the customer can be accommodated.
  • Voice data can also be deciphered ( customers can speak out their issues and get redressal)
  • A single interaction can result in the addressing of several customer issues thus providing both a high degree of efficiency, as well as enhanced customer satisfaction. Since the system is capable of deciphering several formats such as text, images and also voice data, it becomes very easy for customers who are not capable of keying in long bits of data. The system can auto-respond with data from previous interactions and hence resolve issues faster and with greater accuracy. 



  • Memory tickets are the norm, meaning that Assisted closure can always keep in mind the list of tickets managed from the date of inception
  • Issues are immediately accessible
  • Customers can keep track of all their open as well as closed tickets and maintain a complete log of their queries within the app
  • Auto-fill of forms is possible, thus greater accuracy is possible as well as huge time savings
  • Additionally, when agents are needed to intervene, they will have a clear understanding of the problem due to availability of archives


Virtual Salesman

  • Superior cognition skills enable it to assess each customer query and process it in terms of customer behaviour
  • Hence it can anticipate the sales process and help in intelligent cross sell or up-sell of related products, as well as proffer relevant promotional literature
  • It can also recognise when human interaction is needed and can prompt an agent with the potential to cross sell or upsell as the case maybe



  • Many intents can be carried out in a single interaction, this will help save costs
  • Different kinds of data can be utilised – visual, voice, words, etc., hence more capable, more effective operations, [again will impact the bottom line positively] 
  • Can upsell and cross sell, hence can generate revenue
  • End-to-end handling of ticket, superior query resolution, thus saving agent time
  • More queries can be handled each day
  • Superior clarity about which call needs human intervention will leave no open-ended response but clearly intimate an agent, with access to a record of all communication with the customer
  • Reduce the need for human intervention by ensuring 80% tickets are closed at point of contact by customer.



  • Many intents can be carried out in a single interaction
  • For people who can’t key in, voice can be used to connect
  • Memory ensures that data is retrievable and doesn’t have to be furnished each time
    • Autofill – saves up to 60% time, impacting both time and costs

    For the organisation:

  • Intelligent system, hence can improve further and offer superior quality of interactions
  • More data can lead to even more streamlining of offerings and sharper clarity on service issues
  • More satisfied customers (Better CSAT scores)

The future of chatbots

Assisted Closure V/S Present Day Chatbots

Feature AssistedClosure Chatbot
Multiple Intents Yes No
Understanding long unstructured text Yes No
Use email data for training Ample training data with high launch accuracy Less data for training Lower launch accuracy High learning curve for ML with trial and error
Memory of past tickets (Open & Closed) Yes No
Advertisements, campaigns, cross sell / upsell products, marketing material featured between tickets, Banners Yes No
App within App: Not restricted to boundaries / limitations of query / service bot Yes No
Throw / push notifications / sales promotions through or within app Yes No
Query resolution Complete process flow for each query or ticket 2 way chat conversation / decision tree based IVR
Autocomplete Autofill (60% time conservation) Dropdown
Model Deep Learning Pattern recognition NLP + Decision Tree Keyword based
Mahesh Krishnamoorthy

The Product is Really impressive

The quality of the research done in the area of unstructured text & vision by AmyGB is truly impressive, and the product offering has real potential for impact. They have an excellent team that whole heartedly puts their mind into understanding your business challenges, and deriving innovative solutions to solve them. One of their products, MailERA, has helped us treat every customer like royalty. To see such brilliant research beginning in the area of unstructured text in India is exciting.

Mahesh Krishnamoorthy

-Chief Business Officer, Mahindra Integrated Business Solutions Pvt. Ltd
Mrunal Fernandes

Our repeat emails have reduced by 5%

Customer happiness is at the fore front of everything we do at RGI. Engaging with AmyGB has been a great decision, as their unique AI solutions have tremendously helped improve productivity and taken Customer Experience to an altogether new level. Our repeat email volume has reduced by 5% and our FTR accuracy stands at 95%. It has been a wonderful experience working with AmyGB, and are already engaging them for several other crucial projects within RGI.

Mrunal Fernandes

-Associate Segment Head – Customer Service, Reliance General Insurance

As Seen On

All news coverage
interested in getting
a demo
request demo