Rummy Chatbot — Namrata Singh

A chatbot redesign to transform customer support for Rummy and Fantasy products at Games24x7 — moving from fully human-routed tickets to an intelligent self-service system built on Haptik's conversational AI platform, targeting 75% query resolution without human agent intervention.

The Problem

Manual-only support at scale

Customer support can be self-service, especially for the large number of repetitive / redundant queries we get from our customers and we don't have any chat support in our system. For each query we transfer to a human service agent from our CS team.

We want to make customer support more transparent, quicker and easier for end users. Additionally, there are a bunch of experience issues as well listed below.

What We Know

All tickets / calls routed to ticket creation (~1.9 lakh p.m.). Significant number of open tickets with no resolution tracking
There are 3 sections for end users to find support: Help, Contact Us, and Support on the web. This is super confusing for end users
Another confusion is that "Help" on Web is marketing content, whereas on the app it leads to a list of FAQs — there is a need to standardize and clean up the support experience
Players do not have visibility on their issues raised, i.e. there's no tracking on their complaints, etc.
If an article is not helpful, a player should be able to contact support easily. Instead, the user will today have to navigate back and forth to find touchpoints to connect with us.

Existing pages with/without Help and Support

Existing pages — Help and Support touchpoints before redesign

Hypothesis

Target 75% self-service

75%

Self-service target

User queries resolved without intervention of human agents — via chatbot automation.

25%

Human agent escalation

Only complex queries escalated to CS team — open tickets only for genuinely unresolved cases.

We want to eventually target ~75% user query resolution without the intervention of human agents. That means only complex ~25% queries get transferred to human agents (Games24x7 employees who are resolving queries, i.e. open tickets). Also, we want a speedy, easier and simpler UX for end users to find answers to their questions. This will impact CSAT and also long term help us reduce manpower costs.

Approach to Solution

Chatbot as first line of defense

Deploy a chatbot (virtual chat assistant) as the 1st line of defense for end users. The chatbot, via its integration capabilities, is supposed to answer the majority of end-user queries.

Only open tickets that are user queries wherein users do not get a resolution, should then form a ticket for our CS team to handle.

Powered by Haptik — Conversational AI Platform

The Process

Research to ready-for-dev

After seeing the current flow and pages, I mapped where help and support should be added, conducted competitor research, and iterated through design drafts before final stakeholder approval.

01

Map the Existing App Flow

I mapped the existing app flow to identify where Help & Support touchpoints should be added — creating a proposed flow showing exactly where chatbot entry points would live.

Flow where help and support can be added

02

Competitor Analysis

After flow creation I did market and competitor analysis to understand how other apps are solving customer support. First I checked how many apps are using chatbots — starting with well-known examples like Amazon, Zomato, and Swiggy, then looking closely at direct competitors Junglee Rummy and Dream11.

Apps using chatbot support — landscape research

Types of chatbots used by different companies

Junglee Rummy

Dream11

Zomato

Swiggy

Amazon

03

Design Drafts

Once done with all research, I started working on first design drafts covering all use cases and scenarios — five help home options, two chat initiation options, and three CTA/page style variations.

04

Stakeholder Review & Tech Handoff

After months of research, ideation, and design iterations, the final work was shared with stakeholders and approved for handoff to the tech team.

Design Exploration

Exploring the design space

I explored multiple design directions for each key component — ensuring all use cases and edge cases were covered before converging on the final solution.

Help & Support Home Page — Design Options

Chat Initiation Point Designs

Different Styles for CTAs and Pages

Final Outcome

Final designs & flows

After months of research, ideations, and design iterations — the final work received stakeholder approval and was handed over to tech.

Final Help Home Page

Ticket Creation

Inner Pages — Chat & Like/Dislike Actions

Main Menu Update — Before & After

Before

→

After

Lobby Screen — Before & After

Before

→

After

Final Flows

Complete chatbot conversation flows

Four flow states showing the complete user journey through the chatbot — from the Help Home through FAQs, Ticket Creation, Quick Links, and Call/Chat escalation.

FAQs