As part of a push for more creative customer engagement, TDMC's creative and development team proposed Rama capitalise on the AI boom with a chatbot unlike a typical customer service bot - one designed to do two jobs at once. Customers would chat with "Rameezy," a character from the campaign, suggesting everyday meal names in exchange for a "Ramanized" twist on them. Every suggestion doubled as a competition entry and a contribution to the brand's growing bank of meal names for future use. Built as a custom mini web app, trained on OpenAI's API to speak in Rama's own voice, and installed via a bespoke WordPress plugin, Rameezy turned a simple conversation into both a marketing moment and a content engine.
01
Executive Summary
02
Introduction
The brief was simple in ambition, if not in execution: find a more creative way for Rama to engage its customers, and do it by tapping into the cultural momentum around AI. Rather than treating that as a novelty add-on, TDMC's team saw an opportunity to build something that worked harder than a typical chatbot - a tool that could entertain customers in the moment while quietly building something valuable for the brand long after the conversation ended.
03
The Opportunity
A standard AI chatbot can hold a conversation. What Rama needed was a chatbot that could turn that conversation into something useful - capturing genuine, brand-relevant user input at scale, without ever feeling like the customer was filling out a form. The challenge was designing an experience playful enough to invite participation, while building the technical backbone to make that participation count.
04
The Solution
Two jobs, one chatbot
At the heart of the concept was Rameezy - a character from the campaign who'd ask customers for the everyday name of a meal, then return a "Ramanized" version of it. Every name a customer submitted did double duty: it triggered entry into a competition, and it was added to a growing bank of meal names for the brand team to draw on going forward. Engagement and content collection, happening in the same interaction.
Teaching Rameezy to talk like Rama
The chatbot's intelligence was built on the OpenAI API, trained on Rama's existing set of Ramanized dish names alongside the brand's intended conversational style - its lingo, tone and vocabulary. The goal wasn't just accurate name generation; it was a chatbot that could handle an unexpected question or an off-script comment the way a real person would, without breaking character.
Bringing Rameezy to life on Rama's site
To keep things simple to deploy and just as simple to retire when the campaign ended, the chatbot was built as a self-contained mini web application, installed onto Rama's WordPress site through a custom-built plugin. It lived on its own dedicated page, ready to be pointed to from any marketing channel. Our design and development teams then built the page around it - an SMS-style interface that made talking to Rameezy feel like texting a friend, wrapped in campaign visuals that made the whole experience feel unmistakably Rama.
05
Results
Rameezy gave Rama a customer engagement tool that worked on two fronts at once.
- Delivered a fully custom AI chatbot, trained specifically on Rama's brand voice and existing naming conventions.
- Combined marketing engagement and user-generated content collection into a single, seamless interaction.
- Built a reusable bank of customer-suggested meal names for the brand team to draw on beyond the campaign.
- Created a low-friction, easily installable and removable solution via a custom WordPress plugin - built for the lifespan of a campaign, not a permanent platform commitment.
- Delivered a fully branded experience, from AI-trained conversation to landing page design, that felt like talking to a character rather than a company.
06
Conclusion
Rameezy is proof that AI doesn't have to mean a generic, off-the-shelf chatbot bolted onto a website. By pairing creative concept development with genuine technical execution, TDMC gave Rama a tool that entertained customers, stayed true to the brand's voice, and left them with something they didn't have before: a customer-built bank of brand-relevant content, gathered one conversation at a time.
