• Funny Finder, a custom-built AI tool, was created to help comedy fans discover acts and artists from a vast and expanding menu of 700 shows.
  • The conversational-AI analysed and interpreted attendees' comedy preferences to create personalised recommendations.
  • Acting as a personal entertainment concierge it improved through user interactions and was a significant success, reaching 6,000 visits a day during the festival.

 

As the scale of the Melbourne International Comedy Festival (MICF) has soared, so too has the complexity of helping audiences navigate a vast, ever-expanding programme of shows. In 2025, the MICF partnered with Datacom to create Funny Finder, a conversational AI tool designed to make show discovery easy, personal, and fun.

The challenge: Navigating comedy overload

The MICF’s growth has been exponential, with the 2025 festival alone offering 696 shows from more than 2,000 artists across more than 7,800 performances in 133 venues. While this diversity is a hallmark of the festival, it also presents a daunting challenge: “choice overload”.

“Festivalgoers tell us they can be a bit overwhelmed by choice, but there is also a real sense that audiences want to discover someone new, and they are looking for an easy way to do that,” says Denise Damianos, the MICF’s Marketing & Partnership Director.

Traditional methods, such as the guide, online curated lists and social media spotlights help, but they are no longer enough alone to provide the personalised, conversational recommendations that festivalgoers craved. Comedy fans love the experience of chatting with one of the Festival’s knowledgeable customer service and info hub team members and wanted someone that could help them dissect the programme and find the perfect show, but in the digital world.

The solution: Funny Finder, the personal comedy concierge

Datacom and the MICF team co-designed Funny Finder as a “personal entertainment concierge”, leveraging artificial intelligence (AI) to understand users’ comedy preferences, availability, and even their mood. The AI-powered tool with a conversational interface allows festivalgoers to describe what makes them laugh or what they’re in the mood for, and then generates tailored show recommendations, complete with direct ticketing links.

“If you think about the traditional way of finding comedy acts at a festival like the MICF, you would either have to scroll through the program, or you might have a billion browsers open. The Funny Finder is actually a way of streamlining that and allowing you to search and ask multiple things at once.”

How it works

Funny Finder was powered by the Chat GPT-4.0 Large Language Model (LLM) search engine hosted on Microsoft Azure’s cloud infrastructure, enabling advanced contextual understanding and natural language conversation.

Built with React, the front-end JavaScript library, Funny Finder offered an intuitive interface that worked seamlessly across devices. Azure's scalable architecture ensured consistent performance even under heavy demand for the AI-driven tool. Integration of advanced language AI enabled natural dialogue.

This combination of advanced technology and user-centred design elevated Funny Finder’s content discovery into an easier and more personalised experience.

The AI was drawing data primarily from the show information provided directly by artists via the festival’s registration system, striving for robust recommendations and helping ensure that artists retained control over how their shows were presented.

As more users interacted with Funny Finder, the AI refined its recommendations, learning from real audience queries and surfacing valuable insights for future festival planning.

Collaborative development

Datacom’s approach was highly consultative, says Damianos, who admits to initially feeling some trepidation over an AI-powered tool dishing out comedy show recommendations.

“Comedy is subjective, and not all shows are for all people. It's a nuanced artform that a broad range of audiences love, but what makes them laugh can be very different,” she explains.

“Going into it, we had to be a bit courageous about trying something new, and trust that the process would achieve the intended outcome of starting to solve this problem for our festivalgoers.”

Datacom was there on every step of the journey, guiding the MICF team through audience journey mapping, feature prioritisation and data strategy, ensuring the solution fit the festival’s unique needs and tight timelines. Datacom also helped proactively manage the risks identified around the nuanced nature and subjectivity of comedy, and provided training to demystify AI for a small, non-technical festival team.

Datacom’s Director of AI, Lou Compagnone, says the co-design process with the MICF team and participating comedians threw up a serious question: should Funny Finder be funny itself?

“We all decided that its job was not to tell jokes,” says Compagnone. “But there was a moment in the testing phase when it started to become its own sort of comedian. So we had to turn the temperature of the model down and get it back to its original intent.”

“The Datacom team was just fabulous,” says Damianos. “We have a large reach as a festival, but we are a very small organisation with no experience in AI, so Datacom’s willingness to listen, adapt and educate us was crucial to Funny Finder’s success.”

Strong uptake and real engagement

Funny Finder was prominently linked from the MICF website’s homepage, offering a seamless experience with comedy show suggestions served up alongside ticket prices, show times, location and duration.

The tool – which was removed from the website at the conclusion of the festival – saw robust engagement, with up to 6,000 daily interactions during the festival and 130,000 Funny Finder visits across the three and a bit-week festival, clear evidence that audiences were eager for a smarter way to navigate the programme.

Many users embraced the innovation, finding the Funny Finder tool helpful for discovering new acts and planning their festival experience. The project surfaced valuable insights into audience behaviour, including what types of queries people made and where information gaps existed, such as requests for award-winning acts or more nuanced show attributes.

Some users encountered limitations, often related to the nuances of prompting AI or gaps in the data provided, reflecting the learning curve for both users and the technology.

Learning from experience

The MICF team realised the AI’s effectiveness was closely tied to the quality and completeness of the data it received. The team identified opportunities to enrich the dataset for future iterations, making recommendations even more precise and relevant. The experience also highlighted the potential for AI to streamline other festival operations, from artist communications to sales analytics.

“This was a fun project to work on,” says Damianos. “I’m excited about what we’ve started and where we can take Funny Finder.”

“We’ve talked about, in the future, Funny Finder having the ability to become your travel planner,” says Compagnone.

“An AI agent could go beyond booking your ticket to making a reservation at a restaurant near to the show venue, and ordering you an Uber. Eventually, it could be your concierge for the night.”

As the festival team looks to the future, Funny Finder stands as both a successful proof of concept and a springboard for further digital transformation.

Related industries
Media, entertainment and telecoms
Related solutions
Artificial intelligence