From tackling file chaos to tracking food miles, our team delivered real value through AI in just one day.
Kicking off at 10am sharp on Friday 27th June at the New Icon HQ, members of the New Icon team gathered for a one-day internal hackathon,for a chance to do what we do best: tackle real-world client problems with creativity, technical skill, and collaboration.
To keep things fair, we handed over the team selection to ChatGPT. Well, we tried too… in a twist of irony, ChatGPT simply read the names back to us in the exact order we sent them. After a second attempt, the two teams were locked in:
With the teams formed, the challenge was simple in theory but bold in scope:
Design, build, and pitch a new or improved feature for a real or typical New Icon client, ideally implementing AI.
By 4pm, each team had to present:
The goal wasn’t just to build something cool. It was to build something client-ready. That meant showing architectural thinking, user-first design, and commercial impact… and bonus points for making the most of AI.
After a rapid-fire brainstorming session, Team 1 honed in on a common problem nearly every client, and team, struggles with: finding the right file at the right time. It’s safe to say, in fast-paced work environments or shared drives, file names rarely reflect the actual content or intent behind a document, resulting in wasted time, duplicated effort, and frustration.
Their solution was fAIo, an AI-powered file search tool designed to help users quickly locate what they need, even when files are messy or mislabelled. Rather than relying on folder structures or filenames, fAIo uses natural language queries to scan the contents of documents, returning the most relevant results based on context. It can also auto-tag, summarise, and cluster related files to improve discoverability over time.
The team built a lightweight MVP using ChatGPT as the AI engine, avoiding the need for a complex backend. Vue and Node.js gave them the speed and portability they needed to ship a working demo within the day - one that could be easily deployed without additional infrastructure.
A small but powerful tool, fAIo showed how AI can reduce wasted time, unlock siloed knowledge, and create more productive workflows, both for our clients and internally.
Team 2 tackled a challenge rooted in a pressing global issue: the environmental cost of our food. The UK imports over 45% of its food, with everyday staples like avocados from Peru, bananas from Ecuador, and tomatoes from Spain making up much of what ends up in our trolleys. These long supply chains come at a price: high carbon emissions, inflated costs, and a growing disconnect between consumers and the origins of their food.
Their idea, Fair Miles, is an AI-powered tool that brings transparency to food shopping. It helps users understand where their food comes from, how far it has travelled, the carbon emissions it produced, and whether there’s a better, local alternative.
For consumers, it’s a simple post-purchase app that scans your receipt, breaks down the footprint of your shop, and makes greener recommendations for next time. For businesses, it’s a plug-and-play feature that can be integrated into existing supermarket apps - giving shoppers real-time insights and nudging them toward more sustainable, local options.
Beyond environmental awareness, Fair Miles opens the door to smarter supply chains and more meaningful loyalty schemes. Retailers can turn receipts into actionable insights, launching campaigns like “1 million food miles saved last month” or rewarding customers for choosing low-impact swaps. It creates value for three audiences at once: the consumer, the business, and the planet.
To bring it to life, the team used Vue and Vite on the frontend for quick and clean UI work, and a Node.js + Fastify backend for a lightweight API layer. The AI engine was powered by LangGraph and OpenAI, which allowed the team to break complex logic into manageable steps - cleaning product data, identifying food origins, calculating distances, and suggesting alternatives, all in real time.
After a full day of thinking, building, testing, and pitching, Team 2 took the crown with Fair Miles - a polished prototype that blended technical execution with a meaningful environmental mission.
Not only did it compute CO₂ estimates and suggest greener swaps, it even managed to read a pretty hard-to-read M&S receipt… and revealed that, surprisingly, we shop more locally than expected!
While both teams brought sharp thinking and clever ideas to the table, Steve awarded the win to Team 2 for delivering a standout MVP with huge potential for real-world impact.
Same time next quarter?
Subscribe to get our best content. No spam, ever. Unsubscribe at any time.
Send us a message for more information about how we can help you