What is a Hackathon? Complete Guide + Real Case Study
Updated: 10 hours ago
Hackathons went from Silicon Valley novelty to a mainstream strategic tool. Companies across healthcare, education, consumer goods and finance use them to validate ideas fast, upskill teams, and generate solutions that traditional processes would take months to produce. What accelerated this shift? Largely, AI.
This guide covers what a hackathon is, how it works in practice, what changes when AI tools are in the room, and what a real case looks like, based on a two-day event ideafoster organized for Fundación Repsol in Madrid with 25 participating teams.
What is a hackathon?
A hackathon is an intensive, time-boxed collaborative event where people with different skills work together to solve a specific challenge. The output is usually a prototype, a validated concept, or a working solution, built in hours rather than months.
The name combines "hack" (programming ingeniously) with "marathon" (sustained, intensive effort). Originally rooted in software development, hackathons have expanded into healthcare, education, corporate strategy and social impact: anywhere a team needs to move fast and think differently.
What makes them work is the constraint. Deadlines force decisions. Time pressure eliminates perfectionism. Multidisciplinary teams surface blind spots. A well-structured challenge gives everyone a direction without narrowing what's possible.
How a hackathon works: structure and phases
Most hackathons follow a recognizable structure, even if the specifics vary by context and scale.
It starts with a brief: a challenge, a context, and a set of constraints. Good briefs are specific enough to guide without dictating the solution. The Fundación Repsol hackathon focused on sustainability innovation, broad enough for creativity, concrete enough to evaluate outcomes.
Teams form around the challenge. The best compositions mix technical profiles (engineers, developers) with creative ones (designers, communicators) and domain experts. Each lens surfaces different aspects of the same problem.
During the event, teams have access to tools, mentors and feedback loops. The checkpoint structure matters enormously. It is not just support; it is the mechanism that surfaces weak ideas early, before a team has over-invested in the wrong direction.
The final phase is presentation and evaluation. Judges assess innovation, feasibility and impact. But the real value is not always the winning idea. It is the thinking that surfaced, the team dynamics that formed, and the speed at which a complex problem got attacked.
AI hackathons: what changes when every team has a language model
Until recently, hackathons were defined by what a team could build from scratch in 48 hours. That constraint shaped everything: scope, feasibility, what was achievable.
AI changes the baseline. A team with access to a language model, a code assistant and a generative image tool can prototype in hours what used to take days. The ceiling for what is possible in a weekend has shifted significantly.
This creates new dynamics. Technical skill matters less as a barrier to entry. The quality of problem framing matters more. Teams that spend the first hours defining the problem well consistently outperform teams that jump straight to building.
There is also a new failure mode: teams that over-rely on AI output without applying judgment. The AI can generate a landing page, a pitch deck, a working prototype. But it cannot tell you whether the problem is worth solving. That is still human work.
When ideafoster facilitates AI hackathons, we invest significant time on the framing phase before anyone opens a tool. The best outcomes consistently come from teams that know exactly what they are trying to prove before they start generating.
Real case: the Fundacion Repsol hackathon
In 2024, ideafoster facilitated a two-day hackathon in Madrid for Fundacion Repsol. Twenty-five teams participated, each working on a challenge at the intersection of sustainability and innovation.
The format was structured but not rigid. Teams had access to mentors across strategy, design and technology. Each checkpoint was designed to pressure-test assumptions, with the goal of surfacing weak ideas early rather than at the final pitch.
Five teams advanced to the final round. What distinguished them was not the sophistication of their prototypes but the clarity of their problem framing and the specificity of their proposed solution.
The event confirmed what we apply across all our Events and Experiences work: the quality of a hackathon is determined in the planning phase. The brief, team composition, mentor selection and checkpoint structure are not logistics. They are the design of the outcome.
Hackathons across industries: where they work and why
Hackathons started in tech and now run across virtually every sector.
In healthcare, they bring together clinicians, engineers and patients to solve operational problems or accelerate diagnostic tools. Some of the fastest early-stage medical device concepts have emerged from 48-hour clinical hackathons.
In education, they help institutions rethink curriculum, student experience or administrative systems, often with students themselves as participants.
In corporate environments, they break silos, surface internal talent and test ideas before committing budget to a formal project. The speed is valuable not only for output but for organizational learning.
Social impact hackathons connect NGOs, governments and technologists around specific challenges: food waste, housing access, climate adaptation.
The format adapts. What stays constant is the value of focused time, diverse perspectives and a real challenge worth solving.
What makes a hackathon succeed, and what kills it
After facilitating dozens of hackathons across sectors, ideafoster has identified the patterns that consistently predict outcomes.
The briefs that generate the best ideas are specific about the problem and silent about the solution. "Improve our customer experience" generates noise. "Reduce first-contact resolution time in our support team by 30%" generates focus.
Team composition matters more than team size. Five people with complementary skills outperform ten people with overlapping ones. Multidisciplinary is not a buzzword; it is the operational variable that determines whether a team can prototype something real.
Mentor quality determines depth. The role of a mentor in a hackathon is not to answer questions; it is to ask better ones. Mentors who pressure-test assumptions generate better outcomes than mentors who provide answers.
The post-hackathon plan matters as much as the event itself. If you do not decide before the hackathon what happens to winning ideas, who owns them, what resources they get and what the next milestone is, the event produces pitches, not progress.
Planning a hackathon for your team, your clients or your community? We design them from brief to facilitation to post-event follow-through. Get in touch.
The future of hackathons sounds exciting, especially with AI and VR opening up new ways to approach problems and visualize ideas. Remote participation could also bring together people with very different skills and perspectives, which may lead to more creative solutions. The emphasis on combining human creativity with technology is probably the most interesting part. That balance between skill and technology can even be seen in SoFlo Wheelie Life, a motorcycle game where control matters.