Written by Ronald Poon | 10 min read
Generative AI (GenAI) is the talk of the town right now, and for good reason. From being the main feature and point of discussion that’s dominating large tech conferences — to reports from the World Economic Forum suggesting that more than 75% of companies plan to adopt and implement this technology over the next five years.
The impact of this technology is far and large, with the ability to 10x employee output and transform ‘almost every task’. Not to mention the fact that things in this space are moving very, very quickly.
Getting ahead of this trend is paramount to large companies that want to utilize and make the most of the competitive advantage they currently have. Unlike other trends such as blockchain or mobile, GenAI benefits the incumbents, who possess the most significant asset when training and creating new LLMs (large language models): having large stores of data.
In the realm of business, akin to the world of sports, momentum and advantages are ephemeral. The current strategic opportunity presented by GenAI offers a limited timeframe. It is imperative for Fortune 500 corporations to act promptly, prioritizing the upskilling of employees and the development of unique sales assets.
The benefits of running a large-scale global GenAI Hackathon
There are two key reasons why a company should consider running a generative AI hackathon:
- Rapidly develop hundreds of GenAI proof-of-concepts (POCs) to enhance internal workflows, upgrade customer-facing products, or launch new market-ready solutions. For consulting firms, these POCs can serve as instant sales assets for both existing and prospective clients.
- Enable and upskill employees on the latest and rapidly evolving landscape of GenAI tools (such as Google Cloud’s Vertex AI Platform, Cohere’s Command model or Copy.ai’s Workflows product). Hackathons are one of the best ways to take advantage of and learn about any new technology inflection.
We recently achieved these goals with our client Capgemini, who partnered with Google to host their first-ever global generative AI hackathon.
How Capgemini’s Google Cloud based GenAI Hackathon came about
The Capgemini x Google GenAI Hackathon was initiated with a clear mission: to amplify Google Cloud capabilities within Capgemini and its workforce. Together, the two large companies partnered together to launch a brand new Google Cloud GenAI COE (center of excellence). Capgemini's vision was to carve out specific use cases across diverse client industries using Google’s generative AI tools.
Following their Google Cloud training summit, Capgemini was eager to find a way for their employees to put their fresh knowledge into immediate use. That's when they turned to us, Onova — an innovation consultancy specializing in running hackathons for large Fortune 500 companies.
As founders and investors first, operating at the forefront of generative AI, our team members aren’t just your typical hackathon event organizers. Capgemini was convinced of the expertise of our team and chose us to tackle their arduous challenge, which was: to execute the planning for the entire event in just three weeks. Our team was more than ready to deliver.
Planning a successful GenAI Hackathon
The Capgemini Steering Committee and Onova Project Team
Given the condensed timeline, the hackathon's success hinged on having a dedicated and committed project team. Our Onova team was up for the task, often pulling late nights and weekends to ensure that milestones and deliverables were completed on time.
Supporting this core team were four highly skilled developers and designers. Their expertise wasn't just limited to the branding and design for the event. They were instrumental in crafting and managing the hackathon platform, Earth, which was the backbone of the event's operations and the participant experience.
On the client side, it was important to ensure that the team was lean in order to make decisions quickly, but also had the right stakeholders to remove roadblocks. The Capgemini Steering Committee comprised of five key members that helped to approve marketing assets, and key event design decisions, as well as expedite technology access and legal procedures.
Semiweekly Project Meetings
We scheduled semiweekly meetings with the client every Monday and Thursday at 11 AM ET, considering the time zone differences spanning Europe, North America, and India. Each session, lasting 30 minutes, was concise yet highly productive.
This semiweekly cadence, chosen over a weekly or biweekly approach, ensured consistent accountability among team members. It was instrumental in guaranteeing that the hackathon preparations progressed seamlessly, allowing us to commence the event at the end of the three-week period as planned.
A Multi-Stage Project Plan
Drawing from our experience of orchestrating large-scale, two-day in-person hackathons for large companies such as BMO and McDonald’s, we adapted our proven formula to craft a two-week virtual event. This event was not just a hackathon; it was an immersive experience enriched with design-thinking sessions and AI-focused innovation workshops.
Our project plan was meticulously segmented into distinct phases, each with its own set of workstream deliverables. These workstreams spanned a broad spectrum: marketing and design, tools, mentors, judges, challenge formulation, global logistics, learning workshops, legal and finance, and hacker experience. Key milestones included launching registrations, designing marketing materials, recruiting judges and mentors, facilitating the team formation process, conducting workshops, and overseeing participant submissions.
Given the compressed timeline, these stages intertwined seamlessly, demanding a synchronized effort across all fronts. Organizing a hackathon is no small feat, however, having an organized and detailed project plan does make things a whole lot easier.
Compromises involved with a compressed timeline
1. Swag Packs
One of the standout elements of our hackathons, both in-person and virtual, has consistently been the participant swag. Swag isn't just about the tangible items; it's about the sense of belonging and camaraderie it fosters. Receiving swag packs helps participants feel integrated into a larger collective team. Moreover, these items serve as cherished mementos, celebrating and commemorating the hackathon, long after its conclusion.
We collaborate with SwagUp to provide premium branded swag kits, which include items like t-shirts, pickleball paddles, laptop sleeves, drink bottles, and trading pins, to our participants worldwide. However, given the tight three-week planning window for our recent event, it was not viable to design, produce, and distribute the swag before the event's conclusion.
For hackathons that spotlight branded swag, such as 'Burger Hack' or 'Destination Digital', our consulting team requires a lead time of at least three months to ensure the timely design, production, and delivery of these packs.
2. Registration Numbers and Team Formation
Usually, our hackathons benefit from a three-month lead time to garner global participant registration. However, with the recent event constrained to a mere one-week marketing window, we had to recalibrate our expectations regarding registration figures and the number of organically self-formed teams.
Nevertheless, the response to our event was overwhelming. Within just a week of marketing and internal promotion, we witnessed huge interest and demand for the event, with over 1000+ employees registering. This impressive turnout speaks volumes about the dedication and expertise of our combined marketing and design teams. They crafted an engaging hackathon website, highlighted by features like RachelAI, and effectively communicated the event's significance, relevancy and value.
The takeaway from our experience is clear: while a shorter planning period might pose challenges in terms of potentially lower registration numbers and fewer teams forming, with a committed marketing team that possesses exceptional design skills, remarkable outcomes are still achievable.
3. Mentor Matching with Teams
In previous hackathons, we paired hackathon teams with relevant business and technical mentors, taking into account time zones and domain expertise.
For this event, given the shortened timeline, we managed to recruit only 16 mentors for the hackathon. Typically we would have a 10:1 ratio of participants to mentors for matching purposes, however, we were well off that mark in this case at 16 mentors for over 1,000 participants.
In response to the mentorship constraints, we introduced a dual approach: having floating mentors and a series of 30-minute live Mentor Office Hours. Floating mentors were not matched to specific teams, but instead were made accessible to participants and could respond to requests from any team. The office hours provided a platform for participants to seek live business guidance or technical mentorship.
This dual approach ensured participants had consistent access to expertise and assistance.
4. Programming and Number of Workshops
Our recent hackathon's timeline coincided with the July summer holidays in Europe and North America, including the U.S. Independence Day weekend. This overlap with various international holidays posed challenges, as participants worldwide were observing their respective breaks.
Given these constraints, we streamlined our hackathon programming, consolidating our content into four workshops instead of the usual five.
For future reference, we aim to schedule our two-week hackathons away from overlapping international holidays. We've observed that summer months, especially July and August, are less ideal due to vacation schedules. Historically, the Fall months (September and October), early spring (March), and the beginning of June have proven optimal for hosting hackathons.
Trends we noticed
1. Consumer Product Focus
A notable trend was teams' inclination towards consumer products, particularly within the beauty and cosmetics sectors. This suggests a heightened interest in addressing innovation challenges endemic to these industries.
2. Client-Centric Approach
Many participants anchored their projects to familiar client accounts, such as L’Oréal. This strategy underscored their profound grasp of client requirements and their adeptness at tailoring solutions to market needs.
3. Desire for Customization and Personal Choice
Nearly 40% of teams showcased a preference for crafting their own challenges, often blending elements from provided "use case challenges." This approach highlighted their flair for innovation and eagerness to devise distinct solutions.
4. Rise of AI Interfaces
AI-driven tools like chatbots, avatars, and voice ordering were most favoured by participants. This trend underscores a rising inclination to harness AI interfaces for improved user experiences.
Key learnings and takeaways
1. The Importance of Adequate Planning Time:
Our recent experience, while successful, highlighted the challenges of a compressed 3-week planning window. For virtual projects of a modest scale, we believe a 4-week timeframe is the bare minimum, with 6-8 weeks being ideal. For in-person generative AI hackathons, we strongly advocate for a 12-week planning process to adequately address logistics like swag, travel, and AV arrangements.
2. Leveraging Single Technology Partnerships:
Partnering exclusively with Google for this hackathon underscored the advantages of focused technology collaborations. While multiple tech partnerships have their merits, the outcomes here were evident: more cohesive solutions, tangible prototypes, and clearer alignment with client use cases. This approach not only enhances employee proficiency in a particular tool but also streamlines logistics, ensuring a level playing field for participants.
3. Vibrant Virtual Engagement:
The virtual setting didn't diminish participants' enthusiasm. Their global passion and creativity were palpable as they engaged in workshops and delved into Generative AI's intricacies. As a facilitator for the "How to Pitch Like a Pro!" workshop and the "AI-related" trivia master, witnessing participants' joy during live pitches and presentations was immensely gratifying. Their genuine excitement served as a heartening reminder of the joys of working in this dynamic field.
4. Large Companies Must Take Advantage of Being in “Pole Position”:
Being a major player within the technology and consulting industry, companies like Capgemini naturally possess a unique advantage in the realm of generative AI. As alluded to earlier, their edge over smaller entities and startups stems from two main assets: 1. expansive data reservoirs, and 2. robust resources.
It's imperative for corporate CIOs and tech department heads to capitalize on this unique vantage point. By strategically allocating financial and capital resources to the right programs, they can expedite the upskilling of their workforce in generative AI and develop valuable assets for their clientele.
Proactive investment in generative AI initiatives is crucial for these industry giants, ensuring that they stay ahead of agile, nimble competitors. To sit and wait otherwise, risks these large companies falling behind.
Drawing from the insights and reflections shared in this case study, it's evident that the GenAI Hackathon was more than just an event; it was a transformative experience for everyone involved.
The dedication, passion, and creativity displayed by all participants, mentors, judges and organizers have set a new benchmark for future hackathons.
We look forward to seeing how the solutions from this particular event continue to grow within the Capgemini and Google environment and are incredibly optimistic about the use of generative AI in workforces and companies globally.