Articles
Why AI companies sponsor hackathons, and how to measure it
What an AI company gets from sponsoring a hackathon, and a simple plan to measure adoption, feedback and hiring afterwards.
By Starfield LabsPublished
An AI company can sponsor a hackathon to put its model or API in the hands of builders, to watch real people use it under pressure, and to meet people it may want to hire. In our view, a well-planned sponsorship works like a short, focused product test with a community attached. A poorly planned one is a logo on a page. The difference is deciding what you want before the event and measuring it after.
What can an AI company get from sponsoring?
We find it useful to think about the value in three groups.
Product adoption. Builders sign up, get keys and try to ship something real with your product over the event. Some of those projects may keep going afterwards. Even the ones that stop can leave you with developers who know how your product works.
Product feedback. At a hackathon, you can watch people work through your docs in a few hours. You can see where they get stuck, which errors confuse them, which features they reach for and which they ignore.
Hiring and community. You meet builders while they are solving problems, not while they are interviewing. Your engineers can see how people think, communicate and recover when things break. A steady presence at events can also build trust with builders who may try your next release.
There can be smaller benefits too: example projects for your docs, stories for your blog, and a reason for your own team to step outside the roadmap for a few days.
What should you decide before the event?
Pick one main goal and at most two others. Then write down what success looks like for each, in terms you can count.
| Goal | What to decide upfront |
|---|---|
| Adoption | Which product or endpoint you want used, and what counts as "used" |
| Feedback | Which flows you want tested, and who on your team will collect notes |
| Hiring | Which roles you are hiring for, and who will talk to people |
| Community | Which channel you want builders to join afterwards |
Agree these with the organizer early. They shape what you offer builders, what you present and how many of your people need to be there.
How do you design a sponsorship that delivers?
A few choices are worth getting right:
- Make the first step easy. Provide access, a quick-start guide and a sample project that runs quickly. Friction on day one can cost you projects.
- Offer a short workshop near the start, if the event allows it. Show one working example end to end, then stay for questions.
- Send people if you can. Engineers who can answer questions also bring back the best feedback.
- Reward real use. If you set a challenge, frame it around how well teams use your product, not whether they mention it.
- Give builders a next step. A showcase, a community channel or a follow-up call can keep strong projects going.
How do you measure adoption after the event?
Start with what your own systems can see. For the event window and the weeks after, you could track:
- New sign-ups or keys created through the event's code or link
- How many of those made at least one successful call
- How many final submissions used your product
- How many accounts were still active some weeks later
Compare against a normal week so you can see the effect of the event, not just the totals. Agree with the organizer in advance what submission information you will receive.
Look at quality as well as volume. In our view, a few projects that use your product in depth tell you more than many sign-ups that never make a second call. Read the strongest submissions in full and note which features they relied on.
How do you measure feedback and hiring?
Feedback is easiest to measure as a list. During the event, have your team log every confusion, bug or missing feature they see, with the project it came from. Afterwards, group the items, count how often each came up, and turn the top ones into tickets. Then tell builders what you fixed.
For hiring, you could track the people your team flagged as strong, how many you followed up with, how many entered your process and how many you hired. Ask your team to note why each person stood out while it is fresh. Hiring takes time, so check in over several months, not just the following week. Before you contact participants, read the event's rules and ask your legal team what is allowed.
Keep a short write-up of each sponsorship with goals, results and costs. After a few events, you can compare them and see which formats work for you.
Where to go next
Write down one goal, one number to track and one follow-up action before you commit to any event. If you want to put your product in front of builders, partners can sponsor Starfield Labs events, and companies can also book private build days. You can see upcoming events at starfieldlabs.ai/events.