If you’re heading to Dreamforce this year, the Agentforce for Good hackathon has popped up. Maybe a guest invite landed in your inbox, or a teammate forwarded the rules and asked, “Should we do this?”
I read the official rules. Let me walk you through the key dates, eligibility criteria, the two tracks, and how the judging really works.
Table of Contents
What Is the Agentforce for Good Hackathon?
The Agentforce for Good Dreamforce 2026 Hackathon is a Salesforce competition for customers and partners attending Dreamforce in person. You pick one of two tracks, Builder (team, code) or AI Fluency (solo, no code), build something focused on accessibility and social good, and compete for cash prizes at a live Showdown on September 16, 2026.
It’s sponsored by Salesforce and administered by Devpost. The theme of this event is “AI for All: Technology that Empowers Everyone.”
Key Dates you can’t miss (PT and IST)
- Registration opens: August 12, 2026, 12:00 pm PT
- Provisioned orgs delivered: August 24, 2026 (Builder Track)
- Submissions open: August 28, 2026, 6:00 AM PT (6:30 PM IST)
- Registration closes: September 4, 2026, 8:00 PM PT
- Submissions close: September 7, 2026, 5:00 PM PT (Sept 8, 5:30 AM IST)
- Finalists announced: September 14, 2026 (via registered email and the Hackathon Portal)
- The Showdown: September 16, 4:00 PM PT (September 17, 4:30 AM IST)
Who Can Actually Enter? Eligibility Requirements
To enter, you must be registered for and physically attending Dreamforce 2026 in San Francisco with a full conference pass. Dreamforce runs September 15–17 at Moscone Center. No pass, no entry. It doesn’t matter how good your agent is. This is not a virtual hackathon you can win from your desk in Bangalore or Berlin.
Salesforce employees can’t enter this one. The rules exclude anyone who’s been an employee, contractor, or consultant of Salesforce within the past two years. The hackathon is open to every Salesforce Partner and Customer attending Dreamforce – any role, any function, any level of technical experience.
A few more gates worth knowing up front. You need to be 18 or older, and a legal resident of one of the eligible countries. The list includes places like the US, UK, India, Australia, Canada (Quebec is specifically excluded), Germany, France, Japan (20+), and a handful more. Government employees are also out. If you clear all of that, keep reading. If you don’t, you just got your hour back.
Builder Track vs. AI Fluency Track: Which One Should You Pick?
You can only enter one track. Choose what you can realistically ship within the submission window, not what sounds more impressive.
| Builder Track | AI Fluency Track | |
|---|---|---|
| Team size | Up to 4 people | Solo only |
| Code required? | Yes - a working Agentforce agent | No |
| What you submit | Agent on a provisioned org + demo video | A written work product (three types) |
| Best for | Developers, admins, architects, technical PMs | Consultants, analysts, business folks, beginners |
| Top prize | $3,000 | $3,000 |
If you still can’t decide after looking at that table, ask yourself whether you would enjoy building it or would be intimidated by it. Usually the answer is found in this.
Builder Track (for developers, admins, and architects)
In this track, you will build a new, functional Agentforce agent designed to solve a problem related to social good, accessibility, or equity. The emphasis is on it being ‘new’ – it must be created after the submission period begins, not something you build last quarter. It must also run on the org provided to you by Salesforce (which will be available by August 24); do not build it in your own dev org and migrate it later. Judges will test using only the provided org, and using anything else could invalidate your entry.
This submission involves more than just the agent itself. You need to record a demo video (maximum five minutes, though judges might stop watching after three, so showcase the key highlights early on), explain what the agent does and who it helps, provide admin credentials and the Org ID for the judges to test it, and answer two brief questions: why the agent is superior to traditional solutions, and how you addressed responsible AI and agent observability. That second question isn’t just for show; responsible AI actually earns you points, which I’ll explain later.
If you’re newer to Agentforce, this is the track where a little groundwork pays off. Knowing how agents, actions, and topics fit together before the org arrives means you’re building on day one instead of reading docs.
The Builder submission checklist:
Copy this and tick it off before you hit submit:
- A new Agentforce agent, built on the provisioned org, after Aug 28
- Demo video (5 min max, publicly viewable on YouTube, Vimeo, or Google Drive)
- Written description: features, impact, and who it’s for
- Admin credentials for the provisioned org
- The Org ID
- A 1–2 sentence pitch
- Your answer: why is an agent better than a traditional solution?
- Your answer: how did you handle responsible AI and observability?
- Optional but smart: a public GitHub repo link
The AI Fluency Track, deep dive
No code, solo, and grounded in at least one of the AI Fluency curriculum modules from the hackathon portal. What you’re really being judged on is how well you collaborate with AI to solve a real accessibility, trust, fairness, or inclusion problem.
You will pick one of three submission types, and here’s the part worth getting right: the type you choose doesn’t change your score. All three are judged on the same criteria, so choose the format that fits your idea, not the one you think scores higher.
- AI Concept Brief (750–1,500 words): a written proposal for an AI-powered solution. No prototype needed, but the thinking has to be deep enough that someone else could build it from your brief.
- Human-AI Collaboration Showcase (500–1,000 words): a real, working AI workflow you actually use, showing where AI does substantive work alongside you, with evidence like screenshots or before-and-after samples.
- Accessibility Audit + AI Redesign (750–1,500 words): audit an existing tool or process for accessibility gaps, then use AI to propose a redesign, with a clear before-and-after.
How Judging Works, and How to Actually Score Well
Both tracks are scored on weighted criteria. For the Builder Track: Social Impact and Ethical Focus (20%), Responsible AI (20%), Accessibility (20%), Originality (15%), Demonstrability (15%), and Scalability and Adoption (10%). For the AI Fluency Track: Impact and Insight (25%), Application of AI Fluency Skills (25%), Accessibility (20%), Creativity and Originality (15%), and Clarity and Communication (15%).
Now read those numbers as a strategy, not a list. On the Builder side, Responsible AI plus Accessibility is 40% of your total. Add Social Impact, and it’s 60%. So an agent that’s technically outstanding but treats accessibility as an afterthought is leaving most of the points on the table.
One more quiet tell: the tie-breakers. If two Builder entries tie, the higher Accessibility score wins first. For AI Fluency, it’s the AI Fluency Skills score. That ordering tells you exactly what the organizers care about most. Prioritize accordingly.
Prizes and Special Awards
The money is solid for a hackathon of this size, and it’s identical across both tracks:
| Placement | Prize (USD) |
|---|---|
| 1st Place | $3,000 |
| 2nd Place | $1,500 |
| 3rd Place | $750 |
In addition, there are several special awards worth $750 for which you are automatically considered without needing a separate application: Accessibility Excellence, Equality Group Champion, Headless Hero (Builder track only), and Agent Observability (Builder track only). There is also a ‘Community Prize’ for the entire hackathon, decided by the live audience.
How to Register for Hackathon
Register through the Hackathon Portal on Devpost. Registration closes September 4, 2026 at 8:00 PM PT.
Portal: Click here
Questions: support@devpost.com
Final Thoughts
Before you commit, do two quick things. Confirm your Dreamforce 2026 pass is the full conference pass, since that’s the gate everything else hangs on. Then register early; the portal opens August 12, and registering sooner gives you more time with the provisioned org when it arrives on the 24th.
And regardless of the approach you choose, before writing a single line of code or a single word of your brief, make sure to map out your accessibility strategy. That is the crucial part in either case; get that right, and the rest is simply a matter of implementation.
Frequently Asked Questions (FAQ)
Yes, you need a full Dreamforce 2026 conference pass and must attend in person, and finalists present live at the Showdown on September 16, 2026.
No, this edition is for customers and partners. Anyone who’s been a Salesforce employee, contractor, or consultant within the past two years is excluded.
Builder is for teams of up to four building a working Agentforce agent. AI Fluency is solo, no code, and centers on a written work product that applies AI thinking to a real challenge.
Not for the AI Fluency Track. It’s a no-code track. The Builder Track requires building a working agent in a provisioned Salesforce org.
First place in each track is $3,000, second is $1,500, third is $750, plus several $750 special awards you’re automatically considered for.
Submissions close September 7, 2026, at 5:00 PM PT (that’s early morning September 8 in India). Registration closes earlier, on September 4, 2026.
Priya Rastogi
Priya is a 3x Salesforce Certified who believes in the power of continuous learning and collaboration. She’s passionate about exploring how Salesforce can simplify work, boost productivity, and create better user experiences. When she’s not experimenting with new features or automating processes, Priya enjoys connecting with fellow Trailblazers and sharing insights to help others grow in their Salesforce journey.
- Priya Rastogi
- Priya Rastogi
- Priya Rastogi
- Priya Rastogi
