Vanshiv Wins at Dreamforce: CampusCare Takes Community Choice and Builder Track Honours
Vanshiv Technologies is proud to share that our team won not just one but two prizes at the Dreamforce 2026 Agentforce for Good Hackathon. Out of more than 40 entries from across the world, our entry CampusCare earned:
Community Choice Award
2nd Prize in the Builder Track ($1,500)
There were 20+ judges from Salesforce, many of them EVPs, SVPs, VPs and above, leaders responsible for driving innovation at Salesforce at scale. CampusCare came out on top from among a pool of 40+ submissions, and six finalist teams took the Dreamforce stage. For Team Vanshiv, this moment further solidifies our positioning as global AI, Data, and CRM experts.
“I’m super proud of our Vanshiv team for winning not just one but two prizes in the Agentforce for Good Hackathon. CampusCare shows what we believe AI should do: give time back to people doing essential work, keep judgement with the human, and make sure nothing critical gets dropped. Congratulations to Saurabh Kumar, CA Ankur Sharma, and everyone at Vanshiv who stood behind this build.”
- Gaurav Kheterpal, Founder & CEO, Vanshiv Technologies
What Agentforce for Good stands for
Agentforce for Good is Salesforce’s employee-powered hackathon, run with volunteers across Equality Groups and open to employees, customers, and partners. Teams pick a real problem such as accessibility, mental health, housing, or other equity challenges, and build an AI solution that works for everyone, not just the majority.
In 2026 the programme drew 43 submissions from 11 countries. The theme, AI for All: technology that empowers everyone, matched how we approach Agentforce delivery at Vanshiv. We start with the person closest to the problem, then let the technology follow.
CampusCare: an AI copilot for school nurses
CampusCare is a WhatsApp-based AI copilot for school nurses, built natively on Salesforce Agentforce. It was designed and demoed by Saurabh Kumar and CA Ankur Sharma.
A school nurse often sees a child and then completes five separate pieces of admin for one visit: check the care plan, remember whether this child has been in before, decide if the family needs to be told, write the message, and complete the log. Multiply that by twenty visits a day and the step that gets dropped is usually the one only a human could have caught in the first place.
CampusCare exists so nothing important is dropped. The nurse sends one WhatsApp voice note or text message in her own words. No forms. No login. Nothing to install. From that single message, the agent finds the student, checks the visit against the documented care plan, looks for patterns in recent history, files a compliance-ready record in Salesforce, and notifies parents or guardians in the right language when needed.
In the Dreamforce demo, that looked like this in practice:
A diabetes visit logged in one sentence, with the care plan checked and the mother emailed automatically
A quiet message that did not sound unusual, until the agent flagged a third visit in 30 days and recommended a care plan review
A severe asthma attack where 911 had already been called: the agent made no medical decision, but made sure the mother knew immediately on email and WhatsApp
A Spanish-speaking family receiving the full notification in Spanish, including subject line and labels, not as an afterthought
That is what AI for All looks like when it is built for real school health offices, including under-resourced ones where a third of schools in the US do not have a full-time nurse.
How the system works
CampusCare runs on Salesforce using Agentforce, Data Cloud, Apex, and Twilio. Architecturally, trust is split on purpose:
WhatsApp and Twilio carry the nurse’s message in and the confirmation reply out
A Salesforce Site guest user is the untrusted entry point, allowed only to write a minimal intake record
A record-triggered Flow hands off asynchronously once the record already exists inside the org
The CampusCare Agentforce agent reads the message and decides what the visit needs
Apex actions validate the care plan, check visit patterns, file the visit log, and send notifications
Data Cloud grounds the agent in Student, Care Plan, and Visit Log data, and powers pattern detection
Each part of the chain is only permissioned to do the one thing it is meant to do. The agent is instructed to tell a genuine emergency from a routine visit, choose the right language, and hand off actions for care-plan checks, history review, filing, and notification. The nurse’s judgement stays central. CampusCare makes sure that judgement is never walking alone.
Why Community Choice matters
The Builder Track prize recognised technical strength judged by Salesforce executives. The Community Choice Award recognised something else: the project people connected with when finalists took the stage. Winning both says CampusCare worked as engineering and as a story people believed in.
For Vanshiv, that dual win is the point of how we build with Agentforce. Platform depth without empathy is a demo. Empathy without architecture does not survive production. CampusCare is our proof that partners can ship both.
See the build
Watch the Dreamforce demo on YouTube: CampusCare Agentforce for Good presentation.
Explore the open project on GitHub: github.com/saurabhhGitHub/campuscare.
Great work to Saurabh Kumar, CA Ankur Sharma, and the broader Team Vanshiv who made this possible. Thanks also to the organisers who run Agentforce for Good at Dreamforce, including Alexandra and the Equality Group volunteers who keep this programme focused on impact.
If you are exploring Agentforce for mission-driven or operational use cases, or want a partner who can take AI for Good ideas into production Salesforce architecture, talk to us at hello@vanshiv.com or visit www.vanshiv.com.
