CampusCare
Voice-First · Multilingual · Zero Forms
An Agentforce agent for school nurses. One message logs a health-office visit, checks it against the student’s care plan, spots repeat visits, notifies the guardian in their language, and files the compliance record - before the nurse has left the child’s side.
Documentation that happens hours after the care
A school nurse is the only clinician in the building, covering hundreds of students alone. She triages all day and writes it up afterwards, from memory. Every student with asthma, a seizure disorder or a severe allergy has a care plan on file - but nothing checks her action against it while the child is still in the room, and nobody is watching what repeats.
- ✕Visits written up at the end of the day, from memory
- ✕Care plans filed, but never checked at the moment of care
- ✕Repeat visits that signal a worsening condition, visible to no one
A health-office visit logged in one sentence
The nurse says what happened, in her own words, on the channel already open on her phone. CampusCare does the rest: it identifies the student, validates the action against the care plan, checks the visit history, notifies the guardian, and files the record.
Clinical documentation that keeps up with the nurse
Documentation at the point of care
The compliance record is a by-product of the conversation, not a second task at the end of the day. Detail survives because it is captured while it is fresh.
Care plans that get checked, not just filed
A plan is only as good as the moment someone reads it. Every logged visit is compared against the documented protocol automatically, and a mismatch is surfaced to the nurse immediately.
Patterns nobody was watching
A third asthma visit in a month is a signal, not a coincidence. The threshold is set per student by a clinician, so the escalation is a care recommendation rather than a population average.
Built natively on the Salesforce AI stack
Three checks behind every message
CampusCare runs one Agentforce agent with two actions - these are the three checks that fire on every nurse message.
Care Plan Validator
Reads the student’s documented protocol and compares it to what the nurse actually did, quoting the plan back when they diverge.
Pattern Detector
Counts visits for the same complaint in a rolling window against the review threshold on that student’s plan.
Emergency Escalator
Classifies whether the visit was an emergency and, if so, reaches the guardian urgently.
From a spoken sentence to a filed record, end to end
Every visit is orchestrated through Agentforce - from a WhatsApp message to a guardian notice and compliance record in one transaction.
Request flow
How every nurse message moves through CampusCare
Nurse sends a message
Voice note or text over WhatsApp, in her own words.
Twilio webhook
Delivered to a Salesforce Site endpoint.
Session record & Flow
The message is stored and a record-triggered Flow invokes the agent.
Student lookup
Resolves the student and returns the guardian’s language as an instruction.
Care plan validation
The action taken is compared against the documented protocol.
Pattern check
Visit history is counted against the plan’s review threshold.
Notice & record
Guardian emailed in their language, visit filed in the same transaction.
The road ahead for CampusCare
Condition-aware protocol matching
Scope the care-plan check to visits actually related to the plan’s condition, so an unrelated minor visit raises no warning.
District-wide pattern detection
Follow the pattern across schools through Data Cloud, so a student who moves between buildings is still seen as one child.
Multi-turn conversations
Let the nurse correct or add to a visit in the same thread, rather than each message standing alone.
From memory to record
One sentence. Every check. Filed before she leaves the room.
Team Vanshiv - Agentforce for Good Hackathon, Finalist