Students don't drop out.
They disappear, and the signs
are there weeks earlier.
Boom detects the behavioral signals that precede a stop-out, then reaches students over SMS to clear the barrier before the drop happens.
Four weeks between detection and disappearance. That's the window Boom operates in.
37%
of community college students don't return for year two
Wk 4–6
typical gap between first disengagement signal and actual stop-out
1:800
average advisor-to-student ratio at open-access institutions
0
behavioral signals most existing tools act on before grades drop
The argument
Detection, outreach, and review — one continuous flow.
Existing tools hand advisors a dashboard and ask them to act. Boom detects, reaches the student, and loops the advisor in — all before the next staff meeting.
Ingests behavioral data, classifies risk against Tinto's departure framework, surfaces who is disengaging and why.
In practice
A student's LMS logins drop for two consecutive weeks. Signal flags the pattern and identifies a registration hold as the likely barrier.
Signal
detection layerIngests behavioral data, classifies risk against Tinto's departure framework, surfaces who is disengaging and why.
In practice
A student's LMS logins drop for two consecutive weeks. Signal flags the pattern and identifies a registration hold as the likely barrier.
Reach
autonomous SMSStudent-facing text outreach. Clears registration holds, walks students through FAFSA errors, protects them from bursar drops.
In practice
Reach texts the student: their immunization record is missing. The student snaps a photo and texts it back. The hold clears.
Copilot
advisor surfaceStaff review, edit, or approve every action. Every decision trains the system for the next student.
In practice
Copilot routes the immunization record to the registrar, logs the resolution, and surfaces it in the advisor's morning brief.
Workflows in the wild
Flagship workflows, each tied to an office.
The SMS proof
This is what autonomous outreach looks like.
A real exchange — anonymized — between Boom and a student with a FAFSA error. No portal login. No phone tag. No staff time. The barrier gets identified, walked through, and resolved in the same thread.
"I don't even know what I did wrong."
— Maria, student, 9:04 AM
Boom AI
● automated
Hi Maria — your FAFSA has a fixable error on parent income. Want me to walk you through it?
I don't even know what I did wrong
No worries. You left a field blank. Can you text me the number from your parents' 1040?
Hold on... ok, $42,800
Got it. I've submitted the correction. Your aid hold lifts in 24 hours.
You're clear to register for Fall, Maria.
Time to value
Forty-eight hours from signed agreement to first student conversation.
The category is known for six-month integrations. Boom doesn't integrate with your SIS — it ingests a file and starts working.
Hour 0 — drop your CSV
Your registrar or financial aid lead exports a daily CSV of flagged students and drops it in a secure folder. No SIS integration project. No IT ticket.
Hour 48 — first student reached
Boom ingests the data, builds risk profiles, and sends the first outreach. You see the conversations in Copilot. Students are already responding.
Ongoing — autonomous operation
Boom continues monitoring, reaching out, and clearing barriers. Your team reviews and approves. The system learns from every interaction.
Evidence
Early results from a pilot community college partner.
~41%
reduction in summer melt compared to the control group, measured in a quasi-experimental design at a pilot partner institution.
Preliminary and in progress
These are preliminary results from an ongoing quasi-experimental study. The sample is a single institution. The study has not been peer-reviewed. We're sharing it because institutional buyers deserve to see the evidence as it stands, not after it's polished. Full methodology is available on request.
FERPA posture
Student data handled under FERPA-compliant agreements. No data leaves the institutional boundary without written authorization.
Human escalation
Every conversation can be routed to a human in one message. Crisis-language detection triggers immediate handoff to a counselor.
Human-in-the-loop
Staff review, edit, or approve every action before it reaches a student. The system never goes silent on a student without a human able to see what happened.
"What if a student texts back something alarming?" — we answer that question in detail on the trust & safety page.
Read the trust & safety pageMatthew West
Founder · school psychologist, 14 years
I spent fourteen years as a school psychologist on under-resourced campuses in Los Angeles. I watched students disappear. Not drop out — disappear. They'd stop logging in, stop responding, stop showing up. By the time anyone noticed, they were gone. The signs were always there, weeks earlier.
Boom is the detection system I wished existed. It watches the signals. It reaches students over SMS before the drop. And it gives the people on your team the time to do what only humans can do.
Book a demo.
Forty-eight hours from this conversation to your first student reached. Drop your email and we'll send a calendar link.
No commitment. We'll show you the system on your data.