Stop paying aid to students who don't exist.
Fake students took $90 million in federal aid last year. Rollcall is how colleges stop it.
Risk queue
Last sync 9:14 AM · 2,311 enrollments, 18,402 submissions · next sync 10:00 AM
Aid at risk before census
Scheduled Pell and SEOGHold recommended
$371,240 already held
0 automatic drops
Verified real
3.2% of holds reversed
after a live check-in
Needs a decision before Sep 26
Federal guidance already asks for this.
In May 2025, Federal Student Aid told schools to confirm attendance and academic engagement before refunds go out, and to consider live verification for higher-risk cases. A school that finds conflicting information has to resolve and document it before paying more aid. Rollcall is how a financial aid office does that for every roster, before census.
Staff see the evidence and make the call.
Each flagged account opens to the coursework and enrollment records behind it. A hold is only suggested when at least four signals agree, and someone on your team has to approve it.
Marisol Teague
Awaiting decisionW0041-7728 · Sierra Valley College · $3,712 Pell scheduled Sep 29 · opened Sep 22 by Marcus Lee
Risk evaluation
Hold recommendedRegistered in Banner in the same 11-minute window as 22 other accounts, all choosing the same 4 CRNs.
Submission device fingerprint shared with 8 other accounts in C-14.
ENGL 101 Essay 1 matches six other essays in the ring at 96%, shared paragraphs in the same order.
3 submissions in 4 weeks, each within 2 minutes of the deadline. No discussion posts or messages to instructors.
Checked by the aid office at disbursement.
First term at the district.
Evaluated against policy v3 (effective Aug 4, 2026). LMS access logs are not used. Attendance follows the federal definition: submitted work counts, logging in does not.
Submission comparison
ENGL 101 Essay 1 · 96% shared textGrowing up near the river taught me that patience is a kind of strength most people overlook. Every spring the water rose and every spring my grandmother said the same thing: wait, and the ground will dry. I think about her when school gets hard.
When I was a kid by the lake I learned that patience is a kind of strength most people overlook. Every spring the water rose and every spring my grandmother said the same thing: wait, and the ground will dry. That lesson stays with me now.
Activity
Registered · 9:42 AM, inside the C-14 window
Aug 18ENGL 101 Essay 1 submitted · 11:58 PM, due 11:59 PM, 96% match with six ring accounts
Sep 2MATH 110 Homework 2 submitted · 11:57 PM, due 11:59 PM, identical answer order to 5 accounts
Sep 9Discussion 1 missed · no discussion activity in any course this term
Sep 16HIST 120 Quiz 1 submitted · 100% in 11 minutes, same completion time as 4 accounts
Sep 20Pell disbursement $3,712 · will be paid unless a hold is filed first
Sep 29Notes
Marcus Lee Sep 22, 11:05 AM
Spoke with Prof. Alvarez (ENGL 101). She has never had a reply from this student in the discussion board, and the essay reads like the five others she flagged last week.
Term report
Fall 2026 · Aug 18 to Sep 21 · draft until census closes Sep 30 · TR-2026F-SVCCD
Aid held before disbursement
Verifications logged
184Holds reversed after check-in
3.2%Aid re-released within one business daySeats returned to waitlist
970 automatic dropsReferred to the Inspector General
3Every flag closed with a recorded outcomeBy college
Decisions by reviewer
Rings identified
| Ring | Pattern | Accounts | Devices | Held | Outcome |
|---|---|---|---|---|---|
| C-14 | Aug 18, 11-minute registration window, same 4 CRNs, 3 devices | 23 | 3 | $84,960 | Holds pending |
| C-09 | Aug 20, 6-minute registration window, 2 devices | 11 | 2 | $38,120 | Held |
| C-11 | Aug 21, duplicate ENGL 101 essays, no shared device | 9 | 0 | $31,500 | Held |
| C-03 | Aug 24, one device, under review | 6 | 1 | $9,840 | In review |
| C-07 | Aug 19, all five attended check-ins | 5 | 1 | $0 | Cleared |
Signed reports
Signatures recorded with timestamp and IPA report your auditor will accept.
At the end of the term you get a signed record of what was checked, who decided, and how much aid was held, broken out by college.
Pilot results
Sierra Valley CCD · 4 colleges · Fall 2025
$212,900in aid held before it was disbursed. 2.9% of holds were reversed after the student showed up for a check-in.
Bayshore College · single campus · Spring 2026
141 seatswent back to waitlisted students before census. Every one was released by a staff member.
Pacific Online University · Spring 2026
6 ringsfound through batch sign-ups and shared devices. Three were referred to the Office of Inspector General.
of applications to California community colleges in 2024 were fraudulent, according to the system.
in aid was still paid to fraudulent applicants in 2025, even after federal screening blocked over $1 billion.
is enough to run a pilot, from the first day of registration to census.
Nine questions we get asked.
From aid directors, deans, and the people who sign the check.
No automatic drops
Rollcall can suggest a hold. It cannot drop a student. Only a staff member can.
Notice and a way to respond
A student is told when their file is flagged and has a fast way to clear it. No hold takes effect before that notice goes out.
No proxies for who you are
Scores never use IP location, names, writing style, or assistive technology, so a flag cannot track national origin or disability.
No access logs
Your LMS provider has said access logs should not be used for decisions like this, so Rollcall does not use them. It works from coursework and enrollment records.
Plain explanations
Every score shows the reasons behind it in ordinary language, so a reviewer can explain it to a student or a dean.
Read-only
Rollcall only reads from your systems. It cannot change a grade, an enrollment, or a payment. It runs under FERPA, the GLBA Safeguards Rule, and SOC 2 Type II.
Your data stays yours
Each district runs on its own records. Rollcall does not score your students against other schools' data.
Every flag gets an outcome
Schools must refer credible fraud to the Inspector General. Each file closes with a recorded decision and can export as the referral.
Flat pricing
One annual price per college. Never per flag, and never a share of the aid saved.