A first destination survey past the 65% knowledge rate, with every graduate's outcome traced to its source and two names
A knowledge rate past 65%, with every graduate's outcome traced to its source.
See one case, screen by screen ↓A first destination survey records where a university's graduates went after their degree: a job, further study, service or still seeking. First-Destination Survey & Career Outcomes lets career centers verify graduates' jobs and further study from surveys and records, per NACE: agents follow up with graduates, match the national enrollment file, employer confirmations and profiles, and propose each outcome with its source and rule cited. A staff member confirms every outcome, a second person confirms every change, and the director signs the class report.
AI in education: a buyer's guide to agentic AI for colleges, universities and school districts →
Updated
A first destination survey answers one question each December: where did last year's graduates go? The NACE First-Destination Standards and Protocols count every completer from July 1 to June 30, ask a career center to know the outcome of at least 65% of the class, from the graduate or from other sources it has verified, and close the count on December 31. Placement and earnings now carry state and federal consequences, so the numbers have to hold up when a reviewer asks for one graduate.
Most centers still build the number by hand. The survey goes out at graduation and again in the fall, and about a third of the class answers. Staff search public profiles one graduate at a time, email employers and submit the class to the national enrollment match service, and each source lands in its own spreadsheet. In December an analyst merges the sheets, picks one outcome per graduate and codes it to a NACE category; when a survey and an employer disagree, the top row usually wins, and nobody records which source settled which graduate or who decided.
Estimated staff days for one graduating class of about 10,000, by hand and with the agents. Graduates still take their own time to reply; these are the days the office spends. Our estimate, kept modest.
Seven agents gather, match and classify every graduate's outcome; a staff member confirms, a second person checks, the director signs.
Halden State University, class of 2026: 9,812 completers in 148 programs. It is Monday, November 30; the count closes December 31, and the state system's fall outcomes review is due tomorrow. Here is the day, screen by screen, in the working solution.
Tomás opens the class to one number: 61.8%, 6,064 of 9,812 graduates known, 314 short of the dashed 65% line. The column fills in four labeled layers: 3,212 survey answers, 1,236 employer confirmations, 588 verified profiles and 1,028 enrollment matches. Beside it, where the column stood on the first of each month, from 22.4% in August to 57.3% in November.
A pale layer above the fill shows 565 graduates found but not yet countable.
Overnight the Verifier read the national enrollment match return file received November 27 and matched 2,104 of 9,812 graduates on Halden ID, name and birth date. 1,678 were already known; 412 had no known outcome, and the Classifier proposed continuing education for each under NACE §3. Every row shows where the graduate enrolled, full or part time, the start date, the file row and the confidence; 14 name-only or withdrawn rows are held for a person.
Confirm stays disabled until Tomás has checked a 20-row sample against the file. He confirms the 412: the enrollment layer grows from 1,028 to 1,440, the class reaches 66.0%, 6,476 of 9,812, and the programs under 65% fall from 19 to 14.
Past the line by 98.
Jordan Whitfield, B.S. Statistics, answered the survey on August 14: Data Analyst, employed full time, Northfold Logistics, $68,000. The employer's letter of November 24 says a contract through a staffing firm at 20 hours a week. Both sources sit side by side with the passages highlighted, and the Classifier proposes employed part time, temporary/contract assignment, at confidence 0.94: under 30 hours is part time (§3), and the latest known outcome is reported (§5).
Salary is collected for full-time jobs only, so the $68,000 leaves the salary figures.
The citation opens the employment verification letter as received, page 1, paragraph 2: "engaged as a contract data analyst through Brightwater Staffing at 20 hours per week, from October 5, 2026 to March 31, 2027." Paragraph 3 notes that the letter does not state compensation. Opening it is written to the audit trail with Tomás's name.
Tomás resolves the record to part time with reason CO-01, "Employer document outranks the survey". He cannot confirm his own change, so Ruben Castellanos, Assistant Director, Career Outcomes, confirms it at 5:14 PM. Employed full time goes from 3,489 to 3,488, part time from 401 to 402, and reported full-time salaries from 2,214 to 2,213; the median stays $64,500.
From degree conferral on May 16 to the second confirmation: the survey sent by text in English, Lena Kowalczyk's first confirmation on August 17, the public profile of September 30, the employer request under the consent Jordan gave, the letter, the conflict and both decisions. The record is visible to career services and Institutional Research only.
The Collector has drafted the Arts and Sciences follow-up for 327 unknown graduates of B.A. Psychology, Communication and History, by email and text, in English and Spanish, signed by their liaison Owen Marsh. It goes out Tuesday at 11:00 AM, outside quiet hours, and only after a liaison or the assistant director approves. The 187 graduates who opted out are never messaged.
Imogen Ashworth, Director of Career Services, signs the class report at 5:30 PM: knowledge rate 66.0%, career outcomes rate 92.1% (5,883 of 6,389), median full-time salary $64,500 from 2,213 answers. At 5:41 PM Grace Odum checks that the 9,812 completers tie to the IPEDS completions file, that no outcome lacks a confirmed source and that cells under 10 are suppressed. At 5:48 PM it goes to the state review.
Reference SFR-2026-HSU-1130.
Fifteen entries cover the day and what led to it, from the Verifier's 2:14 AM match to the 5:48 PM submission, people and agents alike, each with its reason and evidence. Walter reads everything, changes nothing, and exports a redacted evidence pack for the state, a federal reviewer, an accreditor or NACE.
The same class of 2026, seen by the five people who carry it: what their fall looked like, and what it looks like now.
Seven agents gather, match and classify every graduate's outcome; a staff member confirms, a second person checks, the director signs.
Runs the class from roster to report, routes every item to a named person and keeps the service-level and close-date clocks.
Builds the class from completers between July 1 and June 30 by degree and term, ties it to the completions count, and carries each graduate's consent and directory hold.
Follows up with graduates still unknown, aimed at the programs under 65%, in each graduate's preferred channel and language, and reads the replies into the record.
Matches graduates against the national enrollment match file, employer confirmations and public profiles, and counts only the sources the protocol allows, verified (NACE §2b).
Finds graduates whose sources disagree on status, hours, employer or enrollment, and lays both sources side by side with the passages.
Assigns the NACE category and sub-category with the section cited, applies the latest-known rule, and keeps salary to full-time jobs.
Builds the class report, the program evidence and the export packs, suppresses cells under 10 and redacts what the audience may not see.
Confirms each outcome; a second person checks every change. The agents propose; a named person decides.
Ask about any graduate, program, source or rule in the class, or say what to change. Every answer points to its source and opens the screen it talks about.
Why did the knowledge rate move today?
Because 412 graduates with no known outcome are in the national enrollment match file received November 27, each enrolled in a graduate or professional program this fall. Enrollment after the degree is continuing education under NACE §3. Tomás Ferreira checked a 20-row sample and confirmed them at 4:52 PM, so the enrollment layer grew from 1,028 to 1,440 and the class went from 61.8% to 66.0%, past the 65% line.
Why would Jordan Whitfield be part time?
Two sources disagree. Jordan's survey of August 14 says data analyst, employed full time, $68,000. Northfold Logistics' letter of November 24 says a contract through a staffing firm at 20 hours a week. Under 30 hours is part time (§3), and the latest known outcome is reported (§5), so the Classifier proposes part time, temporary/contract, and takes the salary out of the full-time figures. A person resolves it and a second person confirms.
Which programs are still under 65%?
14 of 148, down from 19 this morning. The largest gaps: B.F.A. Studio Art 47.0%, B.A. History 53.1%, B.A. Communication 56.5%, B.A. Psychology 56.9%, B.S. Civil Engineering 61.2%, M.S. Data Science 61.5% and B.S. Electrical Engineering 63.0%. The Collector has drafted four follow-ups covering 562 of their unknown graduates, each waiting for the college liaison.
Draft a follow-up for engineering graduates we don't know yet
Drafted for the 140 unknown graduates of B.S. Civil Engineering and B.S. Electrical Engineering: by text for the 96 who chose it and by email for 44, in English and, for 9, Spanish, signed by Nadia Petrosyan and timed outside quiet hours. Nobody who opted out is on the list. Nothing goes until Nadia approves.
13 screens from the working solution, on its sample data. Pick one to see it large.
Which sources count, the target, the clocks and who confirms are settings the office owns, applied the same way to the whole class.
| Setting | Default | Choose from |
|---|---|---|
| Sources that count (NACE §2b) | Survey answers, employer confirmations and enrollment matches as received | Public profiles and faculty reports after verification; a fellow graduate or parent only after the graduate confirms |
| Knowledge rate target | 65%, for the class and every program | Stepped, with the programs under the line shown (NACE §2c) |
| A person reads anything below confidence | 0.95 | Stepped; the Verifier holds rows under the floor |
| A batch needs a checked sample | 20 rows | Rows opened against the source before a batch is confirmed |
| Who second-confirms a changed record | Ruben Castellanos, assistant director | Imogen Ashworth, director; never the person who made the change |
| A conflict is resolved within | 5 business days | 3 or 10 business days; then it escalates |
| Suppress cells under | 10 graduates | 5 graduates; applies to any published or exported figure |
| Keep records for | 6 years after the class closes | 3 or 10 years; legal hold |
Every figure is labelled: a target the solution is built to, an estimate, a typical published result, or a proven one.
Sources: NACE, First-Destination Standards and Protocols for undergraduates (approved July 2019; revised 2025), §1–§5: the class of July 1 to June 30, sources and verification, the 65% knowledge rate, the 30-hour full-time line, the career outcomes rate and the December 31 close · Workforce Pell final rule, 2026 (NAICU Washington Update, May 22, 2026; NCAN summary): 70% completion and 70% job placement for programs of 150 to 599 clock hours · "demo" = seen in the working solution, on its sample data · "target" = the design goal, measured in the live solution · "typical" = published figures · "estimated" = our estimate. Halden State University, the companies, institutions and people named on this page are fictional: characters and sample data in the working solution.
A first destination survey records where a university's graduates went after their degree: a job, further study, service or still seeking. First-Destination Survey & Career Outcomes lets career centers verify graduates' jobs and further study from surveys and records, per NACE: agents follow up with graduates, match the national enrollment file, employer confirmations and profiles, and propose each outcome with its source and rule cited. A staff member confirms every outcome, a second person confirms every change, and the director signs the class report.
The knowledge rate is the share of the graduating class whose outcome the institution knows, from the graduate or from other sources it has verified. The NACE First-Destination Standards and Protocols ask institutions to strive for at least 65% (§2c). The solution shows it as one column filled by source, for the class and for each program, so no program hides behind the class average.
The office sets it, following NACE §2b. In the working solution, survey answers, employer confirmations and national enrollment matches count as received; public profiles and faculty or staff reports count only after verification; a fellow graduate or parent counts only once the graduate confirms. What the agents have found but cannot count yet is shown separately, as found but not yet counted.
The Conflict agent lays both sources side by side with the passages highlighted, and the Classifier proposes a category with the rule cited, for example under 30 hours a week is part time (§3) and the latest known outcome is reported (§5). It also takes a part-time salary out of the full-time salary figures. A person resolves it with a reason code, and a second person confirms.
Under NACE §4, the career outcomes rate counts graduates employed full or part time, in voluntary service, in the U.S. Armed Forces or continuing their education, over graduates with a known outcome, leaving out those not seeking. In the working solution, Halden State's class of 2026 reaches 92.1%, 5,883 of 6,389.
It prepares the evidence; the decisions stay with the institution. Certificate programs show placement the Workforce Pell way, the share of completers verified working 180 days after completion against 70%, and each program's outcomes and salary answers by CIP code go to Institutional Research for the earnings review, with small cells suppressed.
Yes, people stay in control. Agents propose and never confirm, send or sign. A college liaison approves every follow-up before it is sent, nobody who opted out is ever messaged, an analyst confirms every outcome, a second person confirms every change, the director signs the class report and Institutional Research checks the totals before submission.
The Agentic Solution Engine builds and deploys it from your requirements and documents: your class definition, your sources and how you verify them, your roles, your systems of record. It goes live once every quality gate has passed.
Every student at risk reached within two days, each warning sign cited to its record.
1-min demoAcademic Advising AssistantAnswers at any hour, cited to the student's own catalog year, and never a promise.
1-min demoAcademic Recovery PlansEvery probation student signs a plan with checked math before registration closes.
1-min demoWe'll run First-Destination Survey & Career Outcomes on a sample of your own class records and sources.