Guides
A practical data-boundary guide for responding to first-party enrollment inquiries without putting FERPA-regulated education records into the workflow.
Last updated
Enrollment teams have a deceptively simple automation question: when can an AI agent follow up with a prospective student, and which data should never enter the conversation workflowWorkflowAn automated, multi-step process — usually triggered by an event (form fill, new lead) and orchestrating one or more voice / SMS / email actions.?
The useful answer is not "FERPA applies to every prospect record" or "FERPA starts on the first day of class, so everything before that is fair game." Both shortcuts are too blunt. The legal status of the record, the institution's definition of attendance, the source of the data, state law, channel consent, and the provider's own contract all matter.
Thoughtly's current public Terms of Service say customers must not provide Customer Data subject to heightened security requirements including the Family Educational Rights and Privacy Act (FERPA). That is the operating boundary for this guide: use Thoughtly for approved first-party inquiry response that does not include FERPA-regulated data, and stop before protected education records enter the workflow.
This guide is operational, not legal advice. FERPA, state student-privacy rules, calling rules, institutional policy, and contract terms can apply differently to each school and workflow. Enrollment, privacy, security, and counsel should approve the actual data map, scripts, triggers, disclosures, and handoffs.
For AI enrollment follow-up, separate the conversion job from the protected-records job. An agent can respond to a known inquiry, provide approved public program information, capture limited preferences, and book or warm-transfer an advisor conversation when the institution has confirmed that the payload is not FERPA-regulated and the contact is permitted.
The agent should not receive, retrieve, infer, or disclose protected education records. Under Thoughtly's current public terms, that includes keeping FERPA-regulated customer data out of Thoughtly entirely.
FERPA applies to educational agencies and institutions that receive funds under programs administered by the U.S. Department of Education. The current FERPA regulations define a student as an individual who is or has been in attendance and about whom the institution maintains education records. An education record is directly related to that student and maintained by the institution or a party acting for it.
For college applicants, the Department's guidance on prospective-student records says applications from people who are not attending are generally not education records at the prospective institution. The same guidance says state law governs the confidentiality of those application materials. It also warns that records such as transcripts received from another school must be protected and may not be redisclosed except under FERPA's rules.
The Department has historically left each postsecondary institution to determine when a person is "in attendance," provided the definition has a reasonable factual basis and is applied consistently. The guidance says that point should be no later than the first day of class. A CRMCRMThe system of record for leads, contacts, deals, and activity. Thoughtly reads from and writes to your CRM continuously. label such as applicant, admitted, deposited, or matriculated is therefore not a universal legal answer.
The practical control is an institution-owned lifecycle classification. Do not ask the agent or a language model to decide whether FERPA applies from conversational clues. The system of recordSystem of recordThe authoritative system where customer, lead, policy, loan, appointment, or account data is stored and updated. should supply a reviewed status such as pre-enrollment/non-FERPA, legal-review, or FERPA-blocked.
This conservative map is for postsecondary enrollment teams. It does not replace institution-specific legal review, and it does not authorize FERPA data in Thoughtly.
| Enrollment moment | Likely FERPA status at the receiving institution | Appropriate AI enrollment job | Required gate |
|---|---|---|---|
| New first-party program inquiry | Generally not an education record when the person has not been in attendance | Provide public program information, capture limited preferences, book an advisor | State privacy, institutional policy, source provenance, and channel permission |
| Application submitted, applicant not in attendance | Generally not an education record at the prospective institution; state law may still protect it | Explain public next steps and route questions without revealing individualized status | Institution-approved applicant-data policy and minimum-data payload |
| Transcript or record received from another school | Protected from redisclosure under the Department's applicant guidance | Keep the record and its contents outside the agent; route to an approved channel | FERPA handling and redisclosure control |
| Admitted, deposited, or pre-start applicant | Depends on the institution's reasonable, consistently applied definition of attendance | Use only approved non-FERPA inquiry context; escalate uncertainty | Deterministic legal classification before any data is loaded |
| Current student in attendance | Directly related records maintained by or for the institution can be education records | Do not pass FERPA-regulated customer data into Thoughtly under current public terms | Block the workflow or move it to an institution-approved protected-record process |
| Former student | Records directly related to prior attendance can remain education records | Keep protected history outside the agent | Record-level classification and approved human handling |
The difficult row is the transition from applicant to student. Treat ambiguity as a stop condition. A slower human review is cheaper than letting a generic status label quietly move protected data into the wrong system.
A strong enrollment workflow begins with a person who raised a hand: a program-information form, a requested callback, a campus-visit request, or another first-party event. That gives the conversation a clear purpose and relevant context. It is a different commercial motion from cold calling a purchased list.
Inbound origin does not, by itself, establish permission for every channel or every later touch. FERPA governs education-record privacy; the TCPATCPAUS federal law governing telemarketing calls and SMS. Thoughtly enforces consent capture, time-of-day windows, and DNC scrubbing automatically., state telemarketing rules, recording-consent laws, carrierCarrierA telecommunications provider that routes phone calls and SMS over its network. Twilio, Telnyx, and Bandwidth are the three most common in the AI voice space. requirements, and the institution's own promises govern other parts of the communication.
The FCC has confirmed that AI-generated voices fall within the TCPA's restrictions on artificial or prerecorded voice calls. The FCC's Declaratory Ruling 24-17 does not create an "inbound lead" shortcut. Before an AI call, the institution still needs the applicable consent basis, disclosures, calling window, opt-outOpt-outA recipient’s request to stop receiving calls or messages. Compliant systems must capture opt-outs and suppress future outreach where required. behavior, and any required suppression checks.
Make contact eligibilityEligibilityThe fit criteria that determine whether a prospect can move forward, such as service area, insurance coverage, loan type, location, age, or program requirements. the first workflow gate. Preserve only the approved evidence needed for that decision:
Relevance does not require the full applicant or student file. The safest useful payload is deliberately boring: enough to recognize the inquiry and complete the next step, with protected records left in the institution's approved systems.
Thoughtly's automation triggers can start from forms, CRM events, incoming webhooks, schedules, and call events. Conditions should evaluate the institution's precomputed eligibility fields before a call, text, email, lookup, or write-back step runs.
Thoughtly's documentation distinguishes temporary call metadata from persistent contact attributes. That distinction is useful for minimization, but it is not a FERPA permission. Do not place protected data in either location under the current public terms.
A field allowlist is more reliable than a field blocklist. Define exactly which values the automation may read, which values it may write, and which state transitions force a human handoffHuman handoffThe moment an AI agent transfers context, call details, and the next step to a human rep, licensed specialist, or support team..
FERPA's definition of a record includes information recorded in any way, including computer media and audio. A recording or transcriptTranscriptThe text record of a voice conversation, used for review, training, compliance audit, and search. is not automatically an education record simply because it exists. But once it is directly related to a student and maintained by the institution or a party acting for it, it can fall inside the definition. See 34 CFR Part 99 definitions.
That makes "we only discussed it on a call" a poor control. The workflow can generate a recording, transcript, summary, extracted variables, CRM note, disposition, and follow-up task from one conversation. Classify each artifact, minimize what is created, and decide where it may be stored before launch.
Thoughtly's security page describes SOC 2 Type II, role-based access, recording, and audit logging. Those controls support a general security review; they do not override the Excluded Data clause in the current Terms or turn a prohibited payload into an approved one.
The FERPA regulations require reasonable methods to identify and authenticate parents, students, school officials, and other parties before disclosing personally identifiable information from education records. A recognized phone number, caller ID, email address, or confident voice is not automatically sufficient identity proof.
A pre-enrollment agent can answer public questions without opening a personal record. When a person asks about an application decision, transcript, financial-aid package, registration hold, academic status, or another individualized record, the agent should route them to the institution's approved authentication and service path.
Do not disclose to a parent, spouse, employer, or other third party merely because the caller knows the student's name and program. At a postsecondary institution, FERPA rights transfer to the student when the student attends, even if the student is under 18.
The handoff should carry only the non-sensitive reason for transfer. The authorized human or system can authenticate the person and retrieve the protected record inside the institution's governed environment.
FERPA permits some contractors and other third parties to be treated as school officials when specific conditions are met. It is not a generic vendor badge, and a security certification alone does not create the exception.
The Department's school-official guidance and the regulations require the outside party to meet the institution's criteria and operate under the institution's control. Before any provider handles education records, the institution should establish that the provider:
The Department also publishes responsibilities for third-party service providers. Institutions should pair that guidance with current contracts, state law, security review, retention rules, incident response, and procurement requirements.
This checklist does not change Thoughtly's current public contract boundary. The Terms say not to provide FERPA-regulated Customer Data. For Thoughtly, keep the deployment on the non-FERPA side of the line unless and until the applicable written terms say otherwise.
Create an institution-approved field such as data_scope with allowed values like pre_enrollment_non_ferpa, review_required, and ferpa_blocked. Set it from authoritative enrollment state and policy, not from model inference.
Use a form submission, requested callback, campus-visit request, or reviewed CRM state. Pass the source, time, purpose, brand, program family, contact permissions, and data_scope. Do not import a cold list and then describe the calls as enrollment follow-up.
If source provenance, channel permission, suppression status, or data_scope is missing or ambiguous, stop. Do not let a default branch place the call or query a broad CRM or SISSISA student information system — the system of record for education programs. Slate and Ellucian are common; Thoughtly integrates with both. object.
Ground the conversation in public program facts, published deadlines, event inventory, advisor availability, and the person's stated pre-enrollment preferences. Use a field allowlist and a narrowly scoped knowledge source.
Confirm what the person requested, which broad program they want, whether they prefer an advisor conversation, and when they are available. Do not probe for grades, test scores, aid eligibility, health information, disability, immigration status, or other sensitive facts simply because the model can ask.
Transfer or schedule when the conversation reaches individualized application status, admissions judgment, transcript review, financial-aid advice, registration, academic records, complaints, identity uncertainty, or any data-boundary question.
Thoughtly variables can capture structured values for branching and follow-up. Restrict write-back to approved non-FERPA facts such as reached, advisor requested, visit requested, preferred callback window, transferred, booked, opt-out, or human review required. Do not summarize protected details into a supposedly harmless note.
An opt-out, wrong-party report, complaint, identity mismatch, or privacy concern should end the automated path and update the institution's suppression or review state before another voice, SMS, or email step can run.
Thoughtly is built for higher-education inbound lead conversionInbound lead conversionThe process of turning opted-in inquiries, form fills, calls, and quote requests into qualified conversations, appointments, or transfers.: responding to enrollment inquiries, providing approved information, booking advisor conversations, and handing warm prospects to the human team. The job is to cover first-party demand the institution already has, not to cold-call strangers.
Inside the reviewed boundary, a CRM or form event can triggerTriggerThe event or condition that starts an automated workflow, such as a new lead, missed call, CRM status change, calendar booking, or completed call. the workflow, conditions can reject ineligible records, the agent can contact the prospective student with minimal context, and the outcome can be written back for the enrollment team. The institution remains the source of truth for eligibility, privacy classification, consent, and the next permitted action.
For adjacent workflow mechanics, see Thoughtly's guides to education enrollment follow-up, campus visit booking, and enrollment no-show recoveryNo-show recoveryThe process of re-engaging leads or patients who missed a scheduled appointment, using automated voice, SMS, or email outreach to reschedule and recover revenue that would otherwise be lost.. Apply the data boundary in this article before using those patterns.
Thoughtly should not become a shadow SIS or a route around the institution's protected-record process. Fast response is valuable because it gets the right inquiry to the right human sooner, not because it makes every student record available to an automated conversation.
Revenue and enrollment teams still need to measure the full path. The clean design joins conversation events to downstream outcomes inside the institution's governed CRM, SIS, warehouse, or analytics environment. Thoughtly does not need the protected record to prove that first-party response contributes to enrollment.
Starts and yield are the final enrollment conversion units. Contact attempts and call minutes are inputs. The measurement plan should preserve the original inbound source while keeping protected details in the systems approved to hold them.
Generally, application records for people who have not been in attendance are not education records at the prospective postsecondary institution. State law and institutional policy may still protect the information. Records received from another school, such as transcripts, carry FERPA redisclosure protections, and the institution must define attendance reasonably and consistently.
Under Thoughtly's current public Terms of Service, customers must not provide Customer Data subject to FERPA. Use Thoughtly only for an institution-approved pre-enrollment workflow whose payload has been confirmed not to include FERPA-regulated data, and route protected-record activity elsewhere.
No. It establishes useful provenance and a relevant enrollment purpose, but the institution still needs the applicable permission for the channel and call type, plus required disclosures, calling windows, recording rules, opt-out handling, and suppression.
Only if the institution has determined the data is outside FERPA, allowed by other law and policy, permitted by the provider contract, and safe to disclose through an approved authentication method. Under Thoughtly's current public Terms, a FERPA-protected status should not be supplied to the platform. The conservative default is to route the person to an approved authenticated channel.
Do not assume so. At the postsecondary level, FERPA rights transfer to the student when the student attends, regardless of age. Personalized record disclosures should move to the institution's approved authentication and authorization process.
On the non-FERPA side of the boundary, automate response coverage, approved public information, limited preference capture, advisor or campus-visit scheduling, warm transferWarm transferA live transfer where the agent connects a qualified caller to the right human while preserving context, instead of sending the caller to a cold queue or voicemail., and factual non-sensitive disposition. Keep education records, admissions judgment, individualized financial aid, academic matters, and exceptions with authorized people and systems.
Student starts and yield are the final outcomes. Completed advisor conversations, applications, deposits, and campus visits are useful intermediate measures. Join those outcomes to the original inbound cohort inside the institution's governed analytics environment.
Respond quickly to the person who raised a hand, but do not let speed erase the data boundary. Let the institution classify the record, let the workflow use only approved non-FERPA context, and let an authorized human or system take over before protected education records appear.
That is the practical trust advantage of Thoughtly for education enrollment: more complete coverage of first-party inquiries without pretending that a lead-conversion agent should become the student-record system.