Questions
The questions we get asked
Several of these answers are no. They are here because the questions a vendor avoids are the ones worth asking.
- Questions
- 15
Can we actually buy this?
Not as a school licence that switches on student access — our code refuses org-level grants as ineffective, and there is no US school price in our billing system. What we would have to write for you is a school-bought per-student seat for everyone in the cohort, at the first US school price we have ever quoted, in writing before a pilot starts. If we cannot show you that invoice line, buy only the analytics or do not buy.
Who would we be contracting with?
Ask us on the first call for our contracting entity and registered address, a W-9, a certificate of insurance and our cyber-liability and indemnity position. Nobody outside our own team has verified those in writing yet, and we are not going to imply otherwise on a website. If we cannot produce them, stop — you would be a public school making an Indian company its first US customer, and you need to know who answers a subpoena.
How do 900 students actually get accounts?
This is the largest implementation risk and we have not solved it. The activation path we ship runs through an India-specific guardian payment or a manual school waiver, so a US charter cannot activate a roster as the product stands. Until we replace it, expect staff-loaded roster files, a manual waiver step, and an activation rate nobody can promise you — a bulk load runs as a preview by default, so an operator who forgets the write flag gets a report rather than hundreds of accounts.
What happens to a student without a paid seat?
After the 7-day trial they are denied the tutor, the mastery loop and document Q&A. Worse, today they cannot complete a study homework assignment their teacher set inside PrepGraph — the completion paths all require tutor activity the entitlement check blocks — and our app still shows them a Start button whose session our server refuses. It is a defect and it fails in front of a child. We will not run a pilot with mixed entitlement in the same class.
Is our students' work used to train AI models?
We cannot answer that provider by provider today, and we will not pretend the answer is no. We have not established no-training or zero-retention terms in writing with the providers on our subprocessor list. Making that answer contractual is a signing prerequisite alongside residency.
Does the tutor record our students?
Yes. Every tutoring session is a live voice session and we keep the audio and transcript for 90 days. That window is a setting on our side rather than a contract term, and recorded student voice is a biometric identifier in Illinois and Texas, which require written notice and consent before collection. We do not have that notice text yet, and no family should be recorded before it exists.
What happens if a child tells the tutor something serious?
We searched our own codebase for self-harm, crisis and safeguarding handling and found none. There is no defined escalation path to a named adult, no human review of sessions, and no teacher or parent transcript view. This is the first question we would expect from a principal and the honest answer is that it is unbuilt. If it is a condition of your adoption — and it reasonably is — it has to be built before a pilot, not during one.
Can a parent see or challenge what PrepGraph says about their child?
Not in our product. Once your staff maintain these bands and flags they are part of the education record you hold, carrying rights of inspection and of amendment that we cannot serve — there is no parent-facing view and no amendment path. Your staff would carry those requests by hand, exporting from our system, until we build one.
If we leave, do we get our data back or destroyed?
Today, neither cleanly. There is no export of a student's record, no contract-end bulk purge and no written retention schedule. Return and deletion on termination are clauses in a DPA we have not written yet, which is why we list that DPA as a prerequisite to signing rather than as follow-on work.
Do you work at the network or CMO level?
No. The top scope in our academic reporting is one school, a user belongs to exactly one school, and campus is not a dimension in any academic view. Central purchasing is fine; central reporting is not built. If a network wants a combined picture across two pilot schools we will assemble it by hand and label it as hand-assembled.
What will you need from our IT?
Devices with working microphones and a live audio connection per session. We have never run a session inside a US school network. Before a pilot we will give you the domains to allowlist and the bandwidth per session, and we would want to test thirty concurrent sessions in one of your labs before either of us counted on it. Assume nothing about managed Chromebooks until we have both watched it work.
Where does this sit on the ESSA evidence tiers?
Tier IV at best — a rationale. A pilot as designed produces a pre-post comparison with no control, which supports no causal claim and would not reach Tier III. If your authorizer or a funding stream needs Tier III or above, we cannot get you there and you should spend the money elsewhere.
Can we talk to a reference and inspect the content?
We work with schools and coaching centers in India and will introduce you to one — we do not name a school without its written permission, so give us a few days to ask. Before any meeting we will send fifty generated grade 7 and 8 items with their tagged distractors for your content leads to judge, and a recorded walkthrough of the real grid on a synthetic class showing the concept names your teachers would actually read. Given that about 8% of our US-linked items have passed a solvability check, you should look at the other 92%.
Why does your product site say something different?
Because it is wrong and it is being taken down. As of 1 September 2026 partners.prepgraph.com still carries four pilot-results figures that measure nothing, placeholder testimonials nobody said, a FERPA sentence nothing implements, a 14-day trial our code does not grant, a state-standards claim we cannot support, and an agent architecture that does not exist. Removing all of it is the condition for this page existing. If you are shown any of it by anyone at PrepGraph, it is wrong.
What should we ask you, and every other vendor?
The same six things about any number you are shown: sample size, comparison group, duration, who funded the work, who reviewed it, and the name of one school where it showed no result. Ask us too. Where we have not run a study we will say so, and you should write down that we did not provide one. Our own usage data is our own usage data and it is never an independent finding.