Verbatim reads the everyday writing students already do in Google Docs and flags who might be showing signs of dyslexia, so schools can step in early, before a child slips through.
These children aren't hard to help. They're hard to find. The signs are right there in their everyday writing, if someone has the time to look.
Around one in ten New Zealanders is dyslexic, one of the most common learning differences.
SPELD NZ, up to 10% of peopleMany finish school and no one ever formally identifies them as dyslexic.
APPG for Dyslexia (UK), 2019Often just one learning support coordinator tracks need across an entire school.
A structural reality in most NZ schoolsYou are one person responsible for hundreds of students. Verbatim reads their writing alongside you, so catching who is struggling no longer waits on someone raising a hand.
Until a teacher flags a student, or a problem surfaces, you have no way to see who is quietly struggling. Verbatim gives you that view, from the everyday writing itself.
When a child slips through, it falls to you. Verbatim widens the net so fewer do, and makes the ones who need help easier to see, and to act on.
One dashboard, built on the writing students already do. See who's struggling, understand why, track whether support is working, and pull the evidence to act.
A whole-school view of who might be struggling, ranked by who to look at first, so every student is on your radar, not just the ones who speak up.
Open any student's profile to see the shape of their writing over time: the kinds of errors they make, and whether things are getting better or worse.
Mark when support starts, and watch whether a student's writing actually responds. Evidence that an intervention is helping or that it isn't.
Turn weeks of everyday writing into an objective record to stand behind a referral, an assessment request, or a case for resourcing.
The detection approach runs today, grounded in published research (including Luz Rello's pioneering work in Spanish) and shaped by New Zealand literacy researchers who study how struggling readers get identified. We haven't validated it on real dyslexic writing yet.
No English or New Zealand corpus of dyslexic writing exists to test against, so we're helping build one: gathering real student writing with consent and proper ethics oversight, then measuring how accurately Verbatim flags who's struggling. That's our main focus right now.
Verbatim won't call itself a screening tool until the evidence supports it. Until then it's an early signal to take a closer look, and a starting point for the people who can help.
Handling children's writing is a responsibility we take seriously. These are the principles we build it on, from the ground up.
Verbatim reads text to find patterns, then discards it. By design, it keeps only the structured signal, the type and frequency of errors, and never the sentences a student wrote.
Student-level detail stays with the school's own learning-support staff. The school owns its data and decides who uses it, not us.
Any use of children's writing for research or product improvement rests on clear, informed consent. We explain it to families in plain English, and they can opt out, before anything begins.
We're looking for a small number of New Zealand schools to help shape Verbatim and build the evidence base alongside us. This is a partnership, not a purchase.
A direct line to the founder. What your learning-support team actually needs drives what we build.
Founding schools are part of the work that turns an honest prototype into a validated tool, done properly, with consent and research oversight.
Early access to the dashboard and the screening features as the research supports them.
Tell us about your school and what your learning-support team needs. We take it from there, at whatever pace suits you.
We're partnering with a small number of schools to begin with. If yours could be one, put your hand up and we'll take it from there.
Takes about two minutesThat's the part we help with. Take a look for yourself, and help us bring Verbatim to your school.