AskBack — The two-minute check

You've read a paragraph that didn't sound like them. Maybe you ran it through one of the AI checkers — the ones that hand back a percentage. Maybe you just had the feeling and let it go.

Those checkers flagged over half of essays by non-native English speakers as machine-written. Every one was written by a human.

Seven of the leading tools, tested at Stanford. Liang, Yuksekgonul, Mao, Wu & Zou (2023), GPT detectors are biased against non-native English writers, Patterns.

The questions come from their paper. Not from a template.

Here's a real Grade 11 essay on the left, and what AskBack asked about it on the right. Notice where the questions land.

What the student handed in

"The Impact of Social Media on the Academic Performance of Filipino Senior High School Students" — extract One of the primary advantages of social media in an educational context is its capacity to facilitate collaborative learning. According to a study conducted by Reyes (2021), students who participated in online study groups demonstrated a fifteen percent improvement in their overall grades compared to those who studied independently. This suggests that when utilized responsibly, social media can serve as a powerful supplementary educational tool.

What AskBack asked

You cite Reyes for the fifteen percent figure — what were they actually measuring?
Your essay says study groups improved grades. How did the study separate that from students who were already doing well?
You called social media a "double-edged sword." Which edge did you find more convincing while writing this?
Where did the five-hours-a-day figure come from, and what did it count as usage?
You mention Messenger and Google Classroom specifically. Why those two and not the others?
If you had to cut one paragraph from this essay, which would go?

A student who did the work answers these in a sentence each. A student who didn't runs out of road on question one.

Two minutes, start to finish.

The whole check fits inside the gap between one activity and the next.

First

Paste, upload, or photograph

Paste the text, drop in a Word file or PDF, or snap a photo of handwritten work. Handwriting isn't a fallback here — it's how most work actually arrives.

Then

Ask, at their desk

Read the questions off your phone. Hit record. Mark each answer as they go: knew it, wobbled, stuck.

After

Keep the receipts

A transcript in the student's own words, editable, yours. Plus which points from their paper they actually covered.

What it does, and what it will never do.

AskBack does

Read the work and write questions only its author can answer well
Keep the transcript in the student's own words — Taglish included
Show which points from the paper the answer covered
Work on your phone, at the desk, mid-lesson
Leave every judgment call with you

AskBack never

Gives an AI percentage
Says who or what wrote something
Marks a student down for their accent or their grammar
Leaves a student alone in a room with an AI
Decides anything. You're the teacher.

Priced for a teacher, not a district.

Your own pocket. Your own decision. Cancel whenever.

Free

5 checks

every month, forever

Enough to try it on a class, show a colleague, and decide for yourself.

Teacher

₱99

a month, 30 checks

Unlimited at ₱179. Pay with GCash or Maya — no credit card needed.

About one milk tea a week.

Founding teacher

Free

6 months, then half price for good

First 100 only. All I ask is a 20-minute call once a month telling me what's broken.

Stop guessing. Start asking.

Open it and run a check on something a student handed in this week. Nothing to sign up for, nothing to install.

Run your first check free Ask about founding teacher

A stuck answer isn't a verdict. It's where the teaching starts.