Home/How we verify
Editorial standards

How We Create and Verify Our Math Guides

A math guide with a wrong answer is worse than no guide at all. Here's exactly how ours are made, what gets checked, and what doesn't.

How guides are written

Our guides are drafted with the help of AI writing tools and then edited and published by the Aheadia team. We use AI to help structure explanations and draft prose — the same way many writers now do — but we don't publish anything on the strength of an AI model's say-so.

We pick topics where students genuinely get stuck, and we write for the specific question a student is asking — "when do I use u-substitution?" rather than "everything about integration" — because a focused answer is a more useful one.

How every answer is checked

Language models can produce mathematics that reads perfectly and is still wrong. So before any guide is published, every worked example and every practice-problem answer is independently checked with a computer algebra system (we use SymPy, an open-source symbolic math library). Specifically:

If the CAS and our worked solution disagree, the guide doesn't go up until we've found out why and fixed it.

What verification doesn't cover

We'd rather be clear about the limits than overclaim:

How the app checks its answers

The same principle runs through the Aheadia app itself: mathematical solutions are checked against independent math engines before they're shown to you, rather than relying on a single AI model's first attempt. Explanations, chat responses, and non-math subjects are generated by AI models and can occasionally be wrong — which is why our terms ask you to review important work yourself, and why we wrote a guide on how to check any math answer.

Found a mistake?

Tell us. Email hello@aheadia.app with the page and the problem, and we'll check it and correct the guide if we got it wrong. When we fix something substantive, we update the page's date — and we only update dates when the content has genuinely changed.