Talk is cheap
It is easy to build a product and claim it works. It is harder to stake your own teaching outcomes on it.
That is what we do. We run Ren Digital Academy, a digital centre that teaches A Level H2 General Paper, entirely on our own platform. Ren marks every essay. Every piece of feedback a student receives passes through the same system we sell to schools.
This is not a demo environment. The students are real, the exams are real, and the stakes are real.
Why we use our own product
The software industry has a term for this practice: dogfooding. You use your own product internally before you ask anyone else to use it. The logic is simple. If you will not use it yourself, nobody else should.
For us, this practice is the foundation of how we build.
A test ground for new ideas
We often form a hypothesis about a feature. It can be a different way to present feedback, a new tag model, or a change to how Ren applies an answer scheme. We do not wait for a partner school's next assessment cycle. We test the idea on the next batch of essays in the academy, and we get a signal in days instead of months.
A proof you cannot fake
We want a product that gives genuinely personalised, high-quality feedback. We must therefore show results.
Because we run the academy, we hold the data directly: student performance across the term, feedback quality, and tutor efficiency. We do not ask a third party to vouch for us. We show the outcomes ourselves.
Fast iteration, honest feedback
Our tutors are also our most demanding users. They use the product every day, and they tell us exactly where it falls short. No delay separates a problem from the team that must correct it.
What we see
Two effects are now clear as the digital centre grows.
Tutors save substantial time. The AI produces the first pass. A tutor then refines and personalises it instead of starting from an empty page. The mechanical work shrinks and the high-value work remains.
Feedback quality rises. This is the counterintuitive part. Most people expect AI-assisted feedback to be a compromise that is merely good enough. The opposite happens. A tutor who starts from a detailed draft produces final feedback that is more thorough and more specific than the feedback they write from scratch under the same time limit.
What we have not solved
We will not pretend that everything works. The interface changes constantly, because the best workflow is not yet obvious. How should a tutor review AI marks? How much detail belongs in a feedback draft? When must the system ask for a human check, and when can it proceed alone?
We can only answer these questions if we use the product ourselves, observe what works, and iterate quickly.
Our mission stays the same
We build Ren to help teachers, and never at the cost of the education a student receives. We want the opposite result. The feedback a student receives must become measurably better than the current standard.
A digital academy that runs on our own product keeps us honest about that mission. If Ren is not good enough for our students, it is not good enough for yours.
Do you want to see the product we use ourselves? Get in touch to book a demo.