We started Graide in 2019 because marking at scale was genuinely broken. I had seen it first-hand. Lecturers drowning in scripts, spending weekends on feedback for students. Students waited weeks for those comments in order to improve. The process was slow, inconsistent, and exhausting for everyone involved. We thought we could fix it.
The problem was volume. When you have a hundred students submitting the same assignment, writing individualised, high-quality feedback for each one is almost impossible to sustain. Something gives, usually it is the turnaround time.
So we built something to help with that. We built a classification AI system, trained specifically on each assessment using data from that institution's own markers. The model learns what a particular marker, at a particular institution, considers the appropriate feedback is for response. Then it applies that consistently across every submission.
Every suggestion still has to be accepted or rejected by an educator. Nothing is graded automatically. An educator can override any suggested grade at any point. A moderation workflow lets a second reviewer check work before it is released. That was a deliberate design choice from the beginning, and it remains the core of how Graide works. The AI makes suggestions. The educator decides.
The results, when institutions gave it a proper run, were significant. Marking time reduced by up to 89%. Feedback per student increased by up to seven times. Those numbers matter, but what I am most proud of is that the feedback quality went up alongside the efficiency. That is the combination that is hard to get right.
Today, I am sharing that Graide has been acquired by Inspera.
Inspera runs end-to-end digital assessment for universities and awarding bodies across more than 160 countries. Question authoring, secure delivery, marking, feedback — they cover the full assessment cycle. More than 20 million submissions go through their platform every year.
When I looked at what they had built and the institutions they work with, it was clear this was the right place for what we have spent seven years building.
What drew me to them, beyond the scale, was how they think about AI in assessment. Michael Ioakimides, their Executive Chairman, put it directly: educators need to remain accountable for every grade, and AI should make human-crafted feedback more efficient rather than replace it. That is not a marketing position for them, it is built into how they approach the product. It is also exactly what Graide was designed around.
Not all AI-assisted assessment is the same. Some tools have a large language model that produces the feedback, and a human clicks to confirm it. That is not a human in the loop. That is a human endorsing an output they did not write, cannot fully explain, and may not be able to defend if a student challenges it. Accountability matters in assessment. It matters legally, it matters pedagogically, and it matters to students who deserve to know that someone actually read their work.
Graide works differently. Because the model is built from the institution's own marked data, the suggestions it makes trace back to real marking decisions by real educators at that institution. The AI's decisions connect to specific features in the student's response, which means they can be explained and challenged. That combination of efficiency and accountability is what made the Inspera conversation straightforward.
For our current customers, the immediate picture is straightforward. You can continue using Graide as a standalone tool, and you will have full support through the transition.
If you are already using Inspera, Graide's marking and feedback capabilities will become part of the platform you work in day to day.
For institutions not yet using either product, this is a good moment to look at what the combined platform offers. End-to-end assessment, from question authoring through to marked, moderated, and released feedback, with AI that keeps educators in control throughout.
To our customers: you shaped this product. The institutions we worked with pushed us to get things right, pointed out where we fell short, and trusted us with something that genuinely matters to their students. I am grateful for that.
To the Graide team: what you built is something to be proud of. The decisions we made about how to build it, not just what to build, are the reason this acquisition happened.
The next chapter starts now, and I think it is a good one.
.png)
