.jpg)
“Hi Ben, we are pioneering the use of natural languageprocessing to auto-mark answers to open-ended questions for exam revision andassessments. Went live globally with a pilot study for a top name inassessments last week;”
While this is the kind of inbound we seem to get weekly now, it was a rarer message to receive on LinkedIn in August of 2021. At the time, the only way to incorporate AI into your platform was to build it yourself. The founders of Sainaptic, Raj and Kavitha studied at Cambridge over a decade before. They then built careers deep in the practices of machine learning and education respectively and subsequently came together to combine their expertise and apply it to the problem of how to give learners better feedback on exams.
This turns out to be a particularly tricky problem to navigate, and it took a while to develop the product in a way that it could be applied easily across multiple exams. Because Sainaptic began at a time when LLMs were not broadly available, and because in high-stakes exams precision is paramount, they took the long way around, building their own models that could then be fit for purpose for any examination or training body, to written and visual feedback in safe and reliable formats.
Building their own models, it turns out, presents Sainaptic with unique advantages when selling to examination bodies. These bodies are understandably eager to preserve their core testing assets, by, among other things, ensuring their data is not used to train generally available platforms. Moreover, Sainaptic's models a re both precise and significantly cheaper than frontier models and thus can be used to provide feedback to learners economically and precisely. Today Sainaptic have gained the trust of many of the largest examination bodies in the UK, enabling these organisations to adopt AI across a diverse array of fields: from accounting to physical training to building trades.
It would be disingenuous to say that we knew how things would play out when we initially invested. At the time, what we saw were two talented founders who had a technical idea that was differentiated and a credible path towards building it. In 2021, AI was too nascent to be more definite than that.
Supporting Raj, Kavitha, Bram and the team as they pivoted from an initial focus on schools/GCSEs through to finding product market fit as core infrastructure in the UK’s vocational education sector has been a real pleasure. They have built the best platform for examination and training bodies to become AI Native on their own terms. And it still feels like they are just getting started.
We are thrilled to continue backing them alongside new investors and eager to see what the future brings as they help redefine vocational education in the UK and beyond in coming years.
