Working as a freelance EQA across multiple awarding bodies, I regularly act as the final checkpoint in the quality assurance chain. By the time student work reaches me, it has already been reviewed by tutors and IQAs. Yet increasingly, I began noticing signs that content was being generated by tools such as ChatGPT and Co-Pilot—often with no indication that these signs had been picked up earlier.

As someone who adopted AI early on, recognising the traits of machine-generated text has become second nature. But what became immediately clear was that this isn’t the case for most people I work with. In fact, a significant proportion—around 68% of FE professionals I support—have yet to engage seriously with AI, often due to apprehension, a lack of guidance, or the absence of institutional frameworks. Meanwhile, students are using these tools with ease and little oversight, operating in a policy vacuum where the rules simply haven’t caught up.

The Ai Detective came out of that gap. It started not as a business idea, but as a response to a real and growing problem I was encountering in my everyday work. From the outset, my intention was simple: to support the FE sector in adapting to a rapidly changing landscape, by offering policy insight, staff development, and strategic planning that keeps academic integrity intact while acknowledging the realities of modern learning. AI isn’t a passing trend—it’s here to stay. The real challenge is not how we stop it, but how we work with it in a way that makes sense for both educators and learners.

At first, I delivered consultancy and training sessions, helping staff identify what AI-generated work looks like and how to respond constructively. But the tools on the market at the time were letting us down. Many were inconsistent, prone to false positives, and raised serious equity issues—particularly for learners with additional needs or those who speak English as a second language. As a parent to three dyslexic children, that resonated with me on a personal level.

That’s what pushed me to build something better. I trained a custom GPT using the criteria I apply as a human assessor—patterns I’d come to recognise through years of professional experience—and built a system that could generate a clear, reasoned report, offering a likelihood scale and suggested next steps. It wasn’t a binary judgment. It was a support mechanism. And it worked. I began using it in my own EQA work, and it soon became obvious that the concept had wider potential.

That’s when Tom at Hoot Digital introduced me to Amaan and Shakeel at Sprouto, a Manchester-based software development firm. With their support, that original idea has grown into a fully functioning tool designed specifically for FE institutions. We're now testing it with a group of early adopters as we prepare for a full market launch later this year.

What makes our software different is its human-first approach. Unlike traditional tools such, it doesn’t deliver a blanket binary verdict. It replicates how a trained professional might interpret AI use and produces a reasoned, printable report that supports tutors in applying their own professional judgment. It’s designed to complement—not replace—existing IQA processes and policies.

This is not about catching students out. It’s about creating a culture of transparency, responsibility, and fairness around AI use. With clear standards and guidance, we can reduce malpractice, support audit readiness, enhance teaching, and shape more robust institutional policy. That’s the real opportunity.

None of this would have been possible without the support of my husband Will, my parents Pam and Jim, Tom at Hoot, and the brilliant team at Sprouto. As we move into the next stage, we’re focused on finalising our pilot phase and launching as a subscription-based service. We're also eager to collaborate with institutions, awarding bodies, and sector leaders who recognise the need for meaningful change.

This is just the start.