From summative to formative assessment: How ‘Listening Trivia’ uses AI to humanise soft skills assessment

In the corporate world, few rituals are as universally dreaded as the mandatory compliance training. It is often a passive, click-next-until-it’s-over exercise designed to generate a certificate rather than genuine behavioural change. But for Mustapha Salaudeen, a learning experience specialist and developer, the problem isn’t just that these tests are boring; it’s that they are fundamentally measuring the wrong thing.

“You’re only testing their knowledge retention level; you’re not forcing them or pushing them to self-reflect,” Salaudeen told me in an interview. His solution is Listening Trivia, a new AI-powered assessment tool that ditches the traditional 0-to-100 scoring model in favour of a psychological nudge toward self-improvement.

By combining the addictive mechanics of gamification with the analytical power of Large Language Models (LLMs), Salaudeen is proposing a radical shift: an assessment engine that is pure, unadulterated self-reflection. It doesn’t judge you but instead helps you understand how you listen.

The “TikTok-ification” of corporate learning

The architecture of Listening Trivia is built for an era of dwindling attention spans. Salaudeen, citing the TikTok-ification of learning, notes that the average attention span has cratered to roughly seven to nine seconds. To combat this, the tool is designed to be low-friction and high-impact, with an entire session lasting only five to seven minutes.

“I was just thinking around: how can we create a skills assessment that is not boring… but at the same time is not meant to just score people based on how well they can remember answers?” Salaudeen said.

The result is a streamlined, gamified interface where users interact with audio prompts.

But unlike a standard quiz, the system isn’t just checking if you heard the word “apple”. It maps user interactions against five distinct facets of listening behaviour: active listening, reflective listening, non-verbal awareness, summarising, and reaction to auditory signals.

What separates Listening Trivia from a standard BuzzFeed-style quiz is its backend intelligence. The platform uses a low-code architecture, combining Articulate Storyline for the frontend interaction with a Netlify serverless function to handle the logic.

Here is where the innovation lies: The AI never sees the raw user data or grades a test in the traditional sense. Instead, every interaction, whether a user clicks the right button or reacts to a sound promptly, is recorded as a Boolean (True/False) entity.

“The AI is not scoring you,” Salaudeen emphasised. “It’s merely interpreting the computer language and translating it into feedback that a human can react to.”

Salaudeen, whose advocacy for Africa’s edtech ecosystem extends beyond practice to actively contributing to global learning communities, including The Learning Guild, engineered the system’s prompt to adopt the persona of a “Friendly Assessor”.

He placed strict containment rules on the model to prevent it from hallucinating or becoming judgmental. Instead of a grade, the AI generates a qualitative behaviour map. It might tell a user: “You are excellent at summarising key details, but you consistently miss non-verbal cues in the tone of voice. Here is how you can practise that.”

This shift from summative assessment (grading) to formative assessment (coaching) is the tool’s standout feature. It leverages the psychological nudge theory, encouraging users to improve their behaviour rather than just their test-taking skills.

“I’ve put another constraint that says, ‘Keep the tone supportive, practical, and concise,’” Salaudeen noted. This ensures the feedback feels like mentorship rather than a reprimand, avoiding the defensive reaction that bad grades often trigger in learners.

Listening Trivia is solving the “certificate culture”

For Salaudeen, a scholar with an MA in Design, the tool also addresses a specific cultural challenge in the African corporate environment: the obsession with certificates over competence.

“The Nigerian corporate world is notorious for compliance-heavy training where employees just want the certificate,” the interviewer noted. Salaudeen agrees, arguing that the goal of edtech in Africa must shift from “compliance” to “culture change”.

By removing the pass/fail binary, Listening Trivia denies the user the easy dopamine hit of a “100% Score” or the shame of a “Fail”. Instead, they receive a qualitative map of their behaviour.

“Rather than just scoring you based on your performance, it is pushing you subtly to want to do better,” Salaudeen said. The data from early testing suggests this approach works; users are more likely to complete the assessment because the game feels fair and the feedback feels personal.

Designing with inter-market constraints in mind, Salaudeen architected Listening Trivia to be mobile-first and lightweight, acknowledging the digital divide and intermittent internet access that are common in emerging markets.

The platform eschews heavy graphics and video for simple, clean audio and text, ensuring it loads quickly even on low-bandwidth connections. The heavy lifting, the AI processing, is outsourced to the cloud via the serverless function, meaning the user’s device doesn’t need to be powerful.

“It shouldn’t be difficult to access,” Salaudeen said. “You can use it on a tablet, or on a mobile, or on a PC, and the experience is essentially the same thing.”

While the current iteration focuses on listening, the underlying technology is modular. Salaudeen envisions a future where this interaction-to-feedback framework is applied to other soft skills that are notoriously hard to grade, such as empathy, writing, or critical reading.

“The module is modular; it’s detachable,” he explained. “If I were going to re-adapt this to the Writing Skills Assessment, the modular framework still works.”

Mustapha Salaudeen’s Listening Trivia uses AI as a mirror that reflects our behaviours to us, kindly, accurately, and without judgment. It’s safe to say that the future of learning is no longer about getting the right answer but about asking the right questions about ourselves and learning to understand, not just passing.

The post From summative to formative assessment: How ‘Listening Trivia’ uses AI to humanise soft skills assessment first appeared on Technext.