How CaseCoach turned AI-powered case interview preparation into a sales differentiator with top MBA programmes

Who CaseCoach is

CaseCoach prepares candidates for case interviews, the format that McKinsey, BCG, Bain, most management consulting firms, and a growing number of employers in other industries use to hire. In a case interview, you get a business problem and have to structure your solution, run the numbers, demonstrate sound judgement and creativity, and defend your recommendation out loud in real time while being evaluated by the interviewer.

Most of CaseCoach's users are MBA and university students prepping for consulting recruiting, often through a school partnership. Its client universities include names like Harvard Business School, MIT Sloan, London Business School, and HEC Paris.

What we built together

CaseCoach's “classic” self-study product runs on video lessons, interactive text-based exercises, and downloadable PDF exercises: read a case, write an answer, check it against a key. While this is highly effective content, based on mock cases and high-quality material prepared by former recruiters from top consulting firms, there are aspects of case interview preparation that are difficult to replicate through self-study alone. One-way communication and students checking their own answers does not capture the experience of an intense interview session in the most realistic manner.

Over the past two years, our team at Trivial Group has joined forces with CaseCoach to build a different kind of exercise: AI-powered practice drills covering the core skills tested in a case interview (math, chart reading, market sizing, structuring, and creative problem-solving). Instead of filling in a worksheet, a candidate talks through a live problem with an AI that follows up on their answers and gives graded, written feedback at the end. It's closer to what an actual interview feels like than anything a text-based exercise can offer.

Each drill type runs its own feedback logic, tuned to CaseCoach's specific methodology, judging their communication, delivery, and final results. Their founder is a former McKinsey interviewer and headhunter for McKinsey, BCG and Bain, so the bar for what counts as a good answer is exact, based on decades of consulting recruitment experience, not merely generic, general feedback. The challenge is that case interview performance cannot be evaluated through generic feedback or answer matching alone: multiple approaches can be correct, and meaningful coaching requires case-specific prompts and expert-informed criteria to assess structure, logic, and problem-solving.

Results

In CaseCoach students' own words:

“It helped me refine my structuring without feeling like I was practicing alone.”

“Much more realistic to talk through a framework than to type it all out.”

The effect showed up in sales numbers too. CaseCoach sells university access on a per-student basis, competing directly against other prep platforms in the same evaluations. More than once, the AI drills were the reason a school picked CaseCoach over an option that looked similar on paper, but lacked the same depth and quality of the AI simulation & feedback.

How we work together

Our team has been working closely with CaseCoach for nearly two years now. In a well-aligned product development partnership, CaseCoach’s expert team owns the content and consulting methodology, while Trivial Group owns the AI research, engineering, and evaluation behind the drills. Together, we have created a product that users love & which helps them achieve their recruitment goals, all while continually keeping up with new possibilities unlocked by ever-evolving large language and speech-processing frontier models, and upgrading the product as we go. By the time a drill leaves R&D, it’s seamlessly integrated into CaseCoach's existing platform without disrupting anything students or the CaseCoach team already rely on.

Getting there was not easy, however. While it’s easy enough to create a voice agent system prompt and ask any AI chatbot to “judge how I did”, this basic feedback is often generic at best, or misleading, all-too-known “AI slop” at worst. One of the key factors enabling the success in our collaboration has been the iteration process and testing system built alongside the drills. Every case has its own behavioral test suite, so when the underlying models, prompts, or content change, regressions get caught before a real candidate sees them. This is constantly being updated with feedback, anonymized production data from usage, and direction from content experts at CaseCoach.

Trivial Group feels like an extension of our team. We work together closely every day, and they bring a level of AI expertise that we simply wouldn’t have access to on our own. Their support has helped us move faster, build better products, and create a real competitive advantage by offering our customers features they can’t find elsewhere.

Enguerran Loos, Founder, CaseCoach