Designing an AI-powered Interview coach

for RAMPED Careers

transforming passive preparation into an interactive learning experience

00 – Project Introduction

An AI Platform to help you land your dream job

Ramped was a GenAI career platform that helped job seekers create resumes, write cover letters and apply for relevant opportunities. In 2025 was bought and absorbed into TEAL.

I worked on Help Me Get Hired, a feature designed to prepare users for interviews based on a job title and description.
The existing tool generated a preparation script with likely questions, interviewer expectations and recommended topics. While useful, research revealed that users treated it as something they needed to memorise rather than a tool that helped them practise.

01 – The challenge

Useful content did not necessarily create confident candidates

The original journey was simple:

During user interviews, I noticed that candidates were focused on remembering the “right” answers.
The tool explained what they should say, but it did not help them understand how effectively they were communicating it.

It feels like an exercise in memorising the right answers.

CLIENT

RAMPED Careers

DATE

September 2024

02 – Reframing the problem

How might we turn AI-generated interview advice into an active learning experience?

Users still needed quick access to useful information, but they also needed:

  • A safe environment to practise.
  • Guidance before answering.
  • Personalised feedback.
  • Opportunities to retry.

new hipothesys

A tailored preparation with a feedback loop: Like a tutor

03 – Solution

I redesigned the experience around two preparation modes

The goal was not to eliminate the existing preparation script. Some users still needed a fast, low-commitment way to understand what to expect. Instead, I proposed to test two different levels of interaction within the same product experience:

Option A

Interview Simulation

A guided mock interview where users answer questions, receive feedback and improve through a bespoke practice. This allowed candidates to choose the level of preparation that matched their time and needs.

Option B

Quick Guide

A structured overview for users who want to understand what to expect before an interview.

3A – Solution

Interview Simulation

A1. Review the interview plan

Before the session begins, Ramped presents an Interview Brief containing:

  • Interview duration.
  • Interview type.
  • Topics covered.
  • Number of questions.
  • Feedback expectations.

Users can accept the plan or adjust it using text or voice.
This makes the AI’s approach visible and gives the candidate control before starting.

A2. Answer with contextual guidance

  • Each question includes a short tip to help the user structure their response.
  • More detailed guidance is available through a What to cover element, allowing users to access support without overwhelming the interface.
  • Candidates can answer through text or voice, creating an experience closer to a real interview.

A3. Receive actionable feedback

After each answer, Ramped highlights:

  • Strengths: What the candidate communicated effectively.
  • What to strengthen: Where the answer needs more clarity, evidence or detail.
  • Coaching tip: A practical recommendation for the next attempt.

The feedback is immediately connected to three actions:

  • Retry the answer
  • View a suggested answer
  • Continue to the next question

This creates the main learning loop:

A4. Review the final report

At the end of the simulation, Ramped generates an Interview Report containing:

  • An overall performance score.
  • Communication and clarity ratings.
  • Strongest answers.
  • Areas to improve.
  • Personalised recommendations.
  • The complete interview transcript.

The final report turns individual feedback into a clear overview of the candidate’s performance

3B – Solution

Quick Guide

For users who do not need a complete simulation, the Quick Guide generates a structured preparation document containing:

  • A suggested interview agenda.
  • What the interviewer may evaluate.
  • Likely questions and answer guidance.
  • Questions to ask the interviewer.
  • A final preparation checklist.

Each section can be refined using Edit with AI, allowing candidates to personalise the generated content.

03C – Solution

From temporary output to saved progress

While designing the final report, I identified another problem:
The generated analysis remained useful after the session ended, but copying and pasting the content was not an effective way to preserve it.

Then I introduced two new options:

01   Download the report as a PDF.

02  Save the session inside Ramped.

The same functionality was then applied to Quick Guides.

This led to a broader redesign of the preparation homepage, introducing a Saved Sessions area where users could reopen, organise and delete previous simulations and guides

04 – OUTCOME

An active learning experience end interview preparation

The redesign changed the feature from a static content generator into a repeatable coaching experience.

Before

Generate → Read → Memorise

After

Practise → Receive feedback → Retry → Improve

The new experience:

  • Supports quick and in-depth preparation.
  • Makes AI-generated plans transparent and editable.
  • Provides guidance without answering for the user.
  • Encourages reflection and repetition.
  • Preserves previous preparation and progress.

Reflections

  • The strongest AI experience was not the one that generated the most content.
  • It was the one that helped users improve their own thinking.
  • By reframing the feature around learning rather than information generation, the AI became a coach instead of simply an answer provider.
  • The need to save a final report also revealed a wider product opportunity, transforming a single session into an ongoing preparation history.

From memorising the right answer
to learning how to build one.

Want to see more?

Check my Dribbble account or Get in Touch!