Getting started
Cadence is an AI-powered interview preparation app. You add a job posting, practise answering tailored questions by voice, and get specific feedback on what you said and how you said it — covering your answer structure, delivery, and communication patterns.
Each session ends with a scored report you can reference again. Over time, a progress view tracks how you're improving across sessions.
All you need is a job posting URL. Paste it in and Cadence will extract the role details and generate tailored questions straight away.
To get the most out of Cadence, we recommend uploading your resume or connecting your LinkedIn. When we know your background, the questions become more specifically tailored to you — and the feedback can reference your actual experience rather than giving generic advice.
Tap Add role from the Jobs page and paste in the URL of the job posting. Cadence fetches the page, extracts the role details — title, company, requirements, responsibilities — and uses this to generate your practice questions.
Most job boards work out of the box. If a posting sits behind a login (some internal ATS portals), Cadence may not be able to scrape it fully — in that case, paste the job description text directly.
You can practise as many times as you like for the same role — repetition is the point. Each session generates a fresh set of questions and its own scored report, so you can track improvement across multiple attempts for the same job.
How questions are generated
How the AI works
Questions are built specifically for you — not pulled from a bank.
Every question set is generated fresh. Here's exactly what the AI considers:
The job posting
The AI reads the full role description: responsibilities, required skills, seniority level, and company context. A VP role at a startup gets different questions to a Manager role at an enterprise.
Your background
If you've uploaded a resume or connected LinkedIn, the AI maps your experience against the role requirements to identify where to probe further and where you're already clearly qualified.
Question category balance
Each set includes a deliberate mix of behavioural, leadership, role-specific, and culture questions. The weighting shifts based on seniority: senior roles get heavier on leadership and strategy; earlier roles on technical and execution.
Gap analysis
The AI specifically targets areas where your background and the job requirements don't obviously overlap. These are the questions a real interviewer is most likely to ask.
A standard session runs 8 questions. You can review them before starting and adjust session settings if needed. Questions are ordered thoughtfully — typically opening with a broad behavioural question and moving into more specific role and leadership territory.
Questions are tagged by category:
Behavioural — "Tell me about a time you..." situations that test how you've handled real scenarios
Leadership — how you develop teams, build trust, manage conflict, and drive results through others
Role-specific — targeted to the skills and domain in the job description
Culture — how you work, what you value, and how you fit the company's way of operating
Intro — "Tell me about yourself" and career narrative questions
Each session includes a coach tip for each question — a short, specific prompt to help you frame the strongest answer before you speak.
You'll get a fresh set each time. The AI uses the same inputs — job description and your background — but generates new questions on each session. This is intentional: real interviews vary, and practising with varied questions builds more adaptable answers than memorising the same set.
The questions will still be tailored to the role — but only from the job description side, not personalised to your specific background. You'll get strong, relevant questions for the role type. The personalisation layer (gap analysis, tailored follow-ups) activates once your background is connected.
How interview answers are scored
How the AI works
Every answer is scored against the STAR framework — and then some.
After you speak, your answer goes through three layers of analysis:
Transcription
Your voice is converted to text via speech-to-text. The transcript is the foundation for everything else.
STAR scoring
The AI evaluates four components: Situation (did you set the context clearly?), Task (was your specific responsibility clear?), Action (did you describe what you specifically did?), Result (did you quantify or clearly state the outcome?). Each component is scored 0–100.
Filler word analysis
Your transcript is scanned for filler words and verbal habits. Primary fillers (um, uh, er, ah) are counted every time. Secondary fillers (like, actually, basically) are only flagged when they appear more than twice in an answer — because single uses are often legitimate.
STAR stands for Situation, Task, Action, Result — the most widely used framework for answering behavioural interview questions.
A strong STAR answer: briefly sets up the context (Situation), explains what you were specifically responsible for (Task), describes what you personally did (Action), and lands with a clear or quantified outcome (Result).
The Result component is where most people score lowest — saying 'it went well' is much weaker than 'we increased conversion by 22%.' Cadence flags this specifically.
Interpersonal Intelligence is Cadence's measure of communication patterns that reflect emotional intelligence — based on how you talk in your answers, not a fixed personality test.
The AI looks for five observable signals in your transcripts:
Self-Awareness — do you own outcomes, or attribute them elsewhere?
Empathy — do you reference others' perspectives and feelings?
Composure — how do you describe pressure or conflict?
Influence — do you show how you moved people, not just what you did?
Growth Mindset — do you frame setbacks as learning?
These are patterns in how you communicate — all trainable. Your scores improve as you practise.
Note: these scores are communication signals, not a clinical assessment.
Filler words are split into two tiers:
Primary fillers — counted every time they appear: um, uh, er, ah, hmm
Secondary fillers — only counted when they appear more than twice in a single answer, because a single use can be legitimate: like, you know, sort of, kind of, basically, literally, actually, I mean, right, okay so
Your filler count appears per-question in the full report, and as a session total with a trend showing whether you're improving across sessions.
Scores are 0–100 and calibrated to interview standards at the level of the role you're practising for:
90–100: Outstanding — an answer that would genuinely stand out in an interview
80–89: Strong — clear, well-structured, and convincing
70–79: Solid — good foundation, some areas to sharpen
60–69: Developing — notable gaps to address
Below 60: Needs work — prioritise the specific feedback
A score of 75 for a VP-level interview means something different to 75 for a graduate role. The AI calibrates to the seniority and type of role you're practising for.
Pitch Practice
How pitch analysis works
A pitch is scored differently to an interview answer.
Instead of STAR, Cadence evaluates your elevator pitch across six dimensions designed specifically for professional introductions:
Hook
Did your opening line grab attention, or start with hedging and filler?
Structure
Does your pitch follow a Present → Past → Future arc, or is it a chronological list?
Relevance
Did you lead with what matters most? If your resume is connected, the AI checks whether you used your strongest credentials.
Concision
Was your pitch tight and purposeful, factoring in your target length?
Delivery
Pace, confidence, and filler word count.
Memorability
Is there a clear 'so what' — a reason to remember you?
Pitch Practice is a dedicated mode for refining your personal elevator pitch — the 30–90 second professional introduction you use in interviews, networking, and career conversations.
Unlike the job-interview practice flow (which runs through 8 questions for a specific role), Pitch Practice focuses on one thing: getting your personal intro sharp. You record it, get feedback, and try again — as many times as you need. Your score across attempts is tracked so you can see the improvement.
You can choose from four contexts when setting up a pitch session:
Job interview — your 'tell me about yourself' answer
Networking — a warm, memorable professional introduction
Career pivot — explaining a career change or transition clearly
General intro — an all-purpose professional introduction
The prompt you're given and the coaching tips adjust based on the context you choose.
After each pitch attempt, Cadence gives you three specific, actionable pieces of feedback:
Keep — one specific thing working well in your pitch, often quoting your own words back to you
Add — one thing you should include that's missing or underused. If your resume is connected, this is cross-referenced with your actual background — so if you're underselling a key credential, it'll be called out specifically
Cut — one thing taking up time your pitch can't afford
One item each — not a list. The point is to make one meaningful change per attempt, not overwhelm you.
Your scores and progress
Your overall score is a weighted average of all your individual question scores. The AI also considers the arc of your session — not just the average — so a session where you warmed up and got stronger as you went is reflected in the feedback.
Your first session gives you a full scored report and Interpersonal Intelligence profile straight away — there's immediate value from session one.
Trend lines and delta comparisons start appearing from session 2. By session 3, the progress charts become genuinely informative. Most people notice a meaningful score improvement within 3–5 sessions of focused practice.
The 'Where to focus next' recommendation on your progress page identifies the single most impactful area to work on based on your sessions to date — the dimension that's lowest, declining, or holding you back the most.
It comes with a specific, concrete tip for your next session — not generic advice, but something actionable you can practise immediately.
Privacy and your data
Your voice is recorded during sessions to produce transcripts. The transcript — not the raw audio — is what's stored and used for scoring. Raw audio is not retained after transcription.
You can request deletion of all your data at any time from your account settings.
Your resume and LinkedIn data is used solely to personalise your experience — tailoring questions to your background and cross-referencing your credentials in pitch feedback.
It is never shared with third parties or used to train AI models without your explicit consent. You can remove it at any time from your profile.
Yes. From your profile settings, you can export all your data or request full deletion of your account and associated data. Deletion requests are processed within 30 days in line with applicable privacy regulations.