Product walkthrough

The whole path, end to end

Eight stages take an assessment from a blank template to a hiring decision. This page walks all of them with illustrative screens — then hands you the real thing.

Screens below are static mockups. The app has real seeded data behind it.

1Reviewer

Design the assessment

Start from a seeded template or write your own. Coding, document, research, and roleplay tasks all run through the same pipeline, so scores stay comparable across roles.

/app/templates

Debug a failing payment webhook

60 min

coding

Vendor evaluation memo

90 min

document

Underperforming report 1:1

45 min

roleplay
76 templates seeded — coding, document, research, roleplay
2Reviewer

Invite the candidate

One tokenized link — no account to create, no password to reset. The clock starts when they open it, and the token only works once.

/invite/[token]

Invite link

/invite/8f2c…a91d

Sana Okafor — Senior Backend Engineer

Single-use token · no account, no password · timer starts on first click

Start assessment
3Candidate

They work the way they actually work

A real editor, a real terminal, and an AI assistant in the same window. Nothing is sandboxed away from them, because the point is to watch how they use it.

/session/[id]
webhook.ts
export async function handle(req) {
  const sig = req.headers["x-sig"]
  if (!verify(sig, req.rawBody))
    throw new Error("bad signature")
  return process(req.body)
}

$ npm test

✗ 2 failing — raw body consumed

The signature check fails but only in prod. Here's the middleware order — what would you look at first?
`express.json()` runs before your verifier, so `rawBody` is already parsed. Capture it with a `verify` callback…
That matches the failing test. Before I apply it — does that change how retries are deduped?
4System

Every action is captured

Edits, commands, prompts, responses, accepted and rejected suggestions, pastes, tab switches, test runs. It is an append-only log, which is what makes the session replayable instead of merely summarized.

session_events
00:02:11reference_viewopened architecture.md
00:04:38ai_prompt142 tokens · included failing test output
00:05:02ai_responsemiddleware ordering fix
00:06:47ai_suggestion_rejectedcandidate pushed back on dedupe claim
00:09:15code_editwebhook.ts +14 −3
00:11:40test_run12 passed
21 event types captured · append-only · full session replay
5System

Two passes, one report

Deterministic metrics are counted from the log first, then a scoring model reads the transcript alongside them. Separating the two keeps the countable parts stable and the judgment parts accountable to quoted evidence.

analysis worker
Pass 1Deterministic metrics

Counted from the event log — prompt count, edit/AI ratio, time-to-first-test, rejected suggestions. No model involved, so it never drifts.

Pass 2LLM scoring

A scoring model reads the transcript and the metrics, then scores each dimension with quoted evidence from the session.

Runs on a background queue — reviewers aren't kept waiting
6Reviewer

Read the evidence, not a vibe

One AI-Q score across five dimensions, each backed by moments you can jump to in the replay. A high score you disagree with is auditable — that is the whole idea.

/app/session/[id]
79AI-Q score68th percentile in cohort
Problem decomposition82
Prompt quality91
Critical evaluation74
Efficiency68
Output quality79

Evidence: pushed back on the model's dedupe claim at 06:47 and verified against the failing test before applying the fix.

7Reviewer

Compare and decide

Same task, same rubric, side by side. Prompt counts and dependency ratios sit next to the score so a fast finisher and a thoughtful one don't look identical.

/app/hiring/compare
CandidateAI-QPromptsOwn work
Sana OkaforStrong verify loop79141.8×
Dev Raman71310.4×
Mira Halvorsen6463.1×
Same task, same rubric — differences are the candidate, not the prompt
8Candidate

Everyone leaves with something

A shareable credential for the candidate, and — in internal L&D mode — a learning path built from the same evidence that produced the score.

/credential/[token]

AI-Q Credential

79

Sana Okafor

Shareable, verifiable, and scoped to the assessment it came from

/path/[token]

Learning path

  • Verify before applying AI output
  • Decompose before first prompt
  • Reduce redundant re-prompting

Generated from the same evidence that produced the score

That's the path

The app is seeded with templates, an organization, and demo accounts, so you can run the same eight stages against real data.