Talent Engage

A candidate lands on your branded careers page (e.g. /careers/your-company), completes a 30-second intake form (name, email, phone, right-to-work, FIFO availability), and then chats with the AI assistant. The assistant uses deterministic tools to record each answer (tickets, recent sites, medical status, salary expectation), evaluates server-side knockout rules in real time, and either ends the conversation politely with a knockout reason or scores the candidate 0–100 against the role. Recruiters see the result in their inbox within seconds of the candidate finishing the chat.

Yes. After the 30-second intake form, the candidate picks "Talk it out" (~8-min spoken interview with the AI recruiter) or "Type instead" (chat). Same screening questions, same knockout rules, same scoring, same recruiter inbox. Mic + headphones recommended for voice; a clear consent prompt fires before any audio is captured. Candidates can switch surface mid-conversation. Voice is recommended because completion rates beat a 12-question form on a phone, and a 9pm-Friday operator in Newman gets a real conversation instead of typing paragraphs.

Yes. Each tenant can clone one or more recruiter voices via a consent-gated wizard — the recruiter reads a 30-second consent script aloud and the provider verifies it matches the training audio (per Australian Privacy Principle 3.3 and ElevenLabs’ verified-voice requirement). Cloned voices are revocable any time and we delete provider-side and locally within 24 hours. Default ships with six preset voices including two Australian-accented (Marin female, Cedar male) so you can launch without cloning anything.

Each tenant gets a branded careers page (/careers/{your-company}/{role-slug}) hosted on drilledger.com.au. Direct applicants, referrals, and LinkedIn drivers come straight to that URL — no Seek ad spend required. You still post on Seek where it’s strategic, but every application that comes in via your careers page, your website, or a referral lands here without a per-ad cost. For a labour-hire firm running 8–15 ads a month at $200–$400 each, that’s $24K–$72K of avoided ad spend per year.

Knockout rules are hard-pass gates evaluated server-side (not by the LLM, so they can’t be hallucinated or talked around). Examples: "right-to-work must not equal no", "FIFO availability must equal true", "tickets must include White Card". Each role has its own rule set configured by the recruiter through the Talent Engage admin UI. When a candidate trips a rule, the chat ends respectfully with a knockout reason and the candidate is marked rejected with a reason recorded for audit.

When a recruiter clicks Promote to FIFO worker on a shortlisted candidate, the system copies the candidate into the Agent #1 FIFO Mobilisation worker pipeline, seeds the compliance tracking rows (DAS, PEM, White Card, site inductions), and creates the 21-item starter checklist (documents, medical, training, PPE, IT). Zero re-keying. This is the integration that Paradox, Sense, and other AI recruitment tools never built — they all hand off to a separate ATS or HRIS, leaving someone to re-enter the same data.

Yes. Every candidate, chat session, role, and interview slot is stamped with the recruiter’s tenant id and protected by Postgres row-level security — a recruiter can only ever see records inside their own tenant. The public careers page resolves the tenant server-side from the URL slug, so the browser never tells us which tenant to write into. A candidate landing on /careers/company-a can never accidentally land in company-b’s pipeline.

From 10 December 2026, the Privacy Act’s new automated decision-making (ADM) rules — APP 1.7–1.9, added by the Privacy and Other Legislation Amendment Act 2024 — require organisations to disclose when a computer program makes, or substantially assists, a decision that significantly affects someone. AI candidate screening is squarely in scope. Talent Engage is built for it: every careers page carries an ADM disclosure, server-side knockout rules and scoring are recorded against the role, and a real recruiter reviews every result before any hire-or-reject decision (human-in-the-loop). You launch ahead of the deadline instead of bolting a disclosure onto a generic screener at the last minute.

That’s the whole point. The AI assistant runs 24/7. A candidate who finishes their roster at 9pm on Friday in the Pilbara can apply, get screened, and have a clean shortlist record waiting for the recruiter on Monday morning — with right-to-work confirmed, tickets recorded, recent sites captured, and a 0–100 score against the role. Compare that to Seek + recruiter screening: 2–5 day delay between application and first contact, by which time the candidate has applied to three competitors.