What Is Expertini's AI-in-ATS Recruitment Technology? in Ouges, Bourgogne-Franche-ComtéFrance
How Expertini ATS handles What Is Expertini's AI-in-ATS Recruitment Technology? for employers hiring in Ouges. View without a localization
What Is Expertini's AI-in-ATS Recruitment Technology?
Where AI is actually used in the hiring pipeline, and — just as importantly — where it deliberately isn't.
"AI recruitment technology" is a broad, often vague claim across the industry. This page states specifically where Gemini-class AI is used inside Expertini ATS, and where the architecture deliberately keeps a human or a fixed formula in control instead.
On this page
- Where AI is used
- Where AI is deliberately not used
- Usage limits, honestly stated
- What the AI never sees
- How to evaluate any vendor's "AI recruitment" claim — including this one
- Model versions and score stability over time
- How this is architected in the platform
- Operational and audit posture
- Frequently asked questions
01Where AI is used
Reading comprehension only: extracting structured evidence from a PII-stripped CV against a job's specific requirements (CMS scoring), drafting a first-pass job description from a title and notes (AI job description drafting), and generating interview question suggestions. In every case, the AI's job is to read and extract — never to produce a final score, ranking, or hiring recommendation on its own.
02Where AI is deliberately not used
The CMS score itself is computed by ordinary, deterministic code from the AI's extracted evidence — not by asking the model "rate this candidate 1 to 100," which is the industry's most common and most fragile pattern, since an LLM asked for a score won't reliably reproduce the same number twice. Hard-blocker enforcement (a missing mandatory requirement pins a dimension to zero) is fixed logic, not a model's judgement call.
03Usage limits, honestly stated
AI Prompts (job-description drafting and similar generation tasks) and CMS scoring are separate, plan-limited budgets — Starter gets 25 AI prompts and 25 CMS scores per month, scaling up through Growth, Professional, Business, and Enterprise — because every Gemini call has a real cost, and pretending otherwise would be dishonest about the unit economics.
04What the AI never sees
Before a CV reaches any AI step, pattern-based stripping removes names, email addresses, phone numbers, physical addresses, and demographic-suggestive language — eleven distinct PII patterns in total. The model evaluates evidence of competence against the job's requirements without knowing who the candidate is, which narrows the surface area for demographic pattern-matching even unintentionally. This isn't a perfect guarantee (no text-processing step is), and the Privacy Policy states its limits plainly — but it's a structural safeguard that runs on every single scoring call, not an optional setting a busy recruiter has to remember to enable.
05How to evaluate any vendor's "AI recruitment" claim — including this one
Three questions cut through most AI-recruitment marketing. First: does the AI produce the final number, or does fixed code? If a vendor can't answer precisely, assume the model produces it — which means the score isn't reproducible. Second: can the vendor show you the formula? A published methodology can be independently checked; "proprietary algorithm" cannot. Third: what does the tool do when a candidate is missing a mandatory requirement — silently filter them out, or flag it for a human to decide? Expertini's answers are: fixed code computes the score, the formula is published in full as a citable paper, and hard blockers are flagged prominently rather than auto-rejected. Ask every vendor on your shortlist the same three questions and compare the answers you actually get.
06Model versions and score stability over time
Determinism means identical inputs through the same extraction model version always produce the identical score. If the underlying extraction model is upgraded, extraction quality can change — which is exactly why the score's dimension-level rationale matters: a recruiter or auditor can always see what evidence was found and how the arithmetic ran, on any date, for any historical score. A black-box tool that changes its model silently leaves you with numbers you can no longer explain; here, the explanation ships with every score.
Platform architecture & operations
A1How this is architected in the platform
What Is Expertini's AI-in-ATS Recruitment Technology? is not a bundle of point products — it is a slice through one platform. The platform is deliberately server-rendered: every view is prepared by the application server and shipped as complete HTML, with no client-side framework, no third-party CDN scripts, and no build pipeline between the data and the page. What renders is what the server computed — the property that makes the interface auditable.
All persistence runs on a single search-native document store; every query carries the organisation's identifier as a mandatory filter at the lowest query layer. Tenant isolation is therefore structural — a property of how every request is composed — rather than a policy that relies on application code remembering to check.
Every capability referenced on this page resolves to a registered tool or connector: the tools directory and the integrations catalogue are renderings of the same registries the application enforces at runtime, so what this page describes and what the product gates can never drift apart.
A2Operational and audit posture
Screening is deterministic and published — the same inputs produce the same outputs, hard requirements block rather than average away, and the methodology is public on the research page. Actions that touch external systems are explicit and journalled per event; usage reporting aggregates the same journals the actions write, not a parallel telemetry system.
Anything that leaves the request path — notification fan-out, webhook delivery, activity journalling, mail — runs in fire-and-forget background threads. A slow external endpoint can never make the interface hang, and a failed side effect is logged rather than silently retried into inconsistency.
Everything written is yours to take: CSV exports and the Data Export app cover the same stores the product itself reads. The exit is as open as the entrance — by design, not concession.
Frequently asked questions
Which AI model does Expertini ATS use?⌄
Does the AI ever reject a candidate automatically?⌄
Why are AI features usage-limited instead of unlimited?⌄
Will the same candidate get the same score if I re-run it next month?⌄
At a glance
- AI performs extraction/reading only, never the final score
- CMS score computed by deterministic code, not an AI-generated number
- Hard-blocker logic is fixed, not a model judgement call
- AI Prompts and CMS scoring are separate, plan-limited budgets
- Eleven PII patterns stripped before any AI reads a CV
- Dimension-level rationale ships with every score, forever auditable
See what is expertini's ai-in-ats recruitment technology? on your own hiring.
Bring a real job description to a 30-minute demo — free trial included.
Book a demo