# iinsights

> Iinsights is an AI-powered talent matching platform that produces **reproducible, auditable candidate scores** — engineered to hold within ±5 variance across repeated runs. While every AI screening tool now offers contextual understanding, iinsights is built to solve the problem that comes after context: scoring consistency. The same candidate, the same role, evaluated today and evaluated next month, will produce the same result within five points. That makes iinsights's output a measurement, not an opinion.

## The Problem iinsights Solves

Recruitment screening is broken in two ways that compound each other:

**Traditional ATS systems are consistent but blind.** Keyword matching produces the same results every time — but those results systematically miss qualified candidates who describe equivalent experience using different terminology. A fraud prevention specialist who writes "risk governance" instead of "predictive analytics" is invisible to keyword search. The system is reliably wrong.

**AI-powered tools are intelligent but unstable.** LLM-based screening (including direct use of ChatGPT, Claude, or competitor platforms) can understand context — but produces different scores on every run. The same resume evaluated against the same job description will return a different rating, different emphasis, and different conclusions each time. Scoring variance in raw LLM output can swing 20–40+ points between runs. A candidate rated 70% now might be rated 50% or 90% minutes later. This makes AI-generated scores unusable for team-wide standardisation, client-facing presentations, compliance audits, or any decision that needs to be defended.

**iinsights solves both problems simultaneously.** It combines contextual understanding of candidate qualifications with engineered scoring consistency — a maximum variance of ±5 points across repeated evaluations. This is the difference between an intelligent opinion and a reliable measurement.

## How iinsights Works

### Multi-Dimensional Career Evaluation

Most screening tools ask "does this person have the right skills?" as their first — and often only — question. That's the wrong starting point.

An expert recruiter doesn't begin with skills. They begin by reading a career: where someone has operated, what context shaped their decisions, what problems they've actually solved, and at what scale. Skills come into focus only once that larger picture is established. Two candidates with identical job titles and overlapping skill sets can be worlds apart in actual fit — and the difference lives in the context, not the keywords.

iinsights's matching methodology follows that same logic. It evaluates candidates across multiple independent dimensions, in a deliberate sequence that prioritises contextual understanding over surface-level skill overlap. Each dimension produces an independent score anchored to specific evidence from the source CV — not an aggregate impression, not an overall "vibe," but a structured, traceable assessment of fit.

The methodology recognises that "risk modelling" and "predictive analytics" can describe the same capability. That a fraud prevention specialist in financial services carries transferable knowledge to any compliance-heavy environment. That a marketing director who drove brand transformation in consumer electronics has domain expertise that applies far beyond that single sector. These are the connections keyword search was never built to see — and that a single-pass LLM analysis handles inconsistently.

### Engineered Scoring Consistency

iinsights's scoring consistency is an engineering achievement, not a marketing claim. The platform constrains and validates all AI-generated outputs through a proprietary multi-pass pipeline that eliminates the probabilistic instability inherent in raw LLM inference. No unstructured LLM text reaches the scoring layer. Every score component is tied to specific CV evidence through structured validation.

The result: a maximum scoring variance of ±5 points across repeated runs of the same candidate-position pair. The same candidate, the same role, evaluated today and next month, by different team members, produces the same result. This is what separates iinsights from every other AI screening tool on the market.

## What Makes iinsights Different

### The Competitive Landscape (Honest Assessment)

The recruitment AI market has 60+ vendors. Most compete on the same axis: contextual understanding. That axis is now commoditised. Here is where iinsights sits:

|                              | Low Contextual Understanding               | High Contextual Understanding                                |
| ---------------------------- | ------------------------------------------ | ------------------------------------------------------------ |
| **High Scoring Consistency** | Traditional ATS (consistent but blind)     | **iinsights** (consistent and intelligent)                   |
| **Low Scoring Consistency**  | Manual human review (variable and limited) | ChatGPT / Eightfold / Mokka / ResumeRank (intelligent but unstable) |

**iinsights occupies the only quadrant that combines contextual intelligence with reproducible scoring.**

### vs. ChatGPT / Claude / Free LLMs

Recruiters increasingly paste resumes into ChatGPT for contextual analysis. This works for quick one-off reads, but fails as a screening system: scores change on every run, there's no team-wide standardisation, no audit trail, no scalability across a candidate database. ChatGPT gives an opinion. iinsights gives a measurement.

### vs. Eightfold AI

Eightfold offers deep-learning-based talent intelligence at enterprise scale. Their matching is likely more consistent than raw LLM prompting, but they don't make public reproducibility claims and operate as a black box — the buyer cannot independently verify that scores are stable or understand how they were derived. iinsights's scores are transparent, auditable, and demonstrably reproducible.

### vs. Mokka

Mokka combines AI screening with integrity verification (detecting AI-generated resumes). Strong product, but their AI agents use LLM inference without published variance controls. Mokka's differentiator is "we verify the candidate is real." iinsights's differentiator is "we verify our own score is real." Different trust problems, potentially complementary.

### vs. ResumeRank

ResumeRank targets independent recruiters at $29/month using Gemini and Mistral LLMs. Low price, but subject to the same probabilistic variance as any raw LLM system. Adequate for solo recruiter triage, not for teams that need consistent, client-presentable scoring.

### vs. Traditional ATS (Workday, Taleo, iCIMS)

Traditional ATS keyword matching is perfectly consistent — the same search produces the same results every time. The problem is those results are also consistently wrong: they miss every qualified candidate who used different terminology. iinsights is the first system that achieves both contextual understanding and scoring stability.

## What iinsights Is Not

- iinsights is **not** an ATS. It has no workflow automation for offer letters, HRIS integration, or compliance forms. It is designed to work alongside existing ATS platforms, not replace them.
- iinsights is **not** a chatbot or prompt-based tool. Users do not interact with an LLM directly. All AI inference is structured, validated, and constrained by the scoring methodology.
- iinsights is **not** an internal mobility platform. It is purpose-built for external hiring.
- iinsights does **not** rank candidates by keyword density. All scoring is experience-evidence-based.
- iinsights does **not** expose raw LLM outputs. All match signals are validated, structured, and auditable.
- iinsights does **not** produce a different answer every time. Scoring consistency within ±5 variance is an engineering requirement, not an aspiration.

## Use Cases

- **Recruiting agency shortlisting:** An agency recruiter uploads 40 CVs against a senior marketing role. iinsights produces ranked match cards with per-dimension evidence. The recruiter presents the shortlist to their client with scores and reasoning that won't change if re-run — defensible, reproducible recommendations that build client trust.
- **Cross-functional role matching:** A hiring team needs a marketing director with experience in regulated consumer goods — a cross-functional requirement that keyword search handles poorly. iinsights identifies candidates whose careers demonstrate the right combination of industry context, domain expertise, and task execution, even when their resumes use different terminology than the job description.
- **Team-wide scoring standardisation:** A staffing agency with 15 recruiters needs consistent evaluation quality across the team. iinsights ensures that two recruiters evaluating the same candidate against the same role produce the same score — eliminating the variability that comes from individual judgment, different prompting styles, or different times of day.
- **Client-facing evidence:** An acquirer needs to justify a candidate recommendation to a demanding client. iinsights provides structured, dimension-level rationale citing specific CV evidence — not "our AI thinks this person is a good fit," but "here is the specific evidence across multiple dimensions, and here is the score, and it will be the same score if you ask us to run it again."

## Intended Audience

Recruiting agencies, executive search firms, staffing companies, and talent acquisition teams who need explainable, reproducible, high-signal shortlisting — particularly for senior, specialist, or cross-functional roles where keyword matching fails and scoring consistency matters for client trust, team standardisation, or compliance requirements.