Case studies on hiring, not slide decks.
Field essays for teams who build.

What we've learned building Mirage: how compatibility matching works, why structured skills beat the CV, and what shifts when AI agents start sourcing. Real platform data, written for the people who hire.

Signals per profile
70+
Skills scored
40 across 7 categories
Built on
Real platform data
Read time
5–11 min

11 pieces, one operating thesis.

AI-Native
Case study·10 min

Matching on DNA, not keywords: how compatible teams form

Trade keyword search for a Compatibility Index that scores skills, motivation, and workstyle together. The shortlist it hands back looks nothing like the one Boolean gives you.

Talent Intelligence
Recruiting
Case study·9 min

The hiring agent that runs your first three interview rounds

Configure the pre-screen, technical, and collaboration interviews once. The agent runs them on every applicant, so your team only meets the candidates who already cleared the bar.

Screening · Automation
Recruiting
Case study·8 min

Skill telemetry as a hiring signal: 40 scored skills beat a scanned CV

Teams that hire on structured skill scores shortlist faster, mis-hire less, and give candidates a fairer read. Here's the mechanism underneath.

Skills · Hiring
Platform · Agentic
Case study·11 min

Profiles as MCP servers: talent data AI agents can actually query

Consent-gated, structured profile endpoints AI recruiting agents query directly — and what changes now that sourcing moves from keyword search to AI-agent discovery.

Platform · Agentic
AI-Native
Case study·9 min

The profile that updates itself: proven learning becomes live talent DNA

Close a gap in Mirage Academy and the proof rewrites your live profile, not a PDF. The next interview already knows what you mastered — and the DNA agents read stays current, because the system learns you once and keeps it in sync.

Talent Intelligence · Academy
Strategy
Case study·9 min

Structured job profiles cut time-to-hire by making fit legible

A complete Organisation Scouting Profile pulls better-fit applicants before the first message goes out. Here's what that looks like in practice.

Hiring · Strategy
People & Org
Case study·7 min

Mentorship at scale: connecting mentor DNA to mentee goals

How the Mirage Mentor Network gets past availability calendars and matches on motivation, expertise, and personal-mastery dimensions.

Community · Mentorship
Career
Case study·8 min

From sprint plan to offer: closing skill gaps before the next role

The 90-day Sprint Plan as a real career asset — how candidates use AI-guided task sequences to go from skill gap to interview-ready.

Career · Planning
Case study / 01·AI-Native · Talent Intelligence

Matching on DNA, not keywords: how compatible teams form

Replace keyword search with a Compatibility Index built on 70+ structured signals, and three things happen at once: the shortlist gets shorter, the hit rate climbs, and the interviews that never have to happen turn into the real saving.

At a glance
SectorB2B + B2C hiring
ScopeCompatibility-driven matching
AudienceTA leaders, people teams
DurationLive platform capability

SituationA sourcing loop that keeps finding the same people

Most funnels are a Boolean search in a trench coat. Write the job description, pull the keywords, run the query, get back a list of people who used the right nouns. Volume was never the problem; signal is. Two candidates can match the same terms and deliver completely different outcomes in the same role.

ApproachScore compatibility instead of running a query

The Compatibility Index rates each candidate–role pair on three independent axes: hard-skill match (skill telemetry against the job's scouting profile), motivation alignment (a candidate motivation survey against the company's culture dimensions), and workstyle fit (remote philosophy, decision pace, operating rhythm). Each axis is scored on its own, and the composite drives the ranking.

  • 40 structured skills scored per candidate across 7 competency categories
  • 70+ organisation-DNA signals pulled from the Scouting Profile
  • Motivation and workstyle scored independently, then combined
  • A full match breakdown per applicant, not a single black-box number

ExecutionThe shortlist is the product

Recruiters open a ranked list with the evidence attached: DNA alignment, hard-skill percentage, and which dimensions actually drove the match. The interesting names are the outliers — high motivation, low keyword overlap — the people a traditional search would have buried on page nine.

OutcomeFewer interviews, better calls

Teams running the Compatibility Index report shorter time-to-shortlist, lower early attrition, and candidates who say the role matched the pitch. The reason sits upstream: when both the job and the candidate are structured, the comparison is real instead of guessed.

Signature outcome
Compatibility becomes the hiring default, in place of similarity.
Signal axes
Skills · Motivation · Workstyle
Evidence per pair
Full match breakdown
Discovery model
Structure-first
Case study / 02·Recruiting · AI Agent

The hiring agent that runs your first three interview rounds

Configure the pre-screen, technical, and collaboration interviews once. The agent runs all three on every applicant — structured, consistent, and on the candidate's own schedule — so your team spends its hours on the shortlist, not the screening.

At a glance
SectorHigh-volume hiring
ScopeAutomated three-stage screening
AudienceHiring managers, TA leaders, job posters
DurationLive platform capability

SituationScreening is the most repetitive work in hiring — and the first thing that gets rushed

The first three conversations with a candidate are largely the same every time: confirm the basics, probe the core skills, then see how the person actually works with others. At volume that is dozens of near-identical calls a week. They get rushed, skipped, or handed to whoever is free, and that inconsistency is exactly where strong candidates fall through and weak ones slip past.

ApproachConfigure the pipeline once; the agent runs it every time

Job posters set up the three-stage interview pipeline per role: what the pre-screen confirms, which competencies the technical round probes, and what the collaboration assessment looks for. The agent then runs that exact pipeline for every applicant — same questions, same rubric, same fairness — and scores each stage against the role you defined.

  • Pre-screening: confirms must-haves, availability, and motivation before anyone spends a minute
  • Technical interview: probes the role's core competencies against the 40-skill rubric
  • Collaboration assessment: shows how the candidate reasons, communicates, and handles ambiguity
  • Fully configurable per role — you set the questions, the weights, and the pass bar
  • Candidates interview on their own schedule; no calendar tetris for your team

ExecutionYour team meets the shortlist, not the slush pile

Every applicant goes through the same three rounds and arrives with a structured result: stage-by-stage scores, the evidence behind them, and a clear pass or hold against the bar you set. Recruiters open a ranked, pre-evaluated list instead of a hundred unscreened CVs. The repetitive screening that used to cap how many people you could fairly consider stops being the bottleneck.

OutcomeMore qualified candidates seen, fewer hours spent seeing them

Because the agent never runs out of hours, you can screen every applicant properly rather than the first twenty who applied. Teams put their time into final-round judgment and candidate experience — the parts that genuinely need a human — while the consistent, structured screening runs underneath at any volume.

Signature outcome
Screening scales to every applicant; your team's time goes to the shortlist.
Stages automated
Pre-screen · Technical · Collaboration
Configured by
The job poster, per role
Coverage
Every applicant, same rubric
Case study / 03·Platform · Agentic

Profiles as MCP servers: turning talent data into agent-native signal

Candidate and company profiles are consent-gated MCP endpoints, so AI recruiting agents reason over structured skills and fit instead of scraping résumés and guessing. This is the layer that turns sourcing into discovery.

At a glance
SectorEnterprise · platform engineering
ScopeConsent-gated profile access
AudienceRevOps engineers, TA leaders, AI teams
DurationLive platform capability

SituationRecruiting agents with nothing structured to query

Today's AI sourcing tools scrape public profiles, parse free text, and hallucinate fit. The agent isn't the weak link; the signal source is. Unstructured HTML was never a talent API. Point an agent at a structured Mirage profile instead and it gets 40 scored skills, a motivation profile, a workstyle map, and a Compatibility Index value. A different class of input entirely.

ApproachConsent-gated MCP endpoints as the protocol boundary

Mirage exposes structured profile data as a remote MCP server. Every tool call is scoped by explicit consent: candidates control what agents can see, and organisations control which connectors can query their job and company data. The agent sees only what a human has permitted, and the audit log proves it.

  • search_candidates — structured filters, consent-gated results
  • get_candidate_profile — redacted by scope, no raw résumé or contact data
  • search_companies — active organisations with published Scouting DNA
  • find_job_matches — ranked candidates from the live Compatibility Index
  • OAuth bearer auth — every call authenticated and audited

ExecutionConnect an agent to real structured data

Platform and RevOps teams point AI scouting agents at real structured data instead of scraped HTML. Early connectors shaped the tool schema, consent model, and scope boundaries. The matching and profile data underneath are live — and so is the MCP transport that exposes them, behind OAuth and a full audit trail.

OutcomeThe talent layer for the agentic era

Profiles as MCP servers move Mirage from a hiring platform toward an infrastructure layer — the source AI recruiting agents query instead of scraping. Once agents read structured, consent-gated profiles, sourcing turns into discovery: they find people who fit rather than people who described themselves with the right words.

Signature outcome
Agents scout real signal, in place of scraped text.
Protocol
MCP · Streamable HTTP
Access model
Consent-gated · OAuth
Status
Live · consent-gated
Case study / 04·Recruiting · Skills

Skill telemetry as a hiring signal: 40 scored skills beat a scanned CV

A structured skill score across 40 competencies in 7 categories, with a radar chart, a market percentile, and a clear level, turns CV parsing into evidence. This is what a sourcing signal looks like when it's built on assessment instead of description.

At a glance
SectorCross-sector hiring
ScopeCandidate evaluation + sourcing
AudienceRecruiters, hiring managers, candidates
DurationLive platform capability

SituationA CV is a marketing document, not an evaluation

Candidates tune CVs for keyword density and recruiter attention. Recruiters scan for word presence, not proof. Neither side learns much about fit. What you get is a process that reliably surfaces people who describe themselves well, which is not the same as people who can do the work.

ApproachSelf-assessed scores, calibrated to the market

Skill telemetry collects self-assessments across 40 skills in 7 competency categories: Technical, Leadership, Communication, Analytical, Creative, Domain, and Interpersonal. Each score is plotted on a radar chart against the market's expected level for the candidate's target role. Out comes a competency score, a market percentile, and an honest read on strengths and growth areas.

  • 40 skills across 7 competency categories
  • Radar chart: your score against market expectation
  • One competency score and level, from Developing to Master
  • Top strengths and growth areas surfaced automatically
  • Market Value card: your percentile for the target role

ExecutionFrom self-assessment to sourcing filter

On the hiring side, recruiters filter by skill score, competency level, and specific gaps rather than keyword frequency. On the candidate side, people see exactly where they rank and which gaps map to mentors who can help close them. The same structured data feeds the Compatibility Index, so skills aren't merely visible; they're scored against the role.

OutcomeEvidence replaces description

When the candidate and the role are both described in structured skill data, matching becomes a comparison instead of a guess. Shortlists built this way surface people a keyword search would miss, and quietly drop the ones who write well but score below the bar on the skills the role actually needs.

Signature outcome
Skills scored, not described — sourcing on evidence.
Skills assessed
40 across 7 categories
Output
Score · Radar · Market percentile
Feeds
Scouting filters + Compatibility Index
Case study / 05·AI-Native · Living DNA

The profile that updates itself: how proven learning becomes live talent DNA

Close a skill gap in Mirage Academy and the proof doesn't end up in a certificate nobody reads — it rewrites your live profile. The next AI interview already knows what you mastered, and the talent DNA agents read stays current, because the system learns you once and keeps that knowledge in sync everywhere it matters.

At a glance
SectorB2B + B2C hiring
ScopeClosed-loop learning → live profile
AudienceCandidates, TA leaders, people teams
DurationLive platform capability

SituationA certificate is a dead end; a profile should be alive

Most learning platforms hand you a PDF and stop. The gap that sent you there is closed, but nothing downstream knows it. Your profile still shows the old you, the next interview re-asks what you already proved, and any system reading your data is reading a snapshot that went stale the day it was written. Learning that doesn't travel is learning you have to keep re-proving.

ApproachOne confirmed skill level, shared by Academy and interview

Mirage keeps a single source of truth for what you can actually do. When the Academy confirms a skill, it doesn't write to its own private record — it updates the same confirmed skill level the interview pipeline reads. There's no second copy to drift out of sync. The level you proved by learning is the exact level the next interview sees, with an advance-only guard so a hard-won gain can't be silently downgraded.

  • Confirmed skill levels live in one shared store — Academy writes it, interviews read it
  • Proof is verified, not self-claimed: your evidence is matched to the rubric before a level moves
  • A learning summary is re-embedded after every milestone, so retrieval always sees the latest you
  • Advance-only guard: a level you earned can't quietly slip backwards
  • No second record to fall out of date — the profile and the proof are the same object

ExecutionFinish a module, and the loop closes on its own

Complete a module and three things happen without you lifting a finger. Your evidence is routed to the rubric and the confirmed skill level is updated. A fresh learning summary — "completed N courses across these skills, most recent first" — is regenerated and re-embedded, replacing the previous one rather than stacking on top of it. From that moment your next interview can retrieve what you just learned. The growth isn't announced; it's already in the signal.

OutcomeLive DNA the next system always reads correctly

Because the proof updates the profile in place and the summary is replaced on every milestone, your talent DNA is never a snapshot — it's the current you. That's what keeps the platform's structured profiles relevant now that they're open to AI agents: profiles are consent-gated MCP endpoints, so an agent querying yours reads skills that were re-confirmed this week, not last year. The freshness isn't a sync job someone has to remember to run; it's built into how learning is recorded.

Signature outcome
Your growth is verifiable — and it travels with you.
Source of truth
One shared confirmed-skill store
Refresh trigger
Every completed module + certificate
Reads it
Interviews and MCP agents

Field essays for operators.

Essay 01Strategy

~6 min · Recruiting

The cost of keyword sourcing.

Keyword sourcing is cheap, quick, and quietly wrong. It surfaces the people who wrote the right nouns, which is a different crowd from the people who can do the job. You pay for the gap later: in interviews that go nowhere, in hires who leave in month four, in teams that underperform the shortlist that built them.

The people hardest to find with keywords are often the best hires.

Better signal changes the maths. When a recruiter can filter on motivation, workstyle, and real skill scores instead of word presence, the longlist shrinks and the hit rate climbs. The saving isn't in the search itself. It's in the dozen interviews that no longer need to happen.

There's a quieter cost, too. The strongest candidates are often the worst at keyword hygiene. They've been doing the work for years and haven't touched a document that nobody reads closely anyway.

Essay 02People & Org

~5 min · Leadership

Why compatibility beats culture fit.

"Culture fit" is a polite word for similarity, and the research on team performance doesn't reward similarity. What predicts good collaborative work is workstyle compatibility: how people plan their day, make decisions, and sit with ambiguity. Whether you'd enjoy the same offsite barely registers.

Compatibility is measurable. Culture fit mostly rehires whoever is already in the room.

Mirage models this with 70+ organisational signals — stability preferences, operating rhythm, remote philosophy, culture dimensions — scored next to the hard-skill requirements. In practice, a candidate who lands at 90% on workstyle and 80% on skills tends to outperform a culture-fit hire who simply used the right words.

Compatibility you can put a number on. Culture fit is a feeling, and feelings in hiring have a habit of hiring the same person twice.

Essay 03Platform

~7 min · Architecture

Structured data vs. free text: the hiring signal gap.

A résumé is a marketing document. It's tuned for searchability and a strong first impression, not for accuracy. That's not a knock on candidates; it's what the format rewards. The trouble is that we keep making real decisions on it.

Free text tells you what someone wants you to believe. Structured data shows what they can do.

The gap shows up in three places: skills that are claimed but never scored, motivation that's assumed but never asked about, and workstyle that nobody mentions until month two. Closing it takes a system that gathers structured signal at every step — assessment, profile, application — and hands it to decision-makers in a shape they can act on.

The teams pulling ahead on talent aren't running faster Boolean searches. They've swapped out the signal source underneath, and that is a product decision long before it's a recruiting one.

Common questions, answered straight.

What is the Mirage Compatibility Index?

The Compatibility Index is Mirage's matching score. It rates each candidate–role pair on three independent axes — hard skills, motivation, and workstyle — using 40 scored skills and 70+ organisational signals. Instead of a keyword hit count, you get a breakdown of why someone fits.

How is skill telemetry different from a CV?

A CV is written to be read; skill telemetry is scored to be compared. Mirage measures 40 skills across 7 competency categories, plots them against the market level for the target role, and returns a percentile. The result is evidence you can filter on rather than prose you have to interpret.

What does "matching on DNA, not keywords" mean?

It means the match is built from structured signal rather than word overlap. Mirage compares a candidate's scored skills, motivation profile, and workstyle against a structured job profile, so two people who list the same keywords can still rank very differently.

Can AI agents access Mirage profiles?

Yes. Mirage exposes consent-gated MCP endpoints so AI recruiting agents can query structured, permissioned profile data instead of scraping résumés. Candidates control what an agent sees, organisations control which connectors can query their data, and every call is authenticated and audited.

How is Mirage different from an ATS or a keyword sourcing tool?

An ATS stores applications; a keyword tool ranks word overlap. Mirage scores the signal underneath — 40 skills, motivation, workstyle — and matches on compatibility, so the shortlist reflects who can do the work rather than who described it best.

Can the AI interview agent be configured per job?

Yes. Job posters set up the three-stage pipeline — pre-screening, technical, and collaboration — for each role: the questions, the competencies probed, the weights, and the pass bar. The agent then runs that exact configuration for every applicant and scores each stage against the role.

Does what I learn in Mirage Academy update my profile?

Yes — that's the point. When the Academy confirms a skill, it updates the same confirmed skill level your interviews read, and re-embeds a fresh learning summary after every milestone. Your proven growth isn't stuck in a certificate; it becomes live talent DNA the next interview already knows about, and that the platform keeps current for the AI agents that query it over MCP.

Who are these insights for?

Talent acquisition leads, people teams, and RevOps engineers who own hiring outcomes. Every piece is a working note from building Mirage, written for operators who care how the result was reached.

From reading to doing

Run your next hire on structured signal.

Insights are easy to nod along to. Putting them to work is the part that moves your numbers. See how Mirage scores skills, models compatibility, and builds a shortlist you can defend.

Newsletter

Every other week, from the talent graph.

One pattern we're watching, one number worth knowing, one thing we shipped. No roundups, no filler.

Unsubscribe any time

Hiring Insights: Skills, Matching & AI Recruiting — Mirage