Match
From vacancy
to shortlist
For recruiters with a vacancy to fill. Paste, upload or link a job description; Jenny turns it into criteria you review, searches your own database — notes, interview feedback and earlier applications included — and LinkedIn Recruiter through a connected seat, then ranks the people it finds with the evidence behind every score.
The problem
A ranked list is only useful if you can see why
Most shortlists start with keyword searches across a database and LinkedIn, followed by an evening of opening profiles to check what the keywords missed. The result is a list of names, but not the reasoning a client will ask for.
The strongest evidence is often not in the CV at all: it sits in an interview note from last year or the feedback on an earlier application. And a score on its own does not help. You need to see which must-haves are proven, which are contradicted and which are simply not mentioned anywhere, before you pick up the phone.
What changes
- Requirements are extracted into must-haves and preferences that you confirm before any search runs; either-or requirements stay grouped and examples are not turned into must-haves.
- Recruiter notes, evaluations, interview feedback and earlier applications count as evidence, each quoted with its date and source.
- Your own candidates and LinkedIn Recruiter profiles appear in one ranked list instead of two separate searches.
- Every candidate shows supported criteria, gaps, risks and quoted evidence, so you can explain a shortlist to a client.
- Unknowns stay unknown: missing information is flagged for your call, never filled in by guesswork.
Senior Financial Controller · Antwerp
Illustrative — fictional dataCandidate A · Finance Manager
- Must-haves
- 75%
- Experience
- 75%
- Skills
- 60%
- Industry
- 100%
- Languages
- 100%
- Location
- 100%
- “Led the group consolidation for six entities” — interview note, March 2024
- “Fluent in Dutch, French and English” — profile
- SAP or Oracle: neither appears in the record
- Team leadership
- Notice period
- CV dated 2021 — confirm the current role
How it works
01
Add the job
Paste the description, upload a PDF, DOCX or TXT file, give the address of a public job page, or start from a job already in Jenny.
02
Review the criteria
Jenny proposes must-haves, experience, skills, industry, languages, location and seniority. Either-or groups and unknowns are kept; you edit and confirm before the job is saved.
03
Search your sources
Jenny searches your records, notes, interview feedback, application history and reviewed CVs and, where a seat is connected, LinkedIn Recruiter, within the page limits of each source.
04
Check the evidence
Open any candidate to see which criteria are supported, failed, contradicted or unknown, with dated quotes and older records marked for checking.
05
Shortlist
Attach the people you want to the job pipeline. Your decisions stay on the record.
Worked example: A Senior Financial Controller role in Antwerp
Illustrative — fictional data01 · Job added
You paste the vacancy of a logistics company in Antwerp. Jenny proposes IFRS consolidation as a must-have, keeps “SAP or Oracle” as one either-or requirement, lists Dutch and English, and keeps “e.g. Power BI” as an example rather than a must-have. No salary is stated, so it stays unknown.
02 · Criteria reviewed
You turn team leadership from a preference into a must-have, confirm the rest and save the job.
03 · Search runs
Jenny searches your workspace and the connected LinkedIn Recruiter seat. Candidate A is already in your database: an interview note from March 2024 says “led the group consolidation for six entities”.
04 · Ranked with evidence
Candidate A gets an evidence-based fit score of 78 with that note quoted and dated. Team leadership requires verification, and the CV, dated 2021, is flagged as more than three years old. Candidate B, from LinkedIn Recruiter, has no evidence on Dutch, so it shows as unknown.
05 · Shortlist made
You attach both to the job pipeline and note what to ask on the first call: current role, team size and notice period.
You provide
- A job description: pasted text, a PDF, DOCX or TXT file up to 4 MB, the address of a public job page, or a job already in Jenny
- Your review of the extracted criteria: what is a must-have, what is a preference and what is still unknown
- Which sources to search: your database, and LinkedIn Recruiter where a seat is connected
- Optionally, a LinkedIn Recruiter CSV export that you import into the job
Jenny produces
- A ranked list of candidates with an evidence-based fit score and a separate confidence figure
- Per candidate: mandatory status (passed, requires verification, failed), strengths, gaps and risks
- Dated, quoted evidence from CVs, profiles, notes and interview feedback behind each supported criterion
- A shortlist attached to the job pipeline
Sources and their status
| Your Jenny database | Available | Candidate records, recruiter notes, evaluations, interview feedback, application history, reviewed CVs in your CV library and people found in earlier searches. |
|---|---|---|
| LinkedIn Recruiter (via Unipile) | Early access | Live search through Unipile on a LinkedIn Recruiter seat connected by Forti8, within an allowance. Not a LinkedIn partner integration; an agency’s own seat cannot be connected per workspace yet, and Jenny sends no LinkedIn messages. |
| LinkedIn Recruiter CSV export | Available | Import an export into a job when you prefer to search inside LinkedIn yourself; the profiles then join the internal search. |
What it depends on
- Claude, the AI provider Jenny uses, for reading the job description; daily AI usage limits apply per workspace.
- Your candidates, notes and application history in Jenny — imported by CSV first if they live in another CRM or ATS.
- For LinkedIn profiles: a LinkedIn Recruiter seat connected in Unipile — today Forti8’s own, within an allowance; an agency’s own seat cannot be connected per workspace yet.
- Reviewed CVs in your CV library if you want CV text to count as evidence.
Approval and privacy controls
- You confirm the criteria before a search runs; nothing is searched on an unreviewed brief.
- The fit score is calculated by fixed rules from recorded evidence. AI can propose a quote as evidence, but it stays pending until you accept it with a reason, and your acceptance is kept as a new assessment.
- A requirement without evidence is marked “requires verification”, not failed. Only a recorded “not met” fails a must-have, and a contradicting note comes to you for review instead of excluding anyone.
- Protected characteristics such as age, gender, nationality or health are not ranking criteria.
- Adding a candidate to a pipeline is always your action.
Current limitations
- The fit score measures how much of the job’s criteria the recorded evidence covers. It is not a prediction of who will be hired.
- LinkedIn Recruiter searches are bounded to a few result pages per run and stop after about a minute. A retrieved profile is not a qualified one: check each before you call.
- Candidates whose records say little will score low on coverage even when they may fit — read the unknowns before ruling anyone out.
- Scanned PDFs need their text pasted, because Jenny does not read text from images. The CV library holds up to 200 reviewed CVs.
- Follow-up questions about the job, such as budget or start date when the description leaves them out, are in development.
- Instructions become criteria by fixed rules, not free interpretation: sentences Jenny can’t turn into criteria are listed for you, and travel or soft skills stay points to check at screening. The optional AI fit review also reads them, as guidance.
Questions about job-to-candidate matching
What does the score mean?
It is an evidence-based fit score: the share of the job’s criteria that the recorded evidence supports, weighted by type. By default must-haves count 40, experience 20, skills 15, industry 10, and language, location and seniority 5 each; these weights are the same for every job. Types without criteria are left out. A confidence figure and the list of unknowns sit next to it, so a low score with many unknowns reads differently from one with failed must-haves.
Does Jenny use my notes and interview feedback?
Yes, when it matches a job. Recruiter notes, evaluations, interview feedback, earlier applications and reviewed CVs count as evidence, each quoted with its date and origin — for example an interview note imported from your previous ATS. Records over three years old are marked “confirm still current”. The free-text search on the Candidates page is different: it covers structured fields such as title, skills, languages and location, not notes.
Can we search with our own LinkedIn Recruiter seat?
Not yet. LinkedIn Recruiter search runs through a seat connected in Unipile, which today is Forti8’s own, used within an allowance. Connecting an agency’s own seat per workspace is not supported yet. You can always run the search in LinkedIn yourself and import the CSV export into the job.
Does Jenny search job boards for candidates?
No. Candidate sources are your own Jenny database, LinkedIn Recruiter through a connected seat, and LinkedIn Recruiter CSV exports you import. Job boards and company data providers are not candidate sources in Jenny.
Is there a way to try this without an account?
Yes: the public demo on the Try Jenny page extracts the requirements from a job description and ranks profiles from a fictional demo database, so you can see the reasoning without sharing real candidate data.
Related features
- Candidate sourcing
Search the people you already have first, then LinkedIn Recruiter, in one list.
Available - Candidate database reactivation
Turn your existing candidate database into new placement opportunities.
Early access - CV generation
Tailored and anonymised client CVs built from recorded facts — nothing added.
Available - Presentations & outreach
Client introductions and personal outreach, drafted by Jenny and approved by you.
Needs setup
See it on your own vacancies and candidates
A short call with the founder: we look at how you work today and show the workflows that fit.