How candidates get into Lovelio

Signed in? Ask Lovelio this question inside the app - it answers from this same page.

Candidates enter Lovelio through six doors: applying on the hosted careers page, applying through the embed widget on the agency's own website, bulk CV import, manual add by a consultant, Ask Lovelio, and capture from the Chrome extension. Every door feeds the same resume parsing and enrichment pipeline, so a candidate looks the same in the talent bank no matter how they arrived.

The apply flow (careers page and embed widget)

The form asks for name and email (both required), phone, LinkedIn, a notice period dropdown, and a cover letter. The cover letter is required: typed (at least 50 characters on the careers page) or uploaded as a file up to 5 MB. Uploading a resume (PDF, DOC or DOCX, up to 10 MB) is optional but does two things: it auto-fills the form, and it powers the full parse and assessment after submit.

The moment a resume lands, a two-phase parse pre-fills the form:

Phase 1 is a fast pattern scan that fills name, email, phone and LinkedIn URL in a fraction of a second.
Phase 2 is an AI read of the whole document that refines those same four fields. It typically takes a few seconds, up to about 15 for long documents.

Both phases run at the same time. Anything the applicant has already typed is never overwritten by a parse result. The form shows "Resume parsed - please review your details" when done. On the careers apply modal, dropping a resume anywhere on the first step starts parsing immediately. If the file cannot be read (a scanned or password-protected PDF, or one that holds no text), neither phase counts as a parse and the form shows "We could not read that file. Try a Word document or another PDF, or type your details below"; the applicant types their details and can still submit. Lovelio tries a second PDF reader before giving up, so a PDF the first reader silently dropped still fills the form.

On submit, Lovelio matches the applicant to an existing candidate by email. An existing live candidate gets a new application on their record; a brand-new email creates a candidate. If a candidate with that email was previously deleted, the application creates a fresh visible candidate rather than vanishing into the deleted record. The applicant receives an acknowledgement email (replies to it thread onto the application), the consultant is notified, and the intake pipeline runs in the background: full resume parse, then AI assessment against the job, then a triage decision. If the parse fails, the assessment still runs on whatever the applicant typed; a bad file never silently parks an application.

What parsing extracts

The full AI parse captures: name, email, phone, location, LinkedIn URL, current title and employer, a summary, full employment history with employer, title, dates and role descriptions, education (degrees, certifications, training), up to 30 skills with years of experience and last-used year where the document states dates, and referees if listed. Salary expectation, notice period and work rights are captured only when the resume explicitly states them; most resumes do not, so blank is normal there. Parsing never invents data - if it is not written in the document, the field stays empty. Parsing never overwrites details a recruiter has already entered by hand; it only fills gaps.

Bulk CV import

From the People page, drag and drop CV files, a whole folder, or a zip archive. Zips are unpacked automatically; anything that is not a PDF, DOC or DOCX, or is over 10 MB, is skipped with a note. Limits: up to 5,000 files per import, 10 MB per file, and a 5,000-CV allowance per rolling 30 days across everything you bring in yourself (imports, and CVs pushed in through the API). CVs from people applying to you never count towards it. Moving a whole database across is a migration Lovelio runs for you rather than something to self-serve: email support@lovelio.ai and the whole book comes over properly, history included.

Import starts the moment files are dropped - there is no confirm step. The rules:

The import never fails as a whole. An unreadable file is marked failed with a plain-language reason in the end report; one bad file never stops the rest.
Duplicates are matched by email. An existing candidate is merged: their CV is updated and gaps are filled, but nothing a recruiter typed is overwritten. A new email creates a new candidate.
A "Sort them into your talent pools" toggle (on by default) judges each arrival against pools that have AI matching enabled.
Everyone is named and contactable as soon as their file lands. The deeper read - skills, work history, education, current role and employer - follows behind: minutes on a small import, up to a day on one of 200 files or more. Either way nothing is lost, and the people are there to work with straight away.
Imported candidates land in the talent bank with no application, so bulk import never triggers per-role AI assessments. Once profiles are indexed, Lovelio ranks the batch against your open roles using similarity matching only - a "looks relevant" shortlist, not a verdict. Assessing against a role is a separate deliberate step a consultant takes from the shortlist.

Import history and progress live at Candidates then Imports.

Manual add

"Add candidate" opens a page where dropping a CV parses it into an editable draft: quick fields (name, email, phone, LinkedIn) appear almost instantly, then the AI pass fills location, current title and current employer in a few seconds; the consultant reviews and saves. A password-protected PDF, or one with no readable text, is refused here with the same message the careers form shows: "We could not read that file", or "That PDF is password-protected", with the ask to upload another copy or type the details. The uploaded CV is kept on the record so "View CV" works later. A small "or enter manually" link skips the CV entirely for typing details by hand; name and email are required. Adding a candidate directly to a job from the job page creates the application and runs the same parse-assess-decide pipeline as a public application. Creating a candidate with an email that already exists in your agency is blocked as a duplicate rather than creating a second record.

Ask Lovelio

Drag a PDF, DOC or DOCX resume onto either Ask Lovelio input, or click its paperclip. The file is staged securely and parsed before the assistant prepares the candidate review. You can also say "add Sally Smith, sally@example.com, ask me for the resume": the review waits for the file and cannot be confirmed until it is attached. The original resume is kept on the candidate record. If typed details and the resume differ, typed details win and the review flags the conflict for you to check. No candidate is created until you confirm.

Chrome extension (summary)

The extension captures candidates from LinkedIn and other profile pages: name, title, employer, location and a professional AI summary. That summary is stored on the candidate's profile and feeds their search profile, so an extension-captured candidate with no CV can still be matched to jobs and pools. Full detail is on the extension's own card.

What happens after arrival

For every door, once the parse lands Lovelio automatically: builds the candidate's search profile and indexes them for AI search (this runs in the background and usually lands within the hour - the profile fields themselves show straight away), resolves their location and normalises their skills and titles for matching, syncs their employment history into the movement-tracking ledger (which can fire job-change alerts), checks them against talent pools with AI matching on, and cross-matches them against your other active jobs - a strong fit for a different role surfaces as an alert. None of this needs setup. Re-uploading the same CV never re-parses it: Lovelio recognises identical documents and reuses the parse it already has.

Common questions

Why does a candidate show no current role right after import? The deep parse runs in the background and takes seconds to a couple of minutes per CV; large imports process in parallel and fill in over minutes.
Why did an applicant not get assessed? Assessment runs for applications only. Bulk-imported candidates have no application by design; assess them deliberately against a role.
Why is salary or notice period blank on the profile? Parsing only captures them when the resume states them in writing. The apply form does ask for notice period, so applications usually carry one.
A resume failed to parse - what now? Check it is a text-based PDF, DOC or DOCX under 10 MB and not password-protected. Scanned image-only PDFs cannot be read. The record and the application show a "CV not read" flag with the reason, and it clears on its own when a readable copy is uploaded.
Do duplicate applications create duplicate candidates? No. Matching is by email; a repeat applicant gets a new application on the same record.