How Craft Resume AI Works: From Building a Resume to Making a Hire

Craft Resume AI is really two connected products. On one side, job seekers build and export ATS-ready resumes. On the other, ShortlistAI gives hiring teams a way to post a role, collect applications, and screen candidates with AI -- without buying a full enterprise ATS. Here's what each side actually does.

For job seekers: building a resume

You start from one of three sources: upload an existing resume PDF, import a LinkedIn export, or connect a public portfolio site. The AI reads whatever you give it and drafts a structured resume -- work history, skills, summary -- that you then edit directly in a split-pane editor, with a live preview on the right that updates as you type.

Three themes are available (Modern, Corporate, Minimal), each a separate component so the layout genuinely changes, not just the color. When you export to PDF, the export renders the exact same template component the preview uses -- not a second, separately hand-built PDF layout -- so what you see while editing is what you get in the downloaded file.

A completion score and an ATS-compatibility check run alongside editing, flagging things like missing sections, unparseable formatting, or a summary that's too short, before you send the resume anywhere.

For hiring teams: ShortlistAI

ShortlistAI is the B2B side, built for a small internal hiring team rather than a staffing agency running hundreds of reqs. It works in roughly this order:

1. Post a role and start collecting applications

You write (or paste) a job description, set a budget range, and get a public apply link -- no login required for candidates. Applications come in with a resume, expected CTC, notice period, and any custom questions you've added to the form.

2. AI screens and ranks every applicant

Each resume is scored against the actual job description and must-have skills, fact-checked against what the candidate claimed on the form (a mismatch between stated experience and what the resume shows gets flagged as a discrepancy, not hidden). Budget-fit is scored separately and weighted in, so the ranked shortlist favors candidates who are both qualified and within range -- not just the most experienced person who's also the most expensive.

3. Technical assessments, drafted from the real job

For roles that need a skills check beyond a resume, you can generate a short assessment (mix of multiple-choice and written questions) drafted by AI directly from that position's job description and must-have skills -- not a generic question bank. You review and edit it before it ever reaches a candidate. Multiple-choice questions grade instantly and objectively; written answers are graded by AI against the same job context, with a score and one line of feedback per answer. Every submission is reviewable question-by-question inside the dashboard -- your answer, the correct answer, the score, and how long the candidate took -- so a technical reviewer can sanity-check the grading, not just trust a number.

4. Interview scheduling and mobile feedback

Shortlisted candidates pick their own slot from times you've published -- no back-and-forth email. After the interview, every interviewer (including outside panelists who don't have a dashboard login) gets a link to submit a quick scorecard from their phone: a 1-5 rating, a recommendation, and optional notes. It takes under a minute, which is the point -- feedback that requires opening a laptop and logging into an ATS tends to just not happen.

5. Tracking past the offer

Most hiring tools stop at "offer extended." ShortlistAI adds a Joined stage and tracks the gap between offer and actual start date -- relevant in India specifically, where notice periods commonly run 30-90 days, plenty of time for a counter-offer to pull a candidate back. If a joining date passes and the candidate is still sitting at "Offer," the recruiter gets flagged automatically instead of finding out a month later that the seat never got filled.

6. The team layer

Everything above works for one recruiter, but most hiring involves more than one person. ShortlistAI supports inviting teammates with distinct roles (Admin, Recruiter, Viewer), and enforces what each role can see server-side -- a Viewer, for instance, never receives compensation data in the response at all, not just a hidden field in the UI. Candidate-facing email can be sent from your own company domain instead of a shared address, and a full hiring report (funnel, scores, budget-fit bands) is one click away instead of a spreadsheet someone has to assemble manually.

Where the two sides meet

They're separate products for separate audiences, but they share the same underlying AI scoring approach: read the actual document (resume or answer), check it against the actual context (job description or claimed background), and surface a reasoned score with evidence -- not a black-box number. Whether you're the person building the resume or the person screening it, the goal is the same: less guessing, more evidence.

If you're job-hunting, Craft Resume AI gets you an ATS-ready resume from whatever you already have. If you're hiring, ShortlistAI gets you a ranked shortlist -- and now, an assessment and an interviewer scorecard -- without the overhead of a full enterprise ATS.