Recruiting4 min read
How staffing agencies screen hundreds of applicants without reading every CV
Collect applications with CV upload, let AI read each CV into a profile, and ask Medusa for a shortlist that fits the role.

Formzed team
formzed.com
Staffing and recruitment agencies often fill several roles at once, and each one can attract hundreds of applicants. Reading every CV in full isn't realistic, so good candidates get missed. This guide shows how to set up Formzed so that every CV is read, and your recruiters start with a shortlist instead of a pile.
The screening problem
Take a role with 300 applicants as an example. A recruiter who spends a few minutes on each CV loses days to first-pass screening, and that's before any calls. In practice, people skim. The first 50 CVs get careful attention and the rest get a glance.
Keyword search helps a little, but CVs describe the same experience in many different ways. One candidate writes "AWS", another writes "Amazon Web Services", and a third lists the projects without naming the platform at all. A strict keyword filter drops good people for the way they wrote their CV.
The goal isn't to remove recruiters from the process. It's to make sure every applicant is considered against the same requirements, and to give your team a ranked starting point.
Start with an application form that collects what you need
Everything starts with structured applications. Formzed includes recruitment templates you can use as they are or adjust. The Job Application Form asks for name, contact details, the role applied for, a required Resume / CV upload, a LinkedIn or portfolio link, years of experience, motivation and notice period.
Agencies often need more than one form. The Job Brief Intake Form helps clients describe a role so every search starts with aligned expectations. The Candidate Submission Form lets recruiters submit candidates with a CV and a short note on why they fit.
You can also describe the form you want in plain English, for example "an application for a senior accountant role with CV upload and salary expectations", and Formzed builds the fields for you.
Every CV is read into an AI profile
When an applicant uploads a CV, Formzed reads it in the background right after they submit. CVs in PDF and Word format work. AI reads the CV together with the applicant's form answers once, and writes a structured profile that includes:
- A headline and a short summary
- Total years of experience and seniority level
- Location
- Skills, with years of experience for each
- Past roles and companies
- Industries, education and languages
- Notable achievements
Reading a CV uses about half an AI credit, and each CV is read once.
Ask Medusa for a shortlist
Once applications arrive, open Medusa, Formzed's AI chat, and pick your application form. Medusa can use every profile to answer questions exactly, so you aren't relying on keyword matches in raw text.
Because the profiles are structured, a question about skills, seniority or location comes back as an exact count or list that covers every CV that has been read. A request for the best candidates works differently: Medusa compares each applicant's answers and CV with the role, then scores the strongest fits.
- 1Paste the job description or the key requirements into the chat. If you ask for a ranking without them, Medusa asks you for them first.
- 2Ask for the best fits, for example "Who are the strongest candidates for this role?" Medusa screens the candidates, reads their answers and CVs, and scores the top matches.
- 3Narrow the list with follow-up questions, such as "How many of them know AWS?" or "List the senior React developers in Germany."
- 4Open the original applications for your shortlist. Answers can point back to the actual responses, so you can check each one before you call.
Send candidates where your team already works
A shortlist is only useful if it reaches the right people. Formzed automations run when a form is submitted, and you can add a condition so a step only runs when an answer matches a rule, such as years of experience above a number you choose.
You can also add an Ask AI step, which lets AI summarize an application, score it or draft a reply as part of the automation. Automations connect to the apps recruitment teams use every day:
- Slack or Microsoft Teams, to alert recruiters about new applicants for a role.
- Google Sheets or Airtable, to keep a running list of every candidate.
- Gmail or Outlook, to send a confirmation from your own connected account.
- Greenhouse, BambooHR or Workday, to pass candidates into your hiring system.
- Google Calendar, Calendly or Cal.com, to book screening calls.
Screen as a team
Agencies rarely screen alone. Formzed team plans give you a shared workspace with roles (admin, member and viewer) and shared templates, so every recruiter starts from the same forms. Team Free covers up to 3 seats, and paid team plans start at 3 seats.
On paid team plans, AI credits are counted per seat. For example, Team Advantage includes 2,500 AI credits per seat, and those credits cover both CV reading and Medusa chats. If you run out, new CVs wait unread until you have credits again, so no application is lost.
A good habit is to write requirements about skills, experience and qualifications, and to keep the final decision with a person. Medusa helps you decide who to look at first. Your recruiters still make the call.
Screening at volume doesn't have to mean skimming. Start with a recruitment template, collect CVs on the form, and let Medusa give your team a ranked starting point for every role. You can try the whole flow on the free plan before you roll it out to your agency.