Skip to content

Fill custom candidate fields from resumes automatically

Standard resume parsing captures the basics.

The gap is usually the detail your team actually screens on — certifications, work authorisation, notice period, or the tools someone has used in production. A custom field parsing agent closes that gap by reading each resume against descriptions you write for your own fields, then populating those fields without manual data entry.

How it works: describe, enable, parse

The workflow has three parts. First, you define custom fields that reflect how your team evaluates candidates. Second, you add a description to each field explaining exactly what information belongs there. Third, you enable the Parsing Agent so that, whenever a resume is parsed, it looks for that information in context and fills the matching fields.

The description is the control point. A vague label such as “Skills” gives the agent little to work with. A specific description — for example, tools and platforms the candidate has run in production, captured as a multi-select — tells it what to find, where it might appear, and how to format it. That is the difference between a field that stays empty and one that fills reliably.

A practical setup checklist

Use this checklist before enabling extraction:

  • Name each field for the decision it supports, not just the data it holds.
  • Write the description as a definition: what counts, what does not, and what format to use.
  • Match the field type to the answer — dropdown for a fixed set of options, multi-select for several items, single line for a short fact, date for a date.
  • Test with a few sample resumes and check whether the extracted values match what a recruiter would expect.
  • Review fields that stay blank: the resume may not contain the detail, or the description may need tightening.

Because settings apply across the account, agree the field set with the teams who will rely on it. One shared definition prevents the same candidate detail being captured differently by different recruiters.

What good extraction looks like in practice

Consider a candidate whose resume lists several certifications, a stated notice period, and right-to-work status for a particular country. With matching custom fields and specific descriptions, each detail can land in its own field during parsing — certifications as a multi-select, notice period as a single line, work authorisation as a dropdown — instead of remaining buried in free text that a recruiter must re-read and retype.

The same pattern applies to fields such as core tech stack, domain expertise, team leadership, and security clearance. The examples differ, but the principle is constant: the description defines the target, and the agent extracts only what the resume supports.

What it does not do

Two distinctions matter. First, this agent complements regular resume parsing rather than replacing it: standard parsing handles the conventional resume data, while the agent adds context-based extraction into your custom fields. Second, it works at parse time. It does not automatically scan or update every existing record in the database; older records are enriched by reprocessing their resumes through the normal workflow.

Accuracy follows from those two inputs — the resume and the field description. Where the resume is silent or unclear, the agent leaves the field unchanged rather than inventing a value. That restraint keeps records trustworthy: a blank field signals missing information, not a hidden guess.

Adjusting extraction as needs change

Hiring criteria change. New roles may require tracking details you did not originally capture, or a field definition may need refinement after testing. Parsing configuration can be updated accordingly, so future resumes are parsed against the revised fields and descriptions. Plan reviews of your field set alongside changes in the roles you recruit for.

Example custom fields and their descriptions
Custom fieldWhat the description should defineField type in the example set
CertificationsProfessional and industry qualificationsMulti-select
Notice periodStated availability to startSingle line
Work authorisationRight-to-work status by countryDropdown
Core tech stackTools and platforms run in productionMulti-select
Domain expertiseIndustries the work sat inMulti-select
Team leadershipLargest team led, if statedSingle line
Security clearanceClearance level, if mentionedDropdown

Examples are drawn from the captured control set; field types shown are the types paired with those examples there.

Frequently asked questions

Does the Parsing Agent replace regular resume parsing?

No. It sits alongside the standard resume parsing step and adds context-based extraction into custom fields, so your existing workflow stays in place while records gain more structured detail.

Can I use the Parsing Agent for existing candidates in my database?

It runs when a resume is parsed, rather than sweeping the whole database automatically. For older records, reprocess the relevant resumes through the normal parsing workflow.

How accurate is the information extracted by the Parsing Agent?

Results depend on what the resume actually contains and how precisely each field description defines the detail you want. Clearer resumes and clearer descriptions produce more reliable extraction.

Can different teams configure different parsing fields?

Settings are shared at account level, so enabled fields are consistent across the workspace. That consistency helps every recruiter work from the same candidate data.

What happens if a candidate’s resume contains conflicting information?

When resume context is unclear or absent, the field is left unchanged instead of being filled with an unsupported guess.

Can I change which information the Parsing Agent extracts later?

Yes. Custom fields and their descriptions can be revised as hiring needs change, and future parsing will follow the updated configuration.