All guides

Sourcing

Searching: queries, JDs and filters

How to brief Talia well, what she does with a JD, how the filter brief works, and what happens when a search is genuinely narrow.

6 min read

Three ways to start a search

  • Type it: describe the role in plain English, including level, location, experience and budget if you have one.
  • Paste or upload a JD: PDF and DOCX both work. Talia reads it as a job description, separating the role being hired from reporting lines, and must-haves from perks.
  • Speak it: the mic button transcribes your brief and runs it through the same understanding.

What makes a great brief

Talia converts your words into a structured search: titles, location, experience range, skills, seniority, and company constraints like funding stage or industry. The more intent you give her, the more precise the net.

  • Name the level: "senior", "head of", "VP" change both the titles searched and the seniority guardrails.
  • Give a budget: "up to 25 LPA" becomes a budget-fit column across the whole pool.
  • Company qualifiers work: "funded startups", "product companies", "D2C brands" are translated into real structured constraints, not keywords.
  • Institution asks work: "any IIT" expands to all 23 campuses, "top tier" builds the full institutional net.

Skip the boolean habit. Writing "java AND spring AND NOT wipro" fights the engine; writing "backend engineers with strong Java, not from IT services" uses it.

The brief card: review before you spend

Before sourcing starts, Talia shows the brief she built: every filter as an editable chip, plus the outcomes she will score against. This is your checkpoint.

  • Remove a title chip and its variants go with it (dropping "marketing manager" also drops "senior marketing manager").
  • A live count shows roughly how many people match the current net, and it updates as you edit.
  • Screening criteria are shown separately from sourcing filters, so you can see what gates the pool vs what ranks it.

When the net is genuinely thin

Some searches are narrow in the real world, not just in the tool. When an exact net holds very few people, Talia widens it intelligently instead of failing silently: broadening exact titles into the equivalent function, moving a company-stack requirement into candidate skills, always preserving seniority, and telling you what she relaxed.

If even the widened net is small, you get the true number with an honest banner rather than a padded list. A shortlist of 12 real matches beats 250 lookalikes; you decide whether to widen further.

Refining mid-search

  • During sourcing, Talia may surface a suggestion card ("widen experience range?", "add these skills?"). Answering adjusts the running search.
  • After a search, "Evaluate more profiles" extends the same net deeper into the market.
  • Every search keeps its full query and brief, so you can revisit and edit rather than retype.

Frequently asked

Can Talia handle a full JD with nice-to-haves and perks?

Yes. The JD route reads structure: it identifies the single role being hired, separates must-have requirements from nice-to-haves, ignores marketing copy and benefits sections, and pulls the experience range the JD actually asks for.

Why did my search return fewer profiles than the count preview?

The preview counts the total matching population; the search evaluates a working pool from it, front-loaded with the best matches. If the population itself is small, the pool is honest about it and Talia offers ways to widen the net.

Can I just set filters manually without a query?

Yes. The manual filter panel offers the same structured fields Talia generates, including curated company pickers and skill inputs, and both routes feed the identical sourcing engine.