The AI Co-pilot explained
An in-project assistant that screens and qualifies candidates against criteria you set, working like a junior recruiter you instruct in plain language.
PUBLIC · Concept Explainer
The Co-pilot is the feature most clients ask about first after their demo. It’s also the one that changes how the day-to-day feels: less profile-reading, more decision-making.
Why it matters
Section titled “Why it matters”Sourcing produces a long list. Qualifying it is the slow part. Even with great filters, you still read each profile and decide whether it’s worth contacting.
The AI Co-pilot takes that off your hands. Ask a question in plain English (“Has this candidate worked in consultancy for at least five years?”). It reads the profile, makes a judgment, and adds the answer as a new column with a one-sentence explanation.
It doesn’t replace your judgment. It makes the obvious calls so you can spend your time on the borderline ones.
How it works
Section titled “How it works”The Co-pilot lives inside Projects. Two ways to use it.
Presets
Section titled “Presets”One-click columns for common signals like:
- Career summary
- Open to work
- Likely to switch
- Job hop risk
- Key skills
Custom Intelligence Columns
Section titled “Custom Intelligence Columns”Type any instruction and the Co-pilot creates a column answering it for every profile. “Is the candidate freelancing?” “Fluent Dutch and French?” “Managed teams of 10+?”
Some instructions filter the existing view. Others create a new column with yes / no / unclear answers plus reasoning. The Co-pilot tells you which it’s doing and confirms before creating columns.
Custom AI Columns per project depend on plan: Explorer 1, Hunter 3, Strategist 5, Oracle unlimited.
When to reach for the Co-pilot instead of filters
Section titled “When to reach for the Co-pilot instead of filters”Some criteria look like filter work but produce bad results when forced into a structured filter. The Co-pilot handles them better because it reads the full profile and reasons about it.
- Languages. Candidates don’t always list every language they speak, especially their native one. A Flemish recruit in Antwerp might have nothing under languages. The Co-pilot infers language from bio language, location, company names, and other signals.
- Role-specific tenure. The Years of Experience filter counts total career years. If you need “5+ years specifically in fintech” or “3 years at their current company”, ask the Co-pilot. It reads the work history and judges per role.
- Domain or subject-matter fit. “Has worked at a subscription business” or “has scaled a B2B SaaS marketing function” can’t be filtered for. The Co-pilot can read company descriptions and role context to assess it.
- Combined criteria. “Senior digital marketer with both e-commerce and SaaS experience who also led a team of 5+” is one Co-pilot instruction. As filters, it’s three or four overlapping criteria that often return zero results.
The pattern: filter broad to get the list (250-400 profiles is a healthy size), then qualify with the Co-pilot to shortlist.
Common misconceptions
Section titled “Common misconceptions”It doesn’t make hiring decisions. It surfaces signals. The decision stays with you.
It doesn’t cost credits. Co-pilot work is plan-gated, not credit-gated. Credits are spent only when contact info is revealed and found.
It only works inside a project. To use it, save the relevant candidates to a project first.
It can’t research the open web yet. Today the Co-pilot assesses profile data in your project. Company-level research (funding, news) is on the roadmap. What it does now, it does fast and at scale.
Related articles
Section titled “Related articles”Still stuck?
Your customer success manager reads faster than any search bar.