Resumes are uploaded as PDF or DOCX, parsed server-side into plain text with section boundaries preserved, and passed to the model as context for the screening conversation. Extracted structured fields (companies, dates, roles) are also written to the candidate record so they are queryable. Most resumes parse cleanly; design-heavy creative resumes occasionally need a manual review.

Yes. Resumes in any of the major European languages (German, French, Spanish, Italian, Portuguese, Dutch, Polish) parse without trouble, and the model handles multilingual context. The bot can be configured to ask follow-ups in the candidate's language while writing structured output in your hiring team's working language.

Yes, with caveats. The bot can output a fit score per qualifier (years of experience, scale, location, authorization) and an overall aggregate. Most teams use these as ranking aids, not gating decisions, because resume parsing and self-reported answers both have noise. Conversation logs let recruiters audit any score that surprises them.

The system prompt is explicitly tuned to ignore name, photo, age, and other protected-class signals when generating follow-up questions or scoring. You can also configure the parser to redact those fields before the model sees them. Bias is never solved by an automated tool alone, but the bot can be configured to surface only role-relevant signal.

Yes. SleekAI fires a webhook with the parsed resume fields, the follow-up answers, the fit score per qualifier, and the full transcript URL. Native or near-native paths exist for Greenhouse, Lever, Workable, Recruitee, and Personio. Custom ATS setups receive the same JSON payload to any endpoint you configure.

Yes, and most teams use it that way. The bot handles the first-pass screen, surfaces a structured profile and transcript, and the recruiter steps in for the second-pass conversation with full context. The recruiter's call is faster and richer than a cold first call because the bot has already covered the basics.

In your own WordPress database, tied to a candidate record post. Resumes live in the media library, structured fields in postmeta, transcripts in a dedicated SleekAI conversation table. No data routes through a SleekAI server. GDPR subject-access and erasure requests work via standard WordPress export and erasure tools.

Yes, with jurisdiction-aware behavior. In jurisdictions where salary-history questions are restricted or prohibited (California, NYC, several EU states), the bot is configured not to ask. It can ask about expected range, framed as a band rather than an exact number, with the same opt-in language a static form would use.