AI recruitment software, with the controls most vendor pages don't show you
This page has two jobs. First, a vendor-neutral explanation of what AI recruitment software actually does at each stage of a hiring pipeline. Second, a specific, verifiable account of how Treegarden's Edera AI is built: which parts you can switch on and off, who reviews what the AI recommends, what gets logged, and which US and EU laws apply to any of it. No invented certifications, no vague "responsible AI" language without a source behind it.
What AI recruitment software actually does
Search this phrase and you mostly get two kinds of page: a marketing page that promises "smarter hiring" without saying what the software actually does, or a dense whitepaper written for compliance teams. Here is the pipeline, stage by stage, independent of any single vendor, so you know what to ask about regardless of who you buy from.
Sourcing
Tools search an internal candidate database, a talent pool, or external channels, and rank people by how closely their profile matches a role's requirements. This is the stage furthest from an actual hiring decision, since it only surfaces candidates for a human to look at, so it carries the least regulatory attention of the five stages here.
Screening
This is where resumes get parsed, skills get extracted, and candidates get scored or ranked against job criteria, sometimes through a chatbot conversation instead of a form. It is also where most hiring bias risk concentrates, because screening filters volume before a human ever reviews an individual application, which is why laws like NYC's Local Law 144 are written specifically around this stage.
Interview
AI can generate structured interview questions tailored to a candidate's background and the role, coordinate interview scheduling across calendars, and summarize interview notes or transcripts. A human interviewer is directly involved at this stage, which lowers risk somewhat, but a biased question set can still shape outcomes before anyone notices.
Decisioning
This is the highest-scrutiny stage, and where most vendor pages get vague on purpose. The question that matters: does the tool recommend a decision, or does it take the decision. Recommend-only tools generally sit outside the strictest audit regimes. Tools that can advance or decline a candidate without a human step are exactly what NYC Local Law 144, Illinois's new AI law, and the EU AI Act are built to constrain.
Analytics
Pipeline funnel metrics, time-to-hire, and source effectiveness are the commercial side of this stage. The compliance side is adverse-impact monitoring: statistical checks for outcome skew across demographic groups, which is less a hiring feature than an operational safeguard against bias the screening stage may have introduced.
Nine switches, plus a per-job override
Edera AI is Treegarden's AI recruitment suite. It is built around a control model, not a single on/off switch: five company-level toggles you enable independently, four further automations layered on top, and one switch that lets you turn AI on for a specific job and off for another, on the same company account.
Five company-level toggles
Candidate ranking
Scores each candidate against the specific requirements of the job they applied for, with a breakdown of matched criteria rather than a bare number.
Screening question generation
Drafts screening questions for a job posting based on the role's requirements, for a recruiter to edit or approve before it goes live.
Interview question generation
Produces structured interview questions tailored to a specific candidate's background and the role, so every interviewer works from a comparable framework.
Candidate chat assistant
A conversational assistant candidates can use for status updates and common questions, reducing the manual back-and-forth a recruiter would otherwise handle by email.
Pipeline auto-move
Moves candidates between pipeline stages automatically once configured conditions are met, instead of a recruiter dragging every card by hand.
Four further automations
Intelligent auto-advancement
Candidates who clear a threshold you set can move forward without a manual click, but only under the rules your company configures, with the review queue able to intercept any action first.
AI interview scheduling
Coordinates interview slots directly against Google or Outlook calendars, removing the email chain that usually goes with finding a shared time.
Decision suggestions
The AI proposes an advance or decline recommendation with its reasoning, for a recruiter or hiring manager to confirm rather than research from scratch.
Offer-letter draft generation
Drafts an offer letter using the role, level, and compensation details already in the system, for a human to review before it is sent.
Plus: a per-job AI switch
All nine toggles above are set at the company level, but AI can also be turned on or off for one specific job at a time. That means a high-volume warehouse-associate role can run with full automation while a confidential executive search runs entirely manually, from the same account, with no separate contract or environment required.
What sits behind every AI action
Every toggle above runs inside the same governance layer, whether it is switched on for one job or every job in the company.
AI Decision Review Queue
A human review workflow built for GDPR Article 22, which restricts solely-automated decisions with legal or similarly significant effect. A reviewer sees the AI's recommendation and its reasoning before any consequential action becomes final.
Automated-processing notices
Candidates see a notice that AI is involved in processing their application, rather than finding out after the fact.
Three separate audit logs
Decision audit logs record every AI-influenced hiring decision. Offer audit logs track offer-letter activity. Toggle logs record when a company or job-level AI setting was changed, and by whom.
Adverse-impact monitoring
Statistical monitoring for outcome skew across demographic groups, the kind of check bias audit laws like NYC Local Law 144 require employers to be able to produce.
EEOC compliance service
EEO-1 reporting and EEO analytics for the federal reporting obligations that apply to many US employers and federal contractors.
FCRA compliance service
Background-check disclosures and the adverse-action flow the Fair Credit Reporting Act requires when a background check contributes to a hiring decision.
ADA compliance service
Accommodation-request handling and medical-inquiry detection, flagging when a screening question strays toward disability-related territory it should not touch.
State-specific compliance rules
Configurable rules for state-level variation, including California, New York, and Illinois among others, since state requirements do not line up with each other.
GDPR consent and retention
Consent tracking, data-subject request handling, and configurable data-retention policies for candidates covered by GDPR.
US AI hiring law in 2026: what is actually in force
This is not legal advice, and Treegarden is not a law firm. It is a summary of what each law regulates, verified against primary sources, followed by how an advisory-first, human-gated, audit-logged design like Edera AI's relates to those obligations. Confirm applicability to your own hiring footprint with employment counsel.
NYC Local Law 144
Requires New York City employers using an "automated employment decision tool" to screen or score candidates to get an independent bias audit within the prior year, publish a summary of the audit results, and give candidates notice. Penalties run $500 to $1,500 per violation, with each day a violation continues counted separately. A December 2025 New York State Comptroller audit found the city's own enforcement had been weak through mid-2025, and the city has since committed to stricter enforcement.
Edera AI's screening and ranking toggles are logged and reviewable per job, which is the kind of record a bias audit and the required public summary depend on. That supports the record-keeping LL144 calls for. It does not substitute for commissioning your own independent bias audit, which the law requires you, the employer, to obtain.
Illinois HB3773 (Public Act 103-0804)
Amends the Illinois Human Rights Act to make it a civil rights violation for an employer to use AI that has the effect of discriminating on the basis of a protected class, or to use zip code as a proxy for a protected class, in recruitment, hiring, or related decisions. It also requires notice to employees and applicants when AI is used in those decisions.
The per-job AI switch and the toggle log give an Illinois employer a clear record of exactly which AI features were active for a given role, which is a reasonable starting point for the notice and documentation this law expects. Whether your specific notice language satisfies the Illinois Department of Human Rights' rules is a question for counsel, since those rules were still being finalized as this page was written.
Colorado SB26-189
Colorado's original 2024 AI Act (SB24-205) was delayed twice and then replaced before it ever took effect. SB26-189, signed May 14, 2026, reworks the framework around consumer notice at the point of interaction, disclosure to a candidate within 30 days of an adverse outcome, a right to request human review, and a minimum three-year record-retention duty, enforced by the Colorado Attorney General. Most provisions take effect August 12, 2026, with a separate developer documentation duty starting January 1, 2027.
A human review queue and a documented right to escalate a decision are close to what SB26-189 asks for by design, not because the platform was built for this specific statute. Colorado's framework changed twice in under two years, so treat this summary as a snapshot and confirm the current text before relying on it.
Colorado General Assembly: SB26-189 → · SB24-205 (superseded) →
EU AI Act (high-risk employment systems)
Recruitment, candidate evaluation, and worker-monitoring AI are classified as high-risk under Annex III, which brings mandatory risk assessment, technical documentation, human oversight, and logging obligations. The headline date for high-risk systems was originally August 2, 2026. The EU's AI Act simplification package, formally approved by the European Parliament and Council in June 2026, pushed the deadline specifically for employment-related high-risk systems to December 2, 2027.
If you hire in the EU, this applies regardless of the later date, since the obligations are still coming. Edera AI's human oversight, audit logging, and candidate-notice features overlap with what Annex III eventually requires, but Treegarden does not claim EU AI Act certification on this page, because no such certification exists yet for anyone.
Related reading: our EU AI Act recruitment compliance checklist and plain-language guide to GDPR Article 22 go deeper on the two AI-specific frameworks above. For the general US employer reporting picture rather than AI specifically, see our EEOC compliant ATS guide.
What this actually costs
Treegarden's base plans are flat and published. There are no per-employee fees on the ATS. The AI Recruiter runs as an add-on to any plan.
Annual billing runs about 20% lower than monthly across all three plans, and every plan includes the full ATS feature set with no per-employee surcharge. Edera AI is available as an add-on on top of any of the three plans above. Because add-on cost depends on company size and which of the nine toggles you turn on, we do not publish a flat add-on number here: it is confirmed on the demo call, the same way a vendor with per-user pricing would confirm your seat count before quoting. See the full pricing page for plan limits and the HR module add-on.
How AI recruiting pages compare
We checked five other AI recruiting pages in July 2026 against four questions a buyer actually needs answered. Cells marked "Not stated" mean we could not find that information on the vendor's own AI or ethics page, not that the vendor necessarily lacks the capability.
| What we checked | Treegarden | Ashby | Zoho Recruit | Paradox | Teamtailor | Manatal |
|---|---|---|---|---|---|---|
| Pricing shown on page | Yes, published | No | No | No | No | Yes, per user |
| Named AI regulations addressed | Yes, with dates and sources | Yes (EU AI Act, NY LL144, CO) | Not stated | Not stated | Not stated | Not stated |
| Human-in-the-loop stated | Yes, with a review queue | Yes, explicitly | Partial | Partial | Partial | Not stated |
| Per-feature AI toggles | Yes, 9 toggles + per-job | Yes, stated | Not stated | Not stated | Not stated | Not stated |
Checked directly against each vendor's own AI or ethics page in July 2026: ashbyhq.com/ai, zoho.com/recruit/ai-recruitment.html, paradox.ai/ethical-ai, teamtailor.com/en/ai-recruiting-software, and manatal.com/pricing. Ashby's page is genuinely strong on this list, it names specific regulations and states a human must be involved in ranking decisions. Vendor pages change; verify current claims directly before deciding.
Questions buyers actually ask
Straight answers, including the ones vendor pages tend to soften.
Does Treegarden's AI ever reject a candidate automatically?
The AI never rejects anyone on its own. Automatic actions - advancing or declining - happen only under rules an employer explicitly configures, using criteria and thresholds they set, with a human review queue available as a gate and every AI-influenced decision audit-logged.
Can we turn off individual AI features?
Yes. Edera AI ships as five independently switchable company-level toggles (candidate ranking, screening question generation, interview question generation, the candidate chat assistant, and pipeline auto-move), plus four further automations you can enable separately (intelligent auto-advancement, AI interview scheduling, decision suggestions, and offer-letter draft generation). On top of that, AI can be switched on for one job and off for another from the same company account, so a high-volume role and a confidential executive search can run under completely different rules.
What US laws apply to AI hiring tools in 2026?
The most relevant are New York City's Local Law 144, in force since July 2023, which requires an independent bias audit and public results for automated employment decision tools; Illinois's amendment to the Human Rights Act (Public Act 103-0804), in force since January 1, 2026, which bans AI that discriminates on protected classes and requires employee notice; and Colorado's SB26-189, which replaces the state's original 2024 AI law and takes effect August 12, 2026 with consumer notice and disclosure duties. The EU AI Act also applies if you hire in the EU, though its high-risk employment obligations were delayed from August 2026 to December 2027. None of this is legal advice. Talk to employment counsel about which laws apply to your hiring footprint.
Does AI decide who gets hired?
No. Hiring decisions are made by your recruiters and hiring managers. Edera AI ranks candidates against the criteria you set, drafts screening and interview questions, and can suggest a decision, but the final call sits with a person, and any automated advance or decline rule only runs inside limits your company configures, with the review queue able to intercept it first.
How much does the AI Recruiter cost?
Edera AI is an add-on to any Treegarden plan (Startup at $299, Growth at $499, or Scale at $899 per month, billed annually, about 20% lower than monthly billing). Add-on pricing depends on your company size and which capabilities you turn on, so it is confirmed on the demo call rather than published as a flat number.
Is there a free trial for the AI Recruiter?
No. Treegarden does not run a self-service free trial for the AI Recruiter or the wider platform. Instead, a guided demo walks through candidate ranking, the review queue, and the audit log using a job similar to one of yours, so you can see exactly what the AI does and does not do before committing.
Ask us the questions this page raised
Bring your own job requisition to the demo. We will show you exactly which toggles apply, what the review queue looks like from a recruiter's seat, and what the audit log records.
Last updated: 13 July 2026. Product details (AI toggles, review queue, audit logs, EEOC/FCRA/ADA services) are drawn from Treegarden's own AI Recruiter module. Pricing is drawn from the Treegarden pricing page. Law dates and requirements were checked directly against primary sources on 13 July 2026: NYC DCWP on Local Law 144, Illinois Public Act 103-0804, Colorado SB26-189, and the European Commission's AI Act timeline. Competitor claims were checked against each vendor's own public page on the same date. This page is not legal advice. Related guides: AI in recruitment: practical guide, how AI resume screening really works, how to audit AI hiring tools for bias, our pricing page, and the applicant tracking system Edera AI runs inside.