AI HR Pilot: shipping the AI-native HR chatbot that actually fits a 50 to 500 person company.
Concept to production in under 90 days. Live at aihrpilot.com under the Portfolio Leverage Company umbrella, deployed into active client organizations and serving as the reference implementation HRTS uses when clients ask for custom AI-native HR tooling.
The problem
Every HR AI vendor on the market is priced for the enterprise. Workativ starts at $349 a month. Leena AI and Moveworks quote six figures. Below that tier the SMB and lower mid-market are stuck with ChatGPT tabs and half-configured HRIS bots: no policy triage, no manager coaching layer, no compliance guardrails and no way to prove ROI to a Board.
The underlying bet: 3x CHRO authority plus working AI engineering means HRTS can ship a better product than the incumbents at one quarter the price, and use it as both a standalone SaaS and a deployable asset inside consulting engagements.
What we built
- Multi-LLM backbone: Claude, GPT and Gemini behind a unified prompt layer so the product is portable across providers and pricing regimes.
- 30+ HR policy domains with verified answer sets spanning PTO, benefits, payroll, performance, compliance, onboarding, offboarding, leaves and accommodations.
- Manager coaching layer: context-aware guidance for 1:1 prep, difficult conversations, performance conversations and comp conversations, trained on 2,300+ executive coaching engagements.
- Ticket classification and routing with employment-law-aware triage: HR vs legal vs manager vs self-service.
- Custom domain deployment under the client company's brand and subdomain, SSO-ready, audit-logged for compliance.
- Tiered pricing: $99 / $399 / $999 per month: the lower tier is designed to beat ChatGPT Team + manual policy Q&A on TCO.
Why this matters for HRTS clients
When a client asks HRTS to build them a custom AI-native HR agent, we're not starting from scratch and we're not outsourcing the build. AI HR Pilot is the reference architecture. The moat is real: we ship, we iterate in production, and we know exactly where the failure modes are.