About PilotCall
PilotCall comes from nearly a decade in sales and sales enablement, thousands of AI evaluation tasks, and public software work across RF machine learning, local LLM systems, automation, and security research.
Who is building this
Before PilotCall, Trevor worked through the day-to-day reality of outbound sales, account research, sales enablement, HubSpot and Salesforce process work, marketing-service discovery, and technical B2B prospecting. In parallel, he built public technical projects around AI evaluation, radio-signal machine learning, edge inference, local model benchmarking, and local automation.
Why PilotCall exists
Reps do not usually lose because the answer is impossible. They lose because the answer is buried in a CRM note, inventory system, SEO audit, pricing sheet, playbook, or previous conversation while the buyer is still talking. PilotCall is built for that live moment.
Trevor has worked across business development, sales enablement, account research, HubSpot/Salesforce workflows, outbound sequencing, voicemail/nurture logic, and technical buyer discovery.
He reviewed more than 5,000 LLM tasks and prompt-response datasets, building judgment around factuality, instruction following, completeness, ambiguity, and failure patterns.
His public technical work includes RTL-ML, local LLM benchmarking on edge hardware, an RF monitoring cyberdeck, Linux troubleshooting, and local agent infrastructure.
unland.dev frames his work around security research, AI training, red-team thinking, prompt engineering, hardware security, and local inference systems.
Public work
PilotCall sits alongside Trevor's broader work on AI evaluation, local inference, edge systems, RF machine learning, and human-controlled agents. That matters because live-call guidance needs both sales empathy and systems judgment.
Founder-built evaluation
PilotCall is best evaluated against a live-call workflow where reps lose time searching systems, handling objections, or translating playbooks into action.