Harmony

Overview

Self-improving voice agent for call centers, built inside monday by nizan-shifman’s team (~16 people as of 2026-05-06). Targets the call-center labor-replacement market, sized at roughly 1.5B ARR is currently captured by incumbents — Nizan’s framing of it as still “early stage.”

Relationships

  • Internal product of monday.
  • Heavy daily user of saar-arbel’s Brain prototype.
  • Competes adjacently with sierra in call-center territory.
  • Built primarily on cursor + Brain + NoteTaker MCP.

Notes

  • The product Nizan keeps citing in the meeting as proof the Brain is already painkiller-grade. Harmony’s whole landing page, pricing strategy, and competitive SWOT were generated from a brand-new Cursor agent that used only the Brain context — without manually attaching docs or transcripts.
  • Generates competitive analyses by pulling categories (CRM-native incumbents, voice-platform competitors, AI-GTM peers like Alta, agent-native challengers) from accumulated brain context across many meetings — used as evidence for context-os-brain.
  • 2026-06-13: also used as evidence for revenue-cycle-brain. The team described Harmony/Monday experiments that push downstream call and customer-lifecycle insight back into acquisition and messaging.
  • 2026-06-14: Nizan added origin context: the global-maximum / revenue-cycle direction was close to Harmony’s original exploration. Saar and Tom built a four-day Monday MVP that acquisition managers found valuable, but the team pivoted toward voice/call centers after the One AI / Vitaly path and Monday’s distribution made that route more attractive.

Mentioned in