Project State: Harmony
Last updated: 2026-04-19 Status: Strategy crystallization complete — SDR-first wedge with brain-layer + agent orchestration as the differentiator. First external customer interview (Atera) validates crm-replacement-thesis and agent-in-call direction.
Overview
Harmony is a new product initiative within monday led as an internal startup (~20 people). Long-term vision: a revenue-platform that achieves global-maximum optimization across the entire customer lifecycle by giving agents (and humans) shared context across departmental boundaries.
Near-term: a self-learning-voice-agent for mid-market B2B teams running Salesforce or HubSpot, ideally with Gong. Sold first as the wedge; the global-maximum / revenue-platform story is the expansion narrative. Per DEC-015, voice is a feature — the brain-layer is the actual product.
Current Phase
- Product definition — converging on one primary wedge: self-learning revenue voice agent for mid-market B2B. See self-learning-voice-agent.
- SDR productization (priority) — engineering ready (acquired voice team has 3 years of battle-tested tech). Gap is product packaging: UX, onboarding, playbook builder, CRM polish, dashboards. AI-016 owners: vitali / ayala.
- Marketing data connectors (second vector) — marketing-connectors. Lightweight ad-attribution data feeding SDR. NOT a full marketing agent. AI-017 owners: tomer-friedman / niv.
- API-first architecture — api-first-architecture. AI-018 owners: saar-arbel / vitali.
- Data infrastructure — account-timeline (CDP) being built on connected sources via connector-architecture (Google Ads, Facebook Ads, Salesforce, Gong). Kremer (BigBrain) for cross-funnel queries.
- POC on monday data — using monday’s own marketing/sales data first. Plan to validate the brain on human workflows before requiring AI agents (DEC-019, AI-019).
- Campaign analysis — active work with tamar and lital on CRM Main campaign. See campaign-analysis.
- Customer research — first external interview complete (atera-2026-04-19). Validates crm-replacement-thesis and agent-in-call. Follow-up: AI-023 (Atera SDR/BDR intro).
- Next steps: productize SDR, build marketing connectors, validate cross-funnel value on humans, design API-first, bring 10 mid-market alpha design partners (AI-012), hire sales reps (2 → 5 by year-end, AI-008).
Team
See harmony-core for the full team page. Quick reference:
| Name | Page | Role |
|---|---|---|
| Vitali | vitali | GM, harmony.ai (voice stack architect) |
| Saar Arbel | saar-arbel | Head of Engineering |
| Nizan Shifman | nizan-shifman | Product / Strategy lead |
| Tomer Friedman | tomer-friedman | Engineering |
| Niv | niv | Engineering |
| Ben | ben | Head of Design |
| Ayala | ayala | Designer (under Ben) |
External / collaborators: nadav-greenberg (advisor), guy-kori (bizapps), tamar / lital (marketing-ops).
Key Constraints
- Timeline: SDR production by July 1, 2026 with paying customers (DEC-006).
- Capacity: ~14 engineers, 5-month window. Drove the lightweight scope on the second vector (DEC-016).
- Runway: Roi (CEO) wants product-market fit + efficiency; not unlimited.
- Distribution: Leverage monday’s brand and customer base, but operate independently. Cannot use monday branding directly.
- GTM: touch-onboarding. Sales reps + solution engineers.
- Competitive landscape: Crowded — see competitive-landscape. Differentiation through self-learning-loop + cross-funnel-intelligence + api-first-architecture.
- Messaging discipline: Avoid generic AI builder / voice API / CRM-replacement framing before the wedge is proven.
- Data access: Salesforce + Gong read API requested via guy-kori; Google Ads + Facebook Ads keys already in hand via marketing-ops.