Roadmap

Last updated: 2026-04-15

Current Milestone: Execution Layer / SDR Agent MVP (Target: July 1, 2026)

  • Ship production SDR/call-center voice agent with paying customers
  • Productization focus (AI-016): UX, onboarding flow, playbook builder, CRM integration polish, customer-facing dashboards. Engineering is ready; product packaging is the gap.
  • Lock the first ICP to mid-market B2B teams with inbound + outbound motion, meaningful call volume, Salesforce/HubSpot, and ideally Gong
  • Self-learning capability (closed-loop feedback from CRM outcomes)
  • Customer-facing proof that the agent improves over time (playbook versions, objection tracking, qualification rate, meeting rate, downstream pipeline quality)
  • Voice quality: 70ms response time (vs competitor 800ms)
  • Lower cost than 11Labs-based competitors
  • CRM integration (Salesforce sync) for data and feedback loop
  • Landing page and sales narrative centered on “self-learning revenue voice agent”
  • Touch onboarding: solution engineer helps customer set up (hours/days, not months — Vision session confirmed ~2 days vs competitor ~1.5 months)
  • Marketing data connectors (DEC-016, AI-017): Build lightweight connectors to feed source attribution data into SDR conversations, proving Harmony’s cross-funnel value
  • API-first architecture (DEC-018, AI-018): Design all capabilities as API-accessible
  • Validate on human workflows (DEC-019, AI-019): Test cross-funnel insights with monday.com’s human SDRs before requiring AI agents
  • Support bring-your-own-agent and create-custom-agent (DEC-017)
  • Related decisions: DEC-005, DEC-006, DEC-007, DEC-010, DEC-011, DEC-015, DEC-016, DEC-017, DEC-018, DEC-019

Near-Term (Q3 2026): Learning Layer

  • Design partners for Global Maximum: Bring 10 mid-market alpha customers; connect via connectors to their ad platforms, CRM, and Gong; build account timeline/CDP for them (AI-012)
  • Sales team build-out: Hire 2 initial sales reps, grow to 5 by year-end (AI-008)
  • Campaign analysis product: Monthly cross-funnel reports connecting marketing spend to sales outcomes (DEC-003, AI-007)
  • Expand SDR depth: From basic qualification toward demo delivery and deeper sales conversations (OQ-006)
  • Learning loop proof surfaces: Show win/loss feedback, objection analysis, and before/after performance by playbook version to customers and prospects
  • Cross-funnel optimization hooks: Feed CRM + Gong outcomes back into SDR guidance so learning is visible and repeatable

Mid-Term (Q4 2026 - Q1 2027): Revenue Memory Layer

  • Marketing agent: AI agent for campaign optimization (ad copy, audience targeting, budget allocation)
  • Local Global Maximum: Cross-department insights between SDR agent and CRM (e.g., SDR improves based on closed-won/closed-lost data)
  • Revenue Intelligence: Insights from Gong calls fed back to marketing and product teams
  • Account Timeline product: Standalone CDP that connects all revenue tools and shows full customer lifecycle
  • Revenue memory UI: Productize the account timeline as the shared memory layer between agents and humans

Long-Term Vision (2027+)

  • Full Global Maximum: All revenue departments (marketing, sales, support, product) connected through shared intelligence layer
  • Agent-to-agent communication: Agents autonomously share insights and optimize across the funnel
  • Deep SDR evolution: Agent that can do full sales cycle — discovery, demo, negotiation, closing
  • Potential CRM replacement: For SMBs, become the primary system of record; for enterprise, remain the intelligence layer on top of existing CRM
  • Buyer agent: Speculative — agent that represents the buyer side in procurement