Decagon
AI-first customer service — autonomous agent handling complex multi-turn support. Enterprise customers include Substack, Eventbrite, Rivian, Notion. Strong on deflection, integration depth, analytics.
Decagon raised a $131M Series C in late 2025 at a $1.5B valuation. Pricing follows an outcome model: roughly $1 per resolved conversation, with floor commitments. The platform layers a custom RAG stack over Anthropic and OpenAI models, swapping based on task. Native connectors for Zendesk, Intercom, Salesforce Service Cloud, and Front mean deployment in under three weeks for most teams. The “AI Operating System” dashboard exposes per-intent resolution rates and CSAT, letting ops teams retire underperforming flows weekly rather than quarterly.
Ada
Longtime leader in AI CX. Widely deployed across industries. Strong multilingual support, no-code bot-building tools, tight third-party CRM integrations (Salesforce, HubSpot, Zendesk).
Ada’s “Reasoning Engine 2” shipped Q1 2026, moving from intent-classification to LLM-driven planning over a knowledge graph of policies. Supports 50+ languages out of the box, which matters for global brands like AirAsia and Verizon. Pricing remains seat-and-volume hybrid: typically $50K-$200K ARR for mid-market, scaling to seven figures. Strongest fit when your CX team owns bot logic and wants no-code authoring rather than engineer-led prompts.
When Specialists Win
CX is primary use case, not CRM-layer. Existing CRM investment you don’t want to replace. Rapid deployment needs. Best-in-class AI chat/voice without building from scratch.
Three concrete signals: contact volume above 50K/month makes outcome pricing pencil out; you have a Zendesk or Intercom contract that won’t be displaced; your in-house team lacks LLM evaluation chops. Specialists also win when regulatory disclosure or escalation logic is complex — Decagon and Ada have audit-trail tooling that platform vendors are still building.
When Platforms Win
You’re already committed to a CRM platform and want one vendor. Tight integration with sales/service/marketing matters more than best-in-class CX. Budget discipline favoring consolidation.
Salesforce Agentforce and HubSpot Breeze ship at marginal cost on existing licenses. If your contact volume is under 10K/month, the integration tax of a third-party specialist outweighs quality gains. Procurement teams increasingly veto net-new SaaS contracts over $100K, pushing buyers toward bundled platform AI even when benchmarks favor specialists by 10-15 points on resolution accuracy.
Decision Criteria
Run a 30-day pilot on a single intent (returns, password reset, billing) with both a specialist and your platform’s native agent. Score on contained-resolution rate, CSAT delta, and engineer-hours per flow change. The winner is rarely ambiguous after 5,000 conversations.
What to Do This Week
Pull your top three intents by volume and request a Decagon or Ada proof-of-concept against your incumbent platform AI on those exact intents.
