Reasoning Engine
Powered by Anthropic's Claude — reasoning, classification, tool use, web search, and multi-turn conversation context.
The intelligence behind every agent
Custom Agents is powered by Claude by Anthropic (Sonnet 4) — one of the most capable large language models available. This isn't a chatbot. It's a reasoning engine that understands context, follows complex instructions, and takes action.
What the reasoning engine can do
- Understand nuanced email context and intent — the difference between a complaint, a question, and a compliment
- Generate professional, on-brand responses — matching your specified tone and style
- Use tools and APIs — schedule meetings, look up order status, search the web
- Maintain multi-turn conversation context — remembering the full history of a thread
- Follow complex behavioral instructions — your agent's instructions are its operating manual
- Reason through multi-step problems — verify an order, check the return window, look up customer history, decide on a resolution, and respond — all in one step
AI model management
You never need to think about AI model versions, prompt engineering, or API management. Custom Agents handles:
- Model selection and optimization
- Automatic upgrades as models improve
- Token management and rate limiting
- Response quality monitoring
Features
Everything your agent comes equipped with — from a reasoning engine powered by Claude to a built-in CRM, native integrations, and configurable autonomy.
Email Identity
Each agent gets a real email address that it owns. Webhook-native architecture with sub-second delivery. Custom domains supported.