The tools in this category
Anthropic's frontier LLM with strong instruction following.
OpenAI's LLM with the widest ecosystem and integrations.
Google's LLM with deep Google Workspace integration.
Search-first AI with citations.
Anthropic's protocol for connecting AI to real tools and data.
AI built into Notion for knowledge base tasks.
Our take on AI models & MCP
AI models are commodity infrastructure in 2026 — but which one you build on matters. Claude wins for writing, reasoning, and following instructions. ChatGPT wins for the broadest ecosystem. Gemini wins for Google-integrated workflows. Perplexity wins for research. And MCP (Model Context Protocol) — Anthropic's standard for connecting AI to real tools — is the layer that turns AI from a chat interface into an operational tool.
For most GTM teams: Claude as the primary model, Perplexity for research, and MCP servers to connect them to your real stack. That combination is what actually makes AI useful in production, not just impressive in demos.
For teams already in the Google Workspace or Notion ecosystem, Gemini and Notion AI are the in-app AIs you're probably under-adopting. Use them there.
How we'd stack these tools together
Claude is the reasoning engine. Perplexity is the research layer. MCP is the connection to the rest of the stack — CRM, data, communication tools. Pair with Claude Code for building AI-in-the-loop automations and workflow automation for glue.