The Signl Blog
Playbooks, teardowns and market notes on agent-native prospecting: MCP servers, buying signals and GTM engineering.
ChatGPT for B2B prospecting: a setup that works
A prospecting workflow that runs inside ChatGPT with live company data, the exact prompts to use, and the failure modes nobody warns you about.
PlaybooksGDPR and B2B prospecting in the UK and EU
No, B2B is not exempt. What the ICO actually says about corporate subscribers, why sole traders are different, and where the usual advice is simply wrong.
PlaybooksLive query vs static database: which data model wins
A static database is a snapshot with a decay curve. A live query is a request with a latency cost. The real trade-off, stated without the vendor spin.
MCPMCP clients compared for GTM work
Claude, ChatGPT, Cursor and VS Code compared on what each actually supports: primitives, auth specs, transports and the result limits almost none of them publish.
MCPMCP for sales: how AI agents get live GTM data
MCP lets an AI agent query live company data, buying signals and verified contacts mid-conversation. How it works, what breaks, and how to evaluate a server.
MCPMCP security for revenue teams
OAuth versus pasted API keys, what a connector can actually reach, how to revoke one client, and the questions to ask a data vendor before connecting.
MCPWhy your prospecting agent silently drops results
MCP tool results are capped at about 150,000 characters in Claude and 25,000 tokens in Claude Code. What that means when you ask an agent for 100 companies.
MCPMCP vs API integration for GTM teams
An API needs someone to write the integration. An MCP server describes itself so the agent calls it directly. When each wins, and where Zapier fits.