[Blog category]
How teams use Claude Code, agents and new AI tools to run GTM.
21 articles
ColdIQ runs the operations of a $7M ARR agency on Claap and Claude Code connected over MCP, and the unlock is that Claude queries labelled call data instead of reading raw transcripts. That structure turns objections, client reviews, delivery handoffs, and call coaching into single queries, and the whole setup took one terminal command and an OAuth click.
Soheil Saeidmehr
Trigify founder Max Mitcham runs his marketing with three OpenClaw agents: Bella orchestrates strategy, Rex executes marketing, and Taz manages product, with a daily 9am standup where they review sales calls and metrics to set priorities. The agents research competitors through the Claude Code Chrome extension, write SEO blog posts from trained skills, monitor performance via Trigify's API, and rewrite their own skills based on what works. In one week the AI-written blogs generated 40 inbound leads, and the full 61-minute breakdown shows how the system was built.
Michel Lieben
Anthropic's marketing team runs four Claude Code workflows that show what a modern GTM team can build. An agentic copy workflow reads hundreds of ads with their performance data, flags the weak ones, and rewrites them inside platform character limits, cutting ad copy creation from 2 hours to 15 minutes. A Figma plugin generates up to 100x ad variations by swapping headlines and descriptions across frames, driving a 10x increase in creative output. An MCP server plugged into the Meta Ads API lets marketers query live campaign performance from Claude Desktop, and a memory system logs hypotheses and test results so every new round of creative learns from the last. The through-line is that tasks that used to need engineers are now handled by marketers, and these four are only the starting point.
Michel Lieben.jpeg&w=3840&q=75)
The Vibe Marketer is an emerging GTM role set to earn more than $300,000 a year by 2027, defined by building the software companies used to buy rather than just running campaigns. Because AI now handles the coding, one operator can design and publish branded content, launch outbound and ad campaigns, build internal CRMs and dashboards, and automate operations across the org, all bound by imagination instead of syntax. The barrier to entry is taste rather than a technical background, and the economics work because a single Vibe Marketer replaces both a slice of the software budget and several specialist hires. It is the natural next step after the Content Engineer, a role defined by what you can build.
Michel Lieben
The GTM stack is splitting into three layers as Anthropic's revenue compounds toward a reported $30B run-rate. Legacy platforms like Salesforce, Zapier, and ZoomInfo cost upwards of $8,000 per month and win on maturity and reliability, but ship slowly and need a full team to operate. AI-native tools like Clay, Cursor, n8n, and Artisan run closer to $3,000 per month, handle around 80% of the work, and deliver 10x output for skilled operators, though the subscriptions are scattered and workflows still break. Claude's own layer (Code, Routines, Managed Agents, Cowork) sits near $200 per month plus tokens, collapses the stack into one interface, and keeps improving, but concentrates control and pricing inside a single vendor. The same consolidation that lowers cost creates a Claude dependence risk worth planning for, and the three layers will coexist for now depending on what you optimize for.
Michel Lieben
Thought Leader, the content agency founded by Naga, runs almost entirely on a content orchestration engine built in Claude Code, connected to Notion and Apify, and the system generated 8.4 million X impressions for Michel Lieben in under six months. The orchestrator is a content agent with a six-part memory layer (client profiles, content library, drafts memory, idea bank, sentinel intel, and a consolidated cache), platform-specific writing skills with voice profiles, and an algorithm scorer forked from Twitter's open-source recommendation code. Guardrails like 45-day topic cooldowns, mandatory fact-checks, and AI-tell scans keep the output credible, while format decisions and idea selection stay human. Replicating the core loop takes one orchestrator, a Notion MCP connection, an Apify token, and meta-prompted writing skills.
Michel Lieben
Claude Code turns cold campaign building from a multi-hour, multi-tool slog into a single terminal session. Here is the seven-step process we use to go from a raw list to a ready-to-send campaign: list building, ICP scoring, intent signals, decision-maker filtering, waterfall enrichment, batched copy generation, and a post-launch feedback loop that lets campaigns auto-improve.
Michel Lieben
Inside the self-learning AI agent stack that replaces lead generation agencies. How Obsidian memory, Claude Code, and Trigger.dev power autonomous outbound.
Michel Lieben
AI agents are breaking seat-based SaaS pricing at a speed most vendors are not ready for. One agent can now perform the work of 1,000 seats, extracting value that the old pricing model never accounted for. Clay shifted to per-action pricing when agents started abusing its free orchestration layer. Usage-based API providers are quietly printing revenue because agents execute autonomously and blame any overspend on the framework rather than the provider. The next pricing model blends a platform fee, a usage fee tied to the actions agents consume, and an outcome fee that captures real value delivered. Not every SaaS should make the jump. Horizontal productivity tools and reporting layers will stay on seats or tiers, but any product where an AI agent directly consumes the unit of value will need to rebuild its pricing logic from scratch.
Michel Lieben
Claude Code can now run a full LinkedIn outbound campaign from one terminal, replacing the need for Clay tables, lemlist sequences, and Instantly campaigns. Othmane Khadri, founder of Earleads, shared his exact setup: a four layer operating system combining Claude.md rules, skills, memory files, and APIs like Unipile, Firecrawl, and Apify. The system scrapes post engagers, runs them through a 7 gate qualification pipeline, scores each lead, drafts outreach copy, and sends messages automatically. The feedback loop is what makes it powerful. After every campaign, the system refines its own ICP criteria based on what converted and what did not. This closes a loop Clay cannot. Claude Code still needs human taste for creative angles and offer direction, but for the mechanical parts of GTM execution, one terminal replaces the entire stack.
Michel Lieben
ColdIQ rebuilt its content engine inside Claude Code using 27 skills that cover every step from voice profile setup to scheduled delivery, and the system added $151K MRR in 87 days with 24 writers posting in their own voice. The architecture runs on a one-time foundation per person (voice profile, I
Michel Lieben
ColdIQ runs outbound campaigns for 70+ B2B clients entirely through Claude Code, using six APIs that cover every step from company sourcing to campaign launch. Kenny Damian built a 300 company campaign in under 20 minutes using AI Ark for the company database, PredictLeads for buying signals, BlitzA
Michel Lieben
Claude Code runs the full outbound campaign workflow from a single terminal session. It scores company lists against ICP criteria, finds verified emails through a waterfall of providers like Apollo and Prospeo, generates personalized email sequences based on proven templates, and uploads the finishe
Michel Lieben
I spent 300+ hours using Claude Code to run GTM activities including building mini-tools, enriching prospect data, running SEO analysis, and automating scraping workflows. The core stack costs under $500 per month and replaces work that would require multiple full-time hires. Claude Code orchestrate
Michel Lieben
ColdIQ manages over $300,000 per month in ad spend across Google, Meta, and LinkedIn Ads entirely from the terminal using custom Claude Code skills built by Head of Growth Ivan Falco. The 12 skills cover keyword auditing, negative keyword management, creative fatigue detection, audience building, bi
Michel Lieben
AI agents, agentic workflows, and non-agentic automation each serve different roles. Non-agentic workflows use AI as a content tool with humans driving every step. Agentic workflows automate repeatable sequences with triggers and fixed logic. AI agents pursue broader goals, adapt to changing context, and persist until objectives are met. The best teams match each approach to the right task.
Michel Lieben
Multi-agent AI setups come in three structures: hierarchical, human-in-the-loop, and sequential. The best GTM systems combine all three, using sequential chains for enrichment, human checkpoints for sensitive communications, and hierarchical managers for complex campaigns.
Michel Lieben
Four agents built in n8n automate lead research with ICP scoring, local business enrichment via Apify and Google Maps, content generation from a single brief to carousel-ready drafts, and Reddit intelligence sourcing that delivers daily pain point reports.
Michel Lieben
AI transformed content marketing from manual writing to automated systems. Content marketers now split into Content Engineers (building AI workflows) and Content Strategists (creating original insights). Winning teams combine both, systems handle volume, humans handle depth.
Michel Lieben
AI transformed the SDR role in 2025 by enabling two distinct career paths: human-focused relationship builders who excel at cold calling and complex conversations, and GTM engineers who build automated prospecting systems using AI tools. The highest-performing teams combine both specialists rather than relying on generalist SDRs doing everything.
Michel Lieben
European AI companies are building highly practical agents for GTM teams in 2026. From CRM automation and web scraping to outbound personalization, voice qualification, and workflow orchestration, these tools remove manual sales and marketing work. Rather than chasing hype, Europe is quietly building the execution layer for AI-powered revenue teams.
Michel Lieben
