[Blog category]
Discover AI-driven methods to boost your sales performance.
68 articles · Page 1 of 3
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
B2B data breaks into 8 distinct layers, each with specialized providers: TAM mapping (Apollo.io, AI Ark, DiscoLike), lookalikes (CompanyEnrich), local business data (Openmart), people enrichment waterfalls (FullEnrich, Prospeo, BlitzAPI), buying signals (PredictLeads, TheirStack), technographics (BuiltWith, TheirStack, Sumble), and live web discovery (Exa, Limadata, Serper). Missing a layer creates a specific blind spot, from undercounted markets to unreachable contacts to late timing. The eighth layer is orchestration: the ColdIQ API Marketplace integrates 40+ providers under one API key so teams can waterfall across every layer and swap providers without code changes.
Michel Lieben
Compare 12 AI prospecting data APIs for 2026 on MCP support, per-request pricing and agent-readiness, with live-verified costs and real corporate risks.
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
The best GTM teams are moving off tools and interfaces onto repos, terminals, and APIs, running full campaigns from a single plain-language instruction. AI agents fail at GTM without context: skills, API keys for 8 to 12 tools, internal SOPs, and past campaign data. The real shift is replacing the operator who stitches tools together with an API that encodes that knowledge once and runs it on demand.
Michel Lieben
Discover the top relevance AI competitors to watch in 2026 that can boost tech sales growth and shape the future of the industry.
Michel Lieben
A GTM or RevOps engineer builds a connected system rather than buying a single platform, and that system breaks into five layers. The signals layer (PredictLeads, CompanyEnrich, Attention) detects which accounts are in market. The data layer (Openmart, Prospeo, FullEnrich) turns those accounts into verified contacts. The action layer (Instantly.ai, lemlist) runs the outreach across channels, the system of record (HubSpot, cobl) tracks every touchpoint and generates proposals, and the revenue layer (Hyperline) handles billing so pipeline becomes recognized cash. The reason it all works is that every tool exposes an API, and Claude Code connects the layers by reading the docs, handling errors, and retrying until the workflow is live. The advantage is not owning the tools, since many teams share the same stack, but wiring signals, data, and execution into one system. After serving 300+ companies and testing 1,500+ tools, this is the architecture we keep returning to.
Michel Lieben
Discover what clay pricing in 2026 means for tech companies today with easy tips from my experience to help you plan smart and save money.
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
Discover nooks pricing in 2026 and how it can save your business money. I’ve seen how right plans boost growth and simplify budgeting.
Michel Lieben
PhoneBurner vs Nooks for agencies in 2026: I tested both and found which tool helps get more leads fast and easy for your team.
Michel Lieben
Eleven GTM tools stand out as worth integrating with AI agents like Claude Code, split across three categories: data feeds that power agents (CompanyEnrich, LeadsFactory, Apollo, Prospeo, PredictLeads), platforms for building agents (Openmart, Nooks, Relevance AI, lemlist), and orchestrators that do both (Clay, Instantly). Each was selected based on API quality, MCP support, coverage, and how cleanly an agent can drive it end-to-end without a human in the loop. The practical playbook stacks them into four layers (data, signals, orchestration, execution), where an agent typically needs one tool per layer rather than all eleven.
Michel Lieben
Expandi analyzed over 100,000 LinkedIn outreach conversations and found seven repeatable patterns behind the highest-performing campaigns. The biggest lifts come from personalization, three-step sequences, and warm connection requests sent right after a profile view or post like. Short messages, conversational copy, multi-channel follow-up, and references to recent activity round out the playbook. The four-tool stack behind it is Expandi for automation and sequencing, Clay for enrichment and intent signals, LinkedIn Sales Navigator for list building, and Zapier for workflow glue. Top performers do not send more messages. They layer the patterns together so the lifts compound.
Michel Lieben
We turned 24 ColdIQ employees into LinkedIn influencers using a five-part playbook: a quarterly internal challenge with cash prizes, full-time designers and content tools to remove friction, 1:1 and group coaching with external LinkedIn experts, hiring people who already wanted to build their brand, and founder-led example setting with the CEO publishing 60+ posts per quarter himself. In under 90 days, the team published 581 posts, gained 34,023 new followers, and added $151,000 in MRR plus 27 new clients. The combined team audience now exceeds 250,000 followers with 15 teammates past 5,000.
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
LinkedIn content got ColdIQ past $550K MRR by feeding every post into a 7-step automated workflow. The system captures engagement, enriches contacts via Clay, qualifies with AI models, deduplicates against the CRM, runs waterfall enrichment for 85%+ email coverage, and sends contextual outreach refe
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
Discover what are the best AI tools for sales GTM in 2026. Explore top categories like workflow builders, sourcing agents, and video tools to scale revenue.
Michel Lieben
