Essays from the GTM engineering desk.
Practitioner writing on AI for GTM, Claude Code, outbound, data, and the GTM rebuilds we run for clients. Read the hubs first, then dig into a spoke.
-
Why every AI SDR pilot is dying
AI SDR vendors trained on the seller's value prop. Recipients pattern-matched it as spam. The category never had a methodology problem hiding inside its tool problem.
-
Claude Code for GTM teams: a practitioner's guide
The shape of a GTM Claude Code repo: where subagents live, how they read CRM data, how CLAUDE.md gates output. Plus the four most-shipped subagents in 2026.
-
AI for GTM in 2026: what shipped, what stalled
Three years into the AI GTM hype cycle, the receipts are in. AI SDRs collapsed. Generic copilots got ignored. The teams that actually moved pipeline shipped Claude Code subagents in their own repos.
-
Recipe: a Claude subagent ICP scorer for inbound
A working recipe for an ICP scorer subagent. Inputs from the CRM, prompt structure, output format, validation against closed-won DNA, deployment notes.
-
Why AI SDRs fail pipeline quality
Pipeline pollution is the failure mode the AI SDR category was built to produce. The math is unforgiving. Here is what works instead and why it scales.
-
Three subagent orchestration patterns for RevOps
Three orchestration patterns cover almost every RevOps subagent system we have shipped. The shape of each, when it wins, and the failure modes that come with it.
-
Anthropic Managed Agents vs self-hosted Claude Code
Managed Agents fits one shape of revenue team. Self-hosted Claude Code fits another. The decision tree, cost math, and how to migrate between them.
-
The outbound data pipeline we ship in Claude Code
Five stages, four subagents, one human gate. The outbound pipeline we deploy into revenue repos and what each step actually does.
-
Replacing BDR seats with subagents: the honest tradeoffs
Subagents replace BDR research, scoring, and drafting cleanly. They do not replace the human work of objection handling and trust. Where the line actually sits.
-
Claude Skills vs MCP vs traditional tools for GTM stacks
Skills, MCP, and traditional tool integrations all do different jobs. Most teams pick the wrong one and pay for it. A decision tree from the builds we shipped this year.
-
SOC 2 / HIPAA trust center buildout: a 60-day plan that unblocks enterprise deals
Enterprise buyers want a trust center URL before they'll book the demo. The 60-day buildout: audit gaps, ship the page, wire questionnaire automation. What the page actually has to contain.
-
ICP discovery from closed-won data: a 30-day rebuild for revenue teams
Your ICP was set in a board meeting 18 months ago. Nobody checked whether the companies you actually close look anything like that profile. The 30-day rebuild from real win data.
-
RFP response automation in 2026: how to win procurement without burning your team
A 200-question security questionnaire takes a sales engineer 20 hours. A Claude Code subagent does the first draft in 12 minutes. The setup that makes it work plus the failure modes to avoid.
-
Event-driven ABM in 2026: a playbook for trade-show pipelines that don't die in week three
Most event pipelines are dead by week three. The ones that work run a subagent chain on the badge-scan list before the booth even closes. Pre-event, day-of, and post-event playbook.
-
HubSpot Breeze deployment in 2026: what works, what breaks, and the gaps to fill
HubSpot shipped Breeze across every hub. Most teams paid for it and never deployed it. Here's the working sequence: audit, configure, fill the gaps with Claude Code, hand off.
-
Salesforce Agentforce review: where it works, where it stalls, and what to build instead
Salesforce is selling Agentforce to everyone. Most deployments stall at configuration. Here's where Agentforce earns its license, where it doesn't, and the cost math nobody put in the AE deck.
-
Triggered Value Asset vs Permissionless Value Proposition: why agents need a different framework
Jordan Crawford's PVP reshaped outbound for human writers. TVA borrows the spirit and adds an agent-native test: would the recipient forward this. Why the shift matters.
-
The PE portco GTM rebuild: a 90-day playbook
Post-close, the GTM stack is always worse than the LOI claimed. The order of operations for the first 90 days, what to kill in week one, and where AI agents fit.
-
Migrating off ZoomInfo in 2026: a build-vs-buy playbook
ZoomInfo costs $100K-$300K and ages out of date the moment you license it. The build-or-replace decision math, plus the migration sequence that doesn't blow up reps mid-quarter.
-
Forward rate: the outbound metric that exposes every AI SDR
Why open rate and reply rate stopped meaning anything once AI lifted both ceilings. What forward rate is, how to measure it, and what good looks like.
-
How to write a CLAUDE.md for your revenue team
A working template for a revenue-team CLAUDE.md. ICP block, voice rules, banned phrases, signal scoring rubric, and the data the subagents are allowed to touch.