A few days after last week’s Warm Intro Guide went out, replies started coming in.

Most were variations of the same question:

• “What about enrichment? “

• ”What about contacts who are not in my LinkedIn connections? “

• ”What happens when someone changes jobs and I miss the window?”

Fair questions. The network audit maps what you have. The cross-connector matching finds warm paths through people you already know.

But two specific signals - work overlap and job changes - sit completely outside what any LinkedIn export and prompt combination can produce.

Today we are covering those gaps.

Gap 1: Work Overlap

Former colleagues who worked at the same company during overlapping years sit in a relationship sweet spot: Close enough to recognise your name and respond. Distant enough to have their own networks and connections you cannot reach through your immediate circle.

That relationship signal is the strongest one available in second-degree prospecting.
And it is completely invisible in a standard LinkedIn export. You cannot calculate it from a connections list. You need employment history data that LinkedIn does not give you.

A mutual LinkedIn connection tells you two people are connected, but nothing about whether they actually worked together, for how long, or in what context. Work overlap is a fundamentally different signal.

Gap 2: Job Change Tracking

When a past contact moves to a new company, there is roughly a 30-day window where they are far more receptive than usual:

• New role, fresh mandate

• Actively evaluating what they need

• Not yet locked into existing vendors or processes

After that window closes, the signal fades.

A manual system has no way to monitor for these changes continuously. You would have to check LinkedIn manually across hundreds of contacts, which nobody does.

The Swarm: Filling Both Gaps

at MoxieGTM pointed out that The Swarm shortcuts a lot of the manual work.

The Swarm is a relationship intelligence platform that maps your professional network beyond what LinkedIn shows you. Former colleagues, work overlaps, second-degree connections your LinkedIn export completely misses.

For this system specifically it fills the two gaps:

• Work overlap detection across 100M plus profiles, surfacing former colleagues based on shared employment history

• Daily job change monitoring so you catch the 30-day window every time a past contact moves to a new role

It also comes with:

• MCP integration to query your relationship graph directly from Claude

• Free Chrome extension showing intro paths and work overlap on any LinkedIn profile

The Swarm gives you an algorithmic version of the S/W/X check; the framework for checking how well your connector actually knows a prospect before asking for an intro: S is strong, W is warm, X is weak.

Its connection strength score automates most of this based on overlap duration and shared context, before you send a single message.

The Swarm makes Layers 1 and 2 of the system stronger. The strategy, the plays, the referral activation: that is the guide.

For founders running this regularly, the $99/month is worth it. For a first run, start with the free Chrome extension.

The Confidence Gap

There is a reason some connectors refer you readily and others, despite liking you and wanting to help, never quite get around to it.

Justin Vajko (via ’s 10x Solo) calls this the Confidence Gap.

People will not stake their reputation on someone whose work they have not experienced firsthand.

Your past clients refer you without being asked because they have seen the output. Your connectors often do not, even when they trust you personally. They are being asked to vouch for someone based on a guess.

The fix: before expecting referrals from a connector, give them a direct experience of your work.

A free 15 to 30 minute slice of what you actually do. That could be:

• A short session helping them think through a problem they are currently stuck on

• A look at something they are building and honest feedback on what is working and what is not

• A conversation where you bring one specific insight relevant to their business right now

After experiencing how you think firsthand, a connector’s referrals carry conviction rather than goodwill.

That distinction is the difference between a prospect who takes the meeting and one who politely ignores the intro.

Where this sits in the system: between the S/W/X check and the actual intro ask. If a connector rates a prospect as S (strong relationship) but has never seen your work, invest that 30 minutes first. The quality of what follows will be noticeably different.

Takeaway

Running this seriously over time exposes the same three limits for almost every founder who tries it.

• You find a warm path to someone who used to work at the same company as your connector, and you have no way to know that from a LinkedIn export.

• You miss a past contact’s job change by three weeks and the window is gone.

• You ask a connector for an intro, they say yes, and the prospect never responds.

The additions in this issue close all three. Which brings us to two things worth doing before this edition leaves your inbox:

1. Install The Swarm’s free Chrome extension and browse five of your top prospects on LinkedIn. See what it tells you that Sales Navigator does not.

2. Think of one connector who rates a key prospect as S but has never experienced your work. Spend 30 minutes with them before you ask for anything.

The warm path was always there. The question is how much of it you can actually see.

Jasper

PS: Want the complete warm intro system that keeps the whole thing automatically updated as your network grows?

The Warm GTM System Guide walks through every layer in full:

• The full network audit prompt with composite scoring across LinkedIn, CRM, calendar, Gmail, Google Drive, and phone contacts

• The Swarm as an optional accelerator: when it is worth it and how to connect it to Claude via MCP

• The referral activation brief builder and all re-engagement plays

• The Claude Code folder structure and CLAUDE.md with single-command workflows

Get instant access here.

Last week, I sat down with Claude and my full LinkedIn data export, just to finally see what my network actually looked like.

One hour later I had an entire map of every warm path to every ICP-fit prospect I could reach through just six people I already trusted.

The data baffled me:

• 1,382 founders and CEOs I had never spoken to, one introduction away from someone I know well.

• 25 prospects connected to two or more of those six people, meaning multiple independent warm paths into the same conversation.

• 5 past clients who had publicly endorsed me on LinkedIn and who I had not contacted in months.

The network was never small. The problem was that I could not hold it in my head.

Read more

A few months back, a founding team asked me for feedback on their email campaigns.

Six replies from ~ 2000 emails. They had already rewritten the subject line three times, swapped the hook, tried a shorter version, tried a longer version, and run the whole thing through Claude twice to make it sound less salesy.

They thought the copy was the problem. I asked to see the list first.

The accounts were fine on paper: right industry, right company size, right job titles. But there was nothing behind any of them. No signal that any of those companies were actively dealing with the problem they solve.

They would pull filtered lists from Sales Navigator and start sending. The copy was not the problem.

Why the Copy Gets the Blame

Copy is visible and easy to change. You can open a doc and rewrite a subject line in ten minutes, or add some fluffy personalisation.

Rebuilding the logic of your targeting however, finding better signals, questioning whether you are actually reaching the right accounts at …

Read more

How come Clay is valued at $5B... when Claude Code exists?

Tim Soulo did not ask that question by accident.

When the CMO of Ahrefs publicly wonders, he is not making a product comparison, but pointing at something structural: a company whose entire value proposition sits between you and data you could access directly.

Rather than a $5B moat, that is a $5B question mark.

And last week, about 100 people had something to say about it on LinkedIn.

The debate in the comments was fascinating: founders, GTM engineers, and operators all weighing in on Clay vs Claude Code.

But underneath all of it, one comment from Alex Bojovich (Ramp) said it all:

“Clay helped train an entire market to value enrichment data. That market is now about to buy it without Clay in the middle.”

What Clay Actually Built

Clay is not a data company. It never was.

What Clay built is a toll road. You pay to pass through it, and on the other side sits data that belongs to someone else entirely. Clearbit, Apollo, Hunter, ZoomInfo, and a dozen other providers own the actual information.

Clay owns the infrastructure that connects them and makes them usable together.

Infrastructure that is genuinely useful:

• Waterfall enrichment: query multiple providers in sequence until you get a verified result

• Deduplication: clean and merge data from different sources automatically

• Workflow logic: trigger actions based on enrichment results without writing code

• Routing: push enriched contacts directly into your outreach tools

For teams who needed all of that without an engineering team, Clay has been the answer. It turned a complex multi-vendor data problem into a visual interface anyone could learn in a week.

But here is the distinction that changes everything:

Clay owns the workflow. The data vendors own the data. You pay for both, bundled together into one credit system that obscures exactly how much each part actually costs.

The moment those two things separate, Clay’s position in your stack becomes a choice rather than a necessity.

And MCP is what separates them.

The Tell Hiding in Plain Sight

Clay made three moves last quarter that most people noticed separately. Put them together and they tell a clearer story.

• Cut data costs by 60%

• Removed AI markup

• Launched MCP support

Each one on its own looks like a product update. Together, they look like a company that can see a threat coming and is repositioning before it arrives.

Think about what each move actually signals:

• The price cut tells you Clay knows their credit margin is vulnerable. If teams can access enrichment data directly through MCP, paying a markup to Clay becomes optional. Getting ahead of that by cutting prices first is smart. It is also defensive.

• Removing the AI markup tells you the same thing from a different angle. Charging a premium for AI features only works as long as the AI layer is hard to replicate. Claude Code running inside your own terminal, connected to the same data sources, removes that premium entirely.

• MCP support is the most interesting move of the three. Clay is essentially building the bridge that makes it easier to route around them. That is either visionary or desperate, depending on how well they execute.

Scott Martinis called it: "Plan for a world where no one ever logs into your app."

Clay is not ignoring the shift. The question is whether they can move fast enough to stay relevant on the other side of it.

Companies that are confident in their moat do not make three defensive moves in the same quarter.

How the Middleman Gets Cut Out

This is not a prediction, but a sequence of events that is already in motion.

Step 1: MCP turns every data vendor into a direct connection

Apollo just launched their MCP server. Crunchbase has one. Brave Search, Serper, Hunter, and a growing list of providers are either live or building.

Inside Claude Code, you connect to them with a single terminal command and query them directly in plain English. No Clay table or credit system. Neither a UI to log into.

Step 2: Data vendors do the maths

Every time a founder queries Apollo through Clay, Clay takes a cut. The vendor gets less, the user pays more, and Clay sits in the middle collecting the margin.

The moment vendors can sell directly to the agent through MCP, that margin disappears and goes back to both sides. Higher revenue for the vendor. Lower cost for the user. The only loser is the orchestration layer.

Step 3: The “add a new vendor” advantage evaporates

One of Clay’s strongest arguments is convenience. Adding a new enrichment provider takes minutes inside their interface. But connecting a new MCP server inside Claude Code also takes minutes.

When the ease-of-use gap closes, the only remaining argument for Clay is the credit system itself. Which is not much of an argument.

Step 4: The Claybros become a liability

Clay’s creator ecosystem is one of its most valuable assets. Consultants and agencies built entire businesses teaching people how to use Clay. But that ecosystem also means Clay’s complexity became a feature rather than a bug.

A tool that requires specialists to run is a tool with an expiration date. When the alternative is plain English instructions in a terminal, the specialist tax disappears.

The irony is almost perfect: Clay spent years training an entire market to value enrichment data. That market is now about to buy it without Clay in the middle.

What This Means for Your Stack Today

This is not a “Clay is dead” take. It is a “know what you are actually paying for” take.

Clay still makes sense in specific situations:

• Large teams that need a UI anyone can pick up when people leave

• Enterprise environments where getting a new MCP integration past a security review takes months

• High-volume operations running scheduled enrichment across tens of thousands of records

• Non-technical teams where the visual workflow builder is the difference between using the system and not using it at all

For everyone else, the honest stack looks like this:

• Claude Code for research and targeting: finding companies that match your ICP, identifying buying signals, pulling structured data from public sources. Faster, cheaper, and more precise than Clay for this specific job.

• Direct MCP connections for enrichment: Apollo MCP is already live. Brave Search and Serper handle web research. You connect once, query directly, skip the credit markup.

• Clay only where it is genuinely irreplaceable: waterfall enrichment at scale, complex routing logic, or anywhere your team needs a visual interface to function.

The stack that made sense 18 months ago was Clay for everything.

The stack that makes sense today is Claude Code for research, direct MCP for enrichment, and Clay only where nothing else fits.

Six months from now, that split will look even more pronounced.

The window to build your muscle in Claude Code before your competition does is right now. Not next quarter.

Takeaway

The GTM infrastructure era is ending. Not with a crash, but with a quiet series of API connections that route around the tools that used to be necessary.

Clay is not the villain in this story. They built something genuinely useful and they are smart enough to see what is coming. The three moves they made last quarter prove that.

But useful is not the same as irreplaceable. And the window between those two things is closing faster than most people realise.

A few things worth remembering:

• The data vendors have no loyalty to the orchestration layer. They will go direct the moment it is economically rational to do so.

• The moat that compounds is the playbook, not the tool. Knowing which signal predicts a deal in your specific market is earned knowledge. It cannot be bought from a data vendor or copied from a GitHub repo.

• The early movers have a real window. Founding teams building their Claude Code muscle now will have systems compounding by the time the rest of the market catches up.

The question is not whether this shift happens, but whether you build before or after your competition does.

Jasper

PS: Want to skip Clay for lead enrichment?

The Claude Code Lead Enrichment Guide shows you exactly how to connect any enrichment provider to Claude Code via MCP and enrich your account list with verified contact data, without Clay as the middleman.

What is inside:

• MCP setup for common enrichment providers: Apollo, Hunter, Findymail, and Clearbit

• The /enrich Skill file: drop it into your project and run it on any account list

• The exact prompt structure that enriched 80 accounts in 20 minutes on a live client call

• A decision guide: when the direct MCP stack is enough and when Clay is still worth keeping

Get immediate access here.

Six months ago, a B2B founder running a specialised dev agency came to me doing outbound the hard way. Every morning, he’d open LinkedIn Sales Navigator, run the same saved searches, scroll through profiles, and copy prospect details into a Google Sheet. Then he’d open each company’s website in a new tab, scan their about page and recent news, and write a personalised email from scratch.

One by one. For every single lead.

He was spending 12 to 15 hours a week on this. Prospecting on Monday and Tuesday. Writing emails on Wednesday. Following up on Thursday. Updating his spreadsheet on Friday. And booking maybe 2 meetings a month. Good meetings, with founders who genuinely needed his team. But not nearly enough to replace the referral pipeline that had quietly dried up over the past year.

He’d looked at agencies. One quoted him $4,500/month with a 3-month minimum. Another wanted $6,000 plus a setup fee. Both pitched “guaranteed meetings” but couldn’t explain what happened to the data and the learnings when the contract ended.

He’d looked at hiring a junior SDR. The maths didn’t work: $60,000 base, 3 to 6 months before they’d book a single meeting, and he’d need to build the entire playbook for them anyway.

We installed this system in his business over 8 weeks. Same founder. Same offer, market, and ICP. The only difference: AI now handles the list building, research, enrichment, and first drafts. He reviews, edits where needed, and approves. Nothing goes out without his eyes on it.

After 60 days: his pipeline went from 2 meetings a month to 8 to 10. His time on outbound dropped from 15 hours a week to under 4. His total tool cost: $580/month. No complex 12-tool stack.

The system you’re about to read is the same one we built together. It’s not a theory. It’s a working pipeline blueprint that B2B founders at $1M to $10M ARR are using right now to generate consistent new business without a sales team, without agencies, and without spending half their week on manual prospecting.

The system has five layers. Each layer does one job. Together, they replace the $5,000/month agency retainer and the $120,000/year sales hire with something that costs less than $1,000/month and gets smarter every time you run it.

• Layer 1: Targeting. AI builds your prospect list from public data based on your exact buyer criteria. No database subscriptions required.

• Layer 2: Research and enrichment. AI researches each company, finds the right contact, and documents why they’re a fit before you write a single word.

• Layer 3: Messaging. AI drafts personalised outreach based on what it found, using your voice, your proven angles, and your rules. You review and approve.

• Layer 4: Sending infrastructure. Emails go out through properly warmed domains with deliverability handled. They land in the primary inbox, not spam.

• Layer 5: Signal-based follow-up. AI tracks who engaged. Hot leads surface automatically. You focus on conversations instead of chasing.

The rest of this guide breaks down each layer: what tools to use, what it costs, how to set it up, and where to go deeper.

The AI Stack: What You Need and What It Costs

Before getting into the layers, here is the full stack. Three tools. That’s it.

• Claude Code ($20 to $100/month): Your AI agent. It builds lists, researches prospects, drafts emails, and manages files. It doesn’t just suggest what to do. It does it. It opens your tools, runs searches, creates files, fixes its own errors, and works through tasks from start to finish.

• Apollo or LinkedIn Sales Navigator ($99 to $200/month): Your contact data source. Contact information, company data, and job change signals. Apollo is better for email-first outbound. Sales Navigator is better if your buyers are active on LinkedIn.

• Smartlead or Instantly ($50 to $150/month): Your email sending infrastructure. Domain warm-up, email sequencing, deliverability monitoring, and reply tracking.

Total: $170 to $450/month for the core stack.

Add Clay ($150 to $300/month) if you need multi-source enrichment at higher volume. Even at the top end, you’re under $750/month.

Compare that to the alternatives:

• A lead generation agency: $3,000 to $5,000/month. Pipeline stops when the contract ends. You own nothing.

• An SDR hire: $80,000 to $120,000/year fully loaded. Takes 3 to 6 months to ramp. No guarantee of results.

• An AI SDR tool (Artisan, AiSDR, etc.): $300 to $1,000/month. Automates sending but has zero context about your business. Generates volume without compounding.

The system below costs less than any of these and compounds because every campaign makes the next one better.

For the complete tool-by-tool breakdown, including what to use when and what to skip, read The only GTM tools you need in 2026.

The Brain: Why This System Gets Smarter Over Time

Every AI SDR tool on the market has the same problem: it knows nothing about your business. It sends emails based on templates and data, but it has no real context about your buyer, your voice, your past campaigns, or what has worked before. Every campaign starts from zero.

This system is different because of one file: your CLAUDE.md.

A CLAUDE.md file is a structured document that serves as the operating memory of your entire GTM system. It is loaded automatically into every Claude Code session and contains:

• Your buyer definition. Not “B2B SaaS companies with 50 to 500 employees” but the specific title, company situation, and breaking point that makes someone a real buyer this week.

• Your voice rules. Three to five concrete rules that make every email sound like you wrote it, not like AI generated it.

• Your proven angles. Subject lines and opening lines that got replies. Messaging approaches that converted. These get added after every campaign.

• Your exclusion list. Company types, industries, and geographies that never convert. Write them down once and the system skips them automatically.

• Your campaign history. What you sent, to whom, what happened. The AI reads this before drafting anything new.

This file starts small. After 30 days of campaigns, it becomes the most valuable sales asset in your business. After 90 days, your AI agent knows your market better than any SDR you could hire.

For the full breakdown on building a CLAUDE.md and the four components of an AI-ready GTM system, read Context engineering: the GTM skillset that makes outbound compound.

For the practical setup walkthrough of Claude Code, start with Claude Code for GTM: the first steps.

Layer 1: AI-Powered Targeting

Traditional prospecting works like this: you log into a database, set some filters (title, industry, company size), export a CSV of 500 names, and start emailing. The list is generic. The targeting is surface-level, and the reply rate reflects it.

AI-powered targeting works differently. Instead of filtering a database, you describe your exact buyer to Claude Code and it searches public sources to find companies that match.

The instruction looks like this:

“Find 30 B2B SaaS companies in the UK that raised a Series A or B in the last 18 months, have 30 to 150 employees, and are currently hiring in their finance function. For each company, record the company name, domain, funding stage, employee count, evidence of the hiring signal, and source URL. Save to companies.csv.”

Claude Code searches Crunchbase, LinkedIn, job boards, company websites, and news coverage. It doesn’t guess. If it can’t verify your criteria from a public source, it skips the company. The result is 30 to 50 companies where every row has a documented reason for being there.

This replaces the $200/month database subscription and produces a higher quality list because every company has verified fit evidence, not just a matching job title.

For the full targeting framework, including the WHO/WHY/WHEN model and the Claude Code prompt you can run today, read How to source high-quality leads.

For the deeper guide on defining a buyer profile that actually converts, read How to find customers that actually buy.

Layer 2: AI Research and Enrichment

You have a list of 30 to 50 target companies. Now the system needs to find the right person at each company, get their contact details, and build a brief on why they’re worth emailing.

In a traditional setup, this is 2 to 3 hours of manual research per batch. A junior SDR does this all day. In this system, Claude Code does it in minutes.

The instruction: “Take companies.csv. For each company, find the CEO or founder. Get their email from Apollo. Research their company website and recent news. Write a one-sentence personalisation note explaining why we’re reaching out. Save everything to prospects.csv.”

Claude Code connects to Apollo via API, pulls contact data, then researches each company using web search. It produces a CSV where every row has: name, title, email, company, personalisation note, and fit score.

If you need deeper enrichment across multiple data sources (combining Apollo, LinkedIn, Crunchbase, and company websites into a single view), Clay is still the best tool for that. The decision depends on volume. If you’re running fewer than 500 prospects per month, Claude Code with Apollo is enough. Above that, Clay adds value.

For a specific comparison, read Do you still need Clay in 2026?

Layer 3: AI-Drafted Messaging

This is where most AI outbound fails. Tools like Artisan and AiSDR generate emails from templates and data. They’re grammatically correct but contextually empty. They read like AI wrote them because AI wrote them without knowing anything about the sender’s business.

The CLAUDE.md file fixes this. Because Claude Code loads your buyer definition, voice rules, and proven angles before drafting anything, the output sounds like you.

The instruction: “Using prospects.csv and the cold email Skill, draft a 3-step email sequence for each prospect. Email 1: reference the specific signal from their personalisation note. Connect it to the pain point from the CLAUDE.md. End with a single ask under 10 words. Total length: under 80 words. Save all drafts to review-drafts.csv. Do not send anything.”

Claude Code writes the sequences, follows your rules, and saves everything for your review. You read through, make edits where needed, and approve. Nothing reaches a real prospect until you’ve signed off.

This is the critical difference between this system and every AI SDR tool: you stay in control of the message. The AI handles the labour. You handle the judgment.

For the detailed breakdown of the 3-step email sequence structure, read The cold email sequence that builds pipeline.

For the messaging principles behind emails that get replies, read Outreach copy that converts.

Layer 4: Sending Infrastructure

None of the above matters if your emails land in spam. Deliverability is invisible work that makes everything else possible.

The setup:

• Separate sending domains. Never send cold outreach from your primary domain. Buy 2 to 3 variants (e.g., if your domain is acme.com, buy getacme.com and acme-mail.com). Cost: $10 to $15/year each.

• Authentication. Configure SPF, DKIM, and DMARC on every sending domain. Without these, inbox providers flag your emails before a human sees them. Your sending tool (Smartlead or Instantly) walks you through this.

• Warm-up. New domains need 2 to 4 weeks of warm-up before sending cold email. Smartlead and Instantly both have built-in warm-up that handles this automatically.

• Volume limits. Stay at 30 to 50 emails per day per inbox. With 2 to 3 domains and 2 to 3 inboxes each, your total daily capacity is 150 to 300 emails. That’s more than enough for a founder-led system.

Load your approved sequences from Layer 3 into Smartlead or Instantly, set the sending schedule, and the emails go out automatically.

For the full deliverability walkthrough, including technical setup and the mistakes that land you in spam, read How to land in the primary inbox.

For what actually drives open rates (it’s not subject lines), read How to get your cold emails opened.

Layer 5: Signal-Based Follow-Up

Most prospects won’t answer to cold emails. But silence doesn’t mean no.

A prospect who opened all three emails, clicked the link in Email 2, and visited your website is a warm lead. They just didn’t reply yet. In a traditional setup, that signal is invisible. In this system, it’s tracked and surfaced automatically.

Your sending tool tracks opens and clicks. When a prospect engages but doesn’t reply, they go into a warm follow-up queue. The follow-up is different from the original sequence: a phone call referencing the resource they clicked, a LinkedIn connection request, or a short direct email acknowledging their interest.

This is where most of the meetings actually come from. Not from Email 1. It comes from intelligent follow-up based on engagement data.

For the complete signal-based follow-up process, read Turning cold emails into warm pipeline.

Adding LinkedIn as a Second Channel

If your buyers are active on LinkedIn, LinkedIn becomes a parallel pipeline channel that amplifies your email outreach.

The system is the same: targeted connection requests, personalised messages based on AI research, and signal-based follow-up. Claude Code handles the research and drafting. You handle the sending using tools like HeyReach or Prosp.ai.

The full LinkedIn pipeline playbook is in The Founder’s LinkedIn GTM Guide.

Supporting deep dives:

• The first step to making LinkedIn pay off: the foundation most founders skip.

• LinkedIn account-based targeting with job alerts: using hiring signals as intent data.

• Collecting LinkedIn leads on autopilot: scaling beyond manual prospecting.

• Turning connections into sales conversations: warm outreach that converts.

The Timeline: From Zero to Pipeline

Week 1: Set up Claude Code, create your CLAUDE.md file, connect Apollo. Buy sending domains.

Week 2: Domain warm-up begins (runs in the background). Build your first target list with Claude Code. Draft your first sequences. Review and refine.

Week 3: First emails go out. Start with 20 to 30 per day. Track deliverability closely.

Week 4: Ramp to full volume (100 to 200 per day). First replies arrive. First meetings booked.

Weeks 5 to 8: Run 2 to 3 campaign cycles. Update your CLAUDE.md with what’s working. Targeting gets sharper. Messaging gets better. The system starts compounding.

Week 9 onwards: The system is operational. You spend 3 to 5 hours per week reviewing AI output, approving sequences, and taking meetings. The AI handles the rest.

When to Hire (And When Not To)

This system is designed to be run by a founder. But at some point, the question becomes: should I hand this off?

The answer is specific. Hire only after you have:

• A CLAUDE.md file with at least 3 months of campaign data.

• A tested messaging sequence with documented reply and conversion rates.

• A working stack that someone else can operate without rebuilding.

• Consistent pipeline generation of 5+ meetings per month.

Without these, you’re paying someone to figure out your outbound from scratch. That’s a $120,000 experiment. With these, you’re handing someone a machine that already works. That’s a $60,000 hire who produces results in month one.

For the detailed hiring framework, read When to hire a GTM engineer and Should you hire a GTM strategist?

Common Questions

Q: How much does this system cost per month?

The core stack (Claude Code + Apollo + Smartlead) costs $170 to $450 per month. Adding Clay for advanced enrichment brings the total to $450 to $750. Even at the top end, the system runs for under $1,000/month. This replaces agency retainers of $3,000 to $5,000/month or an SDR salary of $80,000 to $120,000/year.

Q: How many meetings per month should I expect?

A well-targeted system sending 100 to 200 emails per day to a precise ICP typically produces 5 to 15 qualified meetings per month. The conversion rate from first email to booked meeting is roughly 1% when targeting is sharp. Founders who report lower numbers almost always have a targeting or deliverability problem, not a volume problem.

Q: Do I need Clay or is Claude Code enough?

For most founders running fewer than 500 prospects per month, Claude Code with Apollo handles targeting, research, and enrichment without needing Clay. Clay adds value when you need to combine multiple data sources into a single workflow at higher volume, or when your enrichment logic has complex conditional steps (e.g., “check LinkedIn first, then Crunchbase, then company website, use whichever has the most recent data”).

Q: Why not just use an AI SDR tool like Artisan or AiSDR?

AI SDR tools automate sending but lack context about your business. They don’t know your buyer, your voice, or what worked in previous campaigns. Every campaign starts from zero. The system in this guide compounds because your CLAUDE.md file carries all context and learnings forward. After 90 days, your system knows your market better than any external tool. AI SDR tools are useful only after you have a proven playbook you want to scale beyond what you can manage.

Q: Is this GDPR-compliant for European outreach?

B2B cold email is legal under GDPR when the legal basis of “legitimate interest” applies. This means: only contact people in their professional capacity at their business email, include a clear opt-out in every message, honour unsubscribe requests immediately, and be transparent about who you are and why you’re reaching out. The system described here is designed for targeted, relevant outreach to specific buyers, which is the strongest basis for legitimate interest.

Q: Can I really set this up without technical skills?

Yes. Claude Code operates through natural language. You type instructions in plain English and the agent executes them. “Find 30 SaaS companies in the UK that raised funding in the last year” is a valid instruction. You don’t write code. The initial setup (installing Claude Code, connecting APIs) requires running a few terminal commands, but Claude walks you through each step.

Q: How much time does this take per week once it’s running?

Once the system is operational (after the first 4 to 6 weeks of setup and initial campaigns), most founders spend 3 to 5 hours per week. That time goes to: reviewing AI-drafted sequences before they send, taking meetings with qualified prospects, and updating the CLAUDE.md with new learnings. The AI handles list building, research, enrichment, and drafting.

This guide is updated regularly as tools and tactics evolve. Every section links to a detailed deep-dive you can work through at your own pace.

When outbound stops working, the copy often gets the blame.

Founding teams rewrite subject lines, test new frameworks, switch tools, and eventually conclude that cold outreach is dead.

The list never comes up, even though the list is almost always where the problem started.

Before anything else works, three things have to come together:

1. Targeting: the right message to the right person. WHO is dealing with the problem you solve, right now.

2. Credibility: making it obvious you can actually fix that problem. Why should they trust YOU over anyone else.

3. Your Offer: something worth acting on. Why NOW, not next quarter.

Today we are covering number one, because without it, the other two are a waste of good copy.

Get targeting right and everything downstream gets easier.

Your emails write themselves because you know exactly what pain you are addressing.

Your reply rate starts reflecting the quality of your thinking rather than the luck of your timing.

By the end of this edition you will have a framework for building a hyper-targeted account list and a Claude Code prompt you can run today to pull your first raw list in under an hour.

1: WHO and WHY: How To Define Your Ideal Prospect

Founders usually know their WHO: The title, the industry, the company size. That is a starting point, but not a list. Thousands of companies fit that description right now, and most of them will never buy from you.

The missing piece is the WHY: The specific business context that makes someone a real buyer rather than a profile that looks right on paper.

Think about the last two or three clients you won. Ask yourself what was true about those companies at the moment you reached them. Not their industry, but their situation. What was breaking, shifting, or about to become urgent?

That context is your WHY. Write it down before you open any tool.

• Bad WHY: “Companies that need better reporting.”

• Good WHY: “Finance teams at startups with three to five entities doing month-end close in spreadsheets across multiple currencies. The CFO is reconciling manually. It takes eight to ten days. They have outgrown the process but have not yet hired a Controller to fix it.”

The second version tells you exactly what to search for and what signals confirm a fit. The first tells you nothing.

Three things worth getting specific on:

1. The breaking point. What has to be happening for your product to feel urgent, not just interesting?

2. Who feels the pain daily? Not always the person who signs. Your WHY needs to resonate with the first one to reach the second.

3. What your best customers had in common. Their situation at the moment they said yes, not their industry.

2: WHERE: Unique Prospect Data

Once your WHO and WHY are precise, the right data source becomes obvious. Different buyers leave different digital footprints, and Claude pulls from whichever source fits your ICP.

A practical map:

• Funded startups: Crunchbase, TechCrunch, Sifted, EU-Startups. Funding announcements come with company context, headcount signals, and investor names already baked in.

• B2B SaaS founders and early-stage teams: The YC company directory is underused and filterable by batch, industry, and stage. Product Hunt and Indie Hackers are worth checking if your buyer is earlier stage or bootstrapped.

• Companies using specific tools: G2 and Capterra surface companies actively reviewing tools in your category, which is a strong intent signal on its own. Job postings that mention specific software tell you exactly what a company runs without you having to guess.

• Professional services and SMBs: Companies House and equivalent local registries give you incorporation date, filed accounts, and director names. Industry directories and conference sponsor lists are underrated sources for niche verticals.

These are all key signals that you need include when collecting data.

The principle is straightforward: Go where your buyer already has a digital footprint that reflects their business reality, not just their job title.

3: WHEN: Trigger Events: When Prospects Are Ready To Buy

This part is separate from the WHY and comes after your core list criteria are defined.

Trigger events do not determine whether a company belongs on your list. Your WHO and WHY already do that.

What trigger events tell you is which companies on your already well-targeted list are worth contacting this week rather than next quarter.

A company that fits your WHO and WHY with no recent trigger event is still a valid prospect, just with lower urgency.

A company that fits your WHO and WHY and just hired a new CFO, raised a Series B, or posted three SDR roles at once has a concrete reason to act now. That difference matters when you are deciding where to spend your outreach energy first.

One thing worth being careful about: building a list entirely around trigger events without validating fit first produces companies that are active but wrong. Active and wrong burns your reputation faster than anything else.

Trigger events worth tracking:

• New funding round: growth mode, new budget, new pressure to show results fast

• New C-suite hire in your target function: new leader, new priorities, short window to influence

• First sales leader hire: founder transitioning out of founder-led sales, systems do not exist yet

• Multiple open roles in one function at once: scaling pain is already visible

• Job posting mentions a tool you replace or integrate with: live evaluation happening right now

Think of trigger events as a prioritisation layer. They sit on top of a list that is already well-targeted. Without strong WHO and WHY underneath them, they are just noise.

4: Building the List with Claude Code

You no longer need Apollo or Sales Navigator to build a targeted account list. Claude Code with Brave Search pulls structured prospect data from public sources based on your exact criteria. No exports, no manual tabs or spreadsheet wrangling.

Here is a prompt you can run today inside your project folder. Adapt the criteria to your ICP:

Search for B2B SaaS companies in the UK that:
- Have raised a Series A or B in the last 18 months
- Have between 30 and 150 employees
- Operate across multiple entities or geographies
  (look for clues in their website, job postings,
  or news coverage)
- Are currently hiring in their finance function
  (CFO, Controller, FP&A Analyst, or similar)

Search these sources in order:
1. Crunchbase for funding data and headcount
2. TechCrunch and Sifted for recent announcements
3. LinkedIn for hiring signals and team structure
4. Company websites and careers pages to confirm
   multi-entity structure

For each company found, record:
- Company name
- Domain
- Funding stage and amount if available
- Employee count estimate
- Evidence of multi-entity structure
  (quote the specific source and what it says)
- Finance roles currently posted
- Buying signal: HOT, WARM, or COLD
  HOT = multi-entity confirmed + active finance hire
  WARM = multi-entity likely + no active hire
  COLD = fits WHO criteria but WHY unconfirmed
- Source URL

Save to companies.csv. Aim for 30 to 50 companies.
Skip any company where you cannot find evidence
of the multi-entity structure from a public source.
Do not guess. Write UNKNOWN rather than assume.
After every 10 companies, save the file.

A few things that make this prompt stronger than a generic search:

• The source order matters. Telling Claude where to look first reduces the chance of it pulling thin results from low-quality pages.

• The evidence requirement is explicit. Asking Claude to quote the specific source means every row in your CSV has a documented reason for being there, not just a name that looks right.

• The scoring criteria are defined inline. Without this, HOT and WARM mean something different every session. With it, the ratings are consistent.

• The save frequency is locked. Saving after every 10 companies protects against session timeouts on larger lists.

The result is not a list of 500 names. It is 30 to 50 companies where every row has a documented reason for being there.

The Takeaway

A target account list is not a database of names that match a job title filter, but a collection of companies where you have documented evidence that the business context is right, the fit is there, and the timing makes sense.

• The WHO tells you which door to knock on.

• The WHY tells you whether anyone is home.

• The WHERE tells you where to find them.

• The WHEN tells you whether now is the right moment to knock.

Get all four right before you touch enrichment and everything downstream gets sharper.

Your research produces useful briefings. Your emails land harder.

Your reply rate reflects the precision of your list, not just the quality of your copy.

In the next edition, we take this companies.csv and run it through lead enrichment APIs to fill in contact data, validate signals, and get it ready for the sequence.

Until then,

Jasper

PS: Want the full list building system?

The Claude Code GTM Account List Builder takes you from blank folder to scored companies.csv in under an hour.

What is inside:

• A WHO and WHY worksheet to translate your customer base into precise, filterable criteria

• A pre-written CLAUDE.md template with every block ready to fill in

• Full project folder setup with exact folder structure and terminal commands

• A ready-to-run /list-builder Skill file for your .claude/commands/ folder

• A companies.csv template pre-structured and ready for the enrichment guide

• A buying signal scoring guide with real examples across four common ICP types

• Serper and OpenWeb Ninja setup walkthrough with a decision tree for your ICP type

• A validation checklist to confirm every row is ready before enrichment

• A troubleshooting guide covering the six most common problems with exact fixes

• A list maintenance system with weekly, monthly, and quarterly habits

Get immediate access here.

I have yet to meet a founder who said their AI outbound is compounding.

Most build a campaign, send it out, get inconsistent results, and start over. Six months in, the AI knows nothing more about their business than on day one.

Take your mind back to late 2022. ChatGPT launches and everyone suddenly has an AI writing assistant. A few months later, most have quietly gone back to writing emails themselves. The output was generic. It could have been written for any company selling anything to anyone.

The founders who got real results figured out something everyone else missed: They stopped treating AI like a search engine and started treating it like a new team member who needed proper onboarding before touching anything important.

That shift is called Context Engineering. The difference between outbound that plateaus and outbound that compounds.

Generative vs Agentic AI

Every AI tool that we have used so far, ChatGPT, Gemini, the standard Claude interface, works the same way. You type somet…

Read more

I remember looking at a tool called Clay back in 2022.

At the time, many just saw it as another outbound tool. It felt technical, and frankly, a little intimidating too.

But I saw something else. I saw a tool that was going to fundamentally break the way we build lead lists.

If you listened to me then, you know what happened next. Clay became the operating system for modern GTM.

I have that exact same feeling today.

We are standing at another turning point, and this time it was launched by a single command in a terminal window: Claude Code.

This is not AI hype. It’s a structural shift in how companies are built.

For the last decade, we have been living in the No-Code era. We paid premiums for visual interfaces, drag-and-drop builders, and colourful buttons because they made us feel safe.

We treated the Terminal - that black screen with the white text - as a “Do Not Enter” zone reserved for engineers.

Claude just gave the rest of us keys to the engine room. And if you are willing to get over the initial fear of the black screen, you are about to realise that you don’t need to hire a developer to build your GTM stack.

You just need to learn how to ask the computer to do it for you.

We are going to cover the first steps in getting to work with it:

• What is Claude Code (and why the terminal matters).

• How to install it (even if you have never written a line of code).

• How to run your first agent to build a lead list on autopilot.

What is Claude Code?

People often think using AI means opening a browser tab to chat with a bot. You paste some text, it gives you some text back, and then you have to go do the actual work yourself.

Claude Code is different because it acts as an Agent that lives in your computer’s terminal.

The standard chat interface is passive, but Claude Code is active. When you give it permission, it actually creates files on your desktop, runs scripts, and searches the web. If it runs into an error message, it reads it and fixes its own code without you even needing to intervene.

Think of it like a very fast junior developer who has access to your keyboard. If you tell a standard chatbot to find the pricing page for 50 competitors, it might give you a Python script that you don’t know how to use.

If you tell Claude Code to do the same thing, it just gets to work. It writes the script, runs it, handles the anti-scraping blockers, and saves a CSV file called competitors.csv right to your desktop.

You didn’t write a single line of code, but the work got done.

The Terminal

When a non-technical founder sees a black terminal window, they usually panic. It looks like Matrix code where one wrong keystroke deletes the company database.

But using the terminal just means cutting out the middleman. Here is the entire setup process to get started. It takes about 5 minutes.

The Setup (One Time Only)

Step 1: Install Node.js: Claude Code runs on Node.js. Go to nodejs.org, download the LTS version, and run the installer. Click through the defaults. When it’s done, you can verify it worked by opening any terminal and typing node --version. If you see a number, you’re good.

Step 2: Get the Editor: Download and install VS Code (Visual Studio Code). This is where you’ll manage your files and access the terminal.

Step 3: Open the Command Center: Open VS Code. In the top menu bar, click Terminal > New Terminal. A small window will pop up at the bottom of your screen. That is your cockpit.

Step 4: Install the Brain Copy and paste this exact command into that terminal window and hit Enter:

npm install -g @anthropic-ai/claude-code

It will churn for a moment while it downloads. When it finishes, you’re done. You’ve installed the agent.

You’ll also need an Anthropic API key (pay-as-you-go, costs pennies per task) or a Claude Pro/Max subscription. Claude Code will walk you through connecting your account the first time you run it.

From now on, whenever you want to start work, you just open this terminal and type claude.

Building a Lead List

Let’s look at a concrete example. If you wanted to build a list of the latest Y Combinator startups in the AI sector, you have two bad options: pay for a scraping tool or hire a Virtual Assistant to copy-paste for three hours.

With Claude Code, this becomes a short conversation.

Step 1: Wake the agent. Open your terminal, navigate to your project folder, and type claude to start the session.

Step 2: Give the order. You don’t need code. Just type in plain English:

“Search for the top 20 Y Combinator startups in the Winter 2025 batch that focus on AI sales. Find the names of their founders. Save the company name, founder name, and website into a CSV file called ‘leads.csv’.”

Step 3: Watch the execution. Claude doesn’t just write the script, it runs it too. It will search for the data, parse what it finds, and write the CSV file directly into your project folder.

A note on expectations: this isn’t magic. Sometimes the source it finds is incomplete, or it needs a follow-up nudge from you (”You missed the last 5 — try searching the YC directory directly”). Think of it like managing a fast but junior intern. The first pass gets you 80% of the way there. A couple of corrections gets you to 95%.

When it finishes, a file called leads.csv appears in your folder. You didn’t write a single line of Python. You just managed the intern.

MCP: The Secret Weapon

If you only use Claude to write files on your desktop, you’re scratching the surface. The real power comes from a standard called the Model Context Protocol (MCP).

Think of a standard AI model as a brain in a jar. Smart, but no hands. It can plan a trip, but it can’t book the flight.

MCP gives the brain hands. It acts as a universal plug that lets Claude connect to real-world tools: search engines, databases, APIs.

Instead of buying an enrichment tool, you can plug Claude into Brave Search (a free search API) and have it act as your own private researcher.

How to Set It Up

Step 1: Go to brave.com/search/api and create a free account. The free tier gives you 2,000 searches per month, which is more than enough. Once you’re in the dashboard, generate an API key. Copy it.

Step 2: Open your regular terminal (not inside a Claude session). Navigate to your project folder and run this command, replacing YOUR_KEY_HERE with the key you just copied:

claude mcp add brave-search -e BRAVE_API_KEY=YOUR_KEY_HERE -- npx -y @modelcontextprotocol/server-brave-search

That’s one command. It tells Claude Code to connect to the Brave Search server and use your API key to authenticate.

Step 3: Start a new Claude session by typing claude. Then verify it worked:

“What MCP tools do you have available?”

If you see brave_web_search in the list, you’re live. Claude can now search the internet in real time.

What This Unlocks

Now you can run prompts like this:

“For every company in my ‘leads.csv’ file, search the web for the CEO’s name and their most recent podcast appearance or interview. Add a column called ‘Icebreaker’ to the CSV with a one-sentence reference to that appearance.”

Claude goes out to the web, researches each person individually, and updates your file. No scraping tool. No VA. Just a plain English instruction.

Context Files: The New Prompt Engineering

In the ChatGPT era, you learned “Prompt Engineering.” In the Agent era, you’ll learn Context Engineering.

You don’t paste instructions every time. Instead, you create a “Rule Book” that lives in your project folder. This file is called CLAUDE.md.

Here’s the key insight: when you type claude inside a folder that contains a CLAUDE.md file, the agent reads it automatically. It applies every rule in that file to everything it does — without you saying a word.

This file is the Constitution for your agent. It tells it who it is, what the product does, and most importantly, what it is forbidden to do.

The CLAUDE.md Template

Create a file named CLAUDE.md in your project folder and paste this in:

# GTM AGENT: [YOUR COMPANY NAME]

## 1. IDENTITY & GOAL
- **Role:** You are the Lead RevOps Engineer for [Company Name].
- **Objective:** Identify high-intent prospects in [Industry] and prepare
  technically accurate outreach data.
- **Core Value:** Precision over volume. 10 perfect leads beat 100 mediocre ones.

## 2. THE PRODUCT
- **What it is:** [1-sentence description of your product].
- **Value Prop:** "[Your main promise/benefit]."
- **Killer Feature:** We integrate directly with [key integration or unique capability].

## 3. IDEAL CUSTOMER PROFILE (ICP)
- **Company:** [Size/Revenue range] in [Geography].
- **Title:** [Job titles, e.g., CTO, VP of Sales].
- **Buying Signal:** Look for companies that have [specific trigger, e.g.,
  raised funding, posted a job for X role, installed Y software] in the last 30 days.

## 4. THE "NO-GO" LIST
- **Never target:** [Agencies / consultancies / companies under X employees].
- **Skip these industries:** [Industries that don't fit].
- **Competitors:** Never reach out to [Competitor 1], [Competitor 2].
- **Geography:** Ignore [regions you don't sell to].

Without this file, you explain your business every time you open the terminal. It’s like hiring a new intern every morning.

With CLAUDE.md, the agent has permanent memory.

It knows the Trigger: it doesn’t just look for “CTOs”. It looks for CTOs who just raised funding or posted a specific job.

It knows the Product: it understands your integrations without you spoon-feeding details.

It knows the No-Go List: it skips agencies, wrong geographies, and competitors automatically. This is the part most people miss. Telling the agent who to ignore saves you from burning your domain on bad-fit prospects.

Fill in the brackets. Save the file. Now just type “Find me 5 leads” , and Claude applies every rule without you saying a word.

The Danger Zone

I need to give you a warning before you start running commands.

Claude Code is fast. That is its biggest feature, but it is also its biggest risk. It makes mistakes faster than you can catch them.

Before, manual work was actually a safety mechanism. When you had to copy and paste email addresses one by one, you naturally slowed down. You noticed if a name looked weird or if a company didn’t fit your list. That friction gave you time to think.

Claude removes the friction entirely.

Say you tell it: “Personalise this email based on the company’s latest news.”

It scans Row 42. That company just announced a massive layoff and a lawsuit. Claude reads the headline but lacks the empathy to understand the context.

It writes: “Huge congrats on the restructuring news! Exciting times ahead.”

That draft goes into your outreach file. If you’re not paying attention and you blast it out, that email lands in the CTO’s inbox, along with 49 more just like it.

If you are reckless, you can torch your reputation in sixty seconds.

Treat this like a loaded weapon. Never let the agent fire without a safety check.

The “Safety First” Protocol

Rule #1: The Dry Run. Always ask Claude to draft, never to send. (Claude Code doesn’t send emails on its own unless you’ve specifically wired up a sending tool, but building the habit now protects you later.)

• ❌ Bad: “Email everyone in leads.csv.”

• ✅ Good: “Draft emails for the first 5 people in leads.csv. Save them to a file called drafts.txt so I can review them.”

Rule #2: The Batch Limit. Never process a whole file at once. If you have 500 leads, start with 10. If Claude hallucinates or writes a bad subject line, you want to catch it when it costs you 10 leads, not your entire market.

Rule #3: The Eye Test. Before you approve anything, open that drafts file and read the outliers. Check the companies with recent news. Did Claude mistakenly congratulate a CEO on a lawsuit? Did it misinterpret a “hiring freeze” as “growth”?

The agent does the work. You do the judgment. That’s the deal.

The Takeaway

We are witnessing the end of Software as a Service and the beginning of Service as Software.

You are no longer a User.

For the last twenty years, you paid for software that helped you do the work.

• Salesforce helped you organize data, but you had to type it in.

• ZoomInfo gave you the phone number, but you had to make the call.

• Mailchimp sent the email, but you had to write it.

Claude Code changes the contract. Instead of paying for a tool to help you, you are paying for an agent to do it for you.

This creates a new responsibility for you.

When you use HubSpot, you are a User. You click buttons. When you use Claude Code, you are a Manager. You give orders.

If you are a bad manager who gives vague instructions, your employee will produce garbage. The only difference is that Claude will produce garbage at 1,000 words per minute.

The terminal is now the most direct way to speak to your new workforce.

Jasper

PS: I’ve packaged an entire 4-Phase Claude Code architecture into the The CLAUDE CODE GTM Starter Guide. This transforms Claude from a simple chatbot into a Strategic Analyst.

Inside the Pack:

• 1: The Master CLAUDE.md: The 50-line “Constitution” that forces Claude to adopt your voice, ignore “Agency” leads, and think like a senior partner.

• 2: The Hands Setup (MCP): A 2-minute guide to installing Brave Search so your agent can browse the live web (LinkedIn, Crunchbase, News).

• 3: The Researcher Script: This script visits the company website, reads their Careers page to spot hiring trends, and categorises them by intent.

• 4: The Deep Dive Dossier: A script that takes your top 5 prospects and generates a 1-page briefing doc for each. It summarises their recent news, identifies their tech stack, and writes a “Why Now?” hypothesis for your pitch.

Get access here.

I admitted a hard truth to Chris Hines on his Next Round Ready podcast recently.

Two years ago, if a founder asked me, “How do I fix my sales pipeline?” I would have sold them a GTM Strategy sprint.

You would pay a consultant $15,000. They would interview your team, audit your stack, and three weeks later, they would hand you a beautiful Roadmap.

And back then, it was often worth every penny.

Today, paying for such a Roadmap seems like burning cash. You no longer need to hire a human to answer that question, because you just ask your LLM.

You upload your website to Claude or o1-preview, paste in your value prop, and ask for the ideal GTM strategy.

And honestly? The answer is usually 80% correct. Plus, it does in 30 seconds what used to take us 30 days.

The market value of information has fallen to zero.

The Canary in the Coal Mine

from Moxie GTM shared an honest update last week about the collapse of his agency.

In January 2025, he was doing $30,000 MRR. By April, he was at $0.

He pivoted, t…

Read more

For the last two years, if you asked me how to scale outbound, I had one answer. Get Clay.

Clay basically a spreadsheet with a brain. You give it a list of companies and it automatically finds the CEOs, their emails, and their open job roles. It turns a week of manual research into a 10-minute coffee break.

But if you’re a user, you know it comes with a headache.

You might stare at your credit usage and worry that one wrong setting will burn $1,000 in two minutes. That’s “credit anxiety.” Or maybe you spend your weekend fixing a workflow that broke just because a website changed its layout.

We put up with it because we had to. We paid the “No-Code Tax” because we didn’t know how to code.

That changed in 2026.

When Anthropic launched Claude Code, the barrier to entry collapsed. Now the same Clay experts who sold you on Clay are quietly cancelling their subscriptions. They are moving to the terminal.

Here is why the pioneers are trading expensive tables for affordable agents.

The Three Eras of …

Read more

For the last 15 years, the B2B playbook was simple. It was the “Predictable Revenue” model: To scale sales, you just hired more bodies.

Believe it or not, but most sales orgs are still executing this today. You hire recent grads, give them a script, and tell them to dial.

But while the activity remains the same, the returns have collapsed.

It’s 2026, and we are entering the Agentic Phase of GTM.

You’ve seen the buzz around Claude Code, Cowork, and MCP Servers. We have moved from chatbots that just talk, to agents that can use tools, browse the web, and execute workflows without you.

Now ignore the tech hype for a second. This shift is happening because the old economics are broken.

It comes down to basic math and physics:

• The math: A human SDR costs roughly $80,000 a year. Yet, data shows they spend only 30% of their week actually selling. The rest is lost to admin and research.

• The physics: Humans get tired. The first email of the day is personal. The 50th is a template.

Teams often try to f…

Read more

We need to discuss the ‘Magic Hire’ fantasy.

If you’re a founder grinding through the early days of founder-led sales, you have likely looked at the explosion of GTM automation tools (Clay, n8n, Apify, etc.) and thought:

“If I just hire one brilliant engineer to hook all this up, I can stop doing manual sales, and automate everything!”

This is the most seductive narrative in GTM tech right now. It promises that code can replace headcount and that you can replace a team of 10 burning-out sales development reps with one genius writing Python scripts in a dark room.

But before you spend $200,000 trying to build a Ferrari engine, you need to check if you are still driving a go-kart.

Putting a rocket engine on a sinking ship doesn’t make it fly. It just makes it sink faster.

To make the right decision, let’s understand what this role actually is, where it came from, and why you most likely do not need one yet.

What is a GTM Engineer?

The GTM Engineer differs from a rebranded developer. Clay effect…

Read more

2025 GTM stack advice on LinkedIn has turned into engineering cosplay: complex workflows that look impressive but ignore business reality.

These 12-step diagrams cost $3,000 a month and require a full-time GTM engineer to maintain.

Complexity is profitable for the sellers. SaaS platforms get paid per seat and workflows grow longer every month, but you are the one left carrying the burden of the maintenance.

Before you buy a single tool, map your workflow to the 3 Core Categories below.

If a tool doesn’t fit cleanly into one of these buckets, cut it.

Category 1: Data Sourcing

1. The Accounts (Who fits?)

Founders often default to LinkedIn. But before you buy a new SaaS tool, ask yourself where your target accounts actually live.

• Standard B2B: If you sell to standard tech or service companies, they live on LinkedIn. → You need LinkedIn Sales Navigator. The gold standard for title, industry, and headcount filtering.

• Bootstrapped or early-stage companies: If you sell to founders who have not ra…

Read more

I speak to founders every week who are drowning. They are building the product, managing the team, and trying to run sales all at once.

The result is usually a calendar that looks like a crime scene:

You spend 45 minutes researching a prospect before a call.
You spend hours reading competitor reviews to find an angle.
You spend your Sunday evenings trying to summarise financial reports to find a budget trigger.

This is low-leverage work. If you’re doing manual data synthesis in 2025, you’re voluntarily lowering your hourly rate to $15.

Today I’m sharing three GTM ‘Analyst’ prompts from my personal library. These aren’t prompts to replace your thinking. They exist to fuel it.

AI for reading, not just writing

Founders love asking ChatGPT to write emails, but that is actually its weakest skill. It defaults to the average.

AI’s superpower is synthesis.

An LLM can read a 100-page annual report, scan 500 reviews, and analyse a LinkedIn profile in seconds. A human takes hours.

If you want leverage, l…

Read more

A founder told me last week she was considering an AI SDR.

“Why not automate the whole thing? If AI can do the work of three SDRs for a fraction of the cost, it’s a no-brainer, right?”

I get the fantasy. You dislike sales, hiring someone is exhausting, and expensive consultants rarely deliver.

So when a vendor promises an “AI Agent” that will automate your entire outbound motion for $300 a month, you buy it.

You imagine a world where you push a button, and the AI does the thinking, writing, and strategy while you just need to close the deal.

Here’s the thing: Outsourcing your thinking to an LLM will simply scale your own mediocrity.

Plenty of teams are currently using the most powerful tech in human history to send average spam at a speed we have never seen before.

Let’s look into why the AI SDR is failing, and how the top 1% of founders are using AI to actually build pipeline.

The Data

Let’s look at the scoreboard.

Kyle Poyar’s Growth Unhinged recently surveyed 195 GTM leaders. The results were brutal:

• 75% feel top-down pressure to “adopt AI” immediately.

• 53% see zero or limited impact from it.

The disillusioned outcome is most acute with “AI SDRs.” A founder confessed:

“We tried an AI SDR for six months and were unable to generate a single opportunity.”

Zero.

Founders often rush to plug in every new AI tool, hoping tech will solve the problem. But without a strategy, three months fly by and the sales pipeline stays empty.

The Psychology

Why is the failure rate so high? from Cannonball GTM sees the specific psychological mechanism at play.

Humans are elite pattern recognisers. If we are not interested in one generic email, we are not interested in any of them.

AI models are trained on the “average” of all sales emails ever sent. When you ask an AI to write a sequence, it reverts to the mean. It produces a perfectly average message.

But humans punish “average.” We subconsciously filter it out.

The “Zero Opportunity” stat isn’t a tech failure but a pattern recognition success.

Your prospects know it is a bot within milliseconds, so they delete it.

The Operations

So is AI useless? No. (SaaStr) clarifies the nuance here.

He notes that AI agents do work for them. He has 6+ in production right now. But there is a catch that most “set it and forget it” founders miss:

The AI vendors actually succeeding right now have ACVs of $50k-$100k.

Why does the expensive stuff work while your cheap tool fails?

Because at $100k, the vendor provides humans to train, fine-tune, and orchestrate the AI for you.

The $300/month self-serve tool fails because it relies on you to train it, and you are treating it like a magic wand instead of a new employee.

You can replace most outbound SDRs with AI, but only if you orchestrate it daily. You cannot just push a button.

Be the Thinking Partner

So what does “AI orchestration” look like?

It means you provide the ingredients. The AI provides the scale.

Here is an example of a strategy prompt that I use to find high-intent leads without buying expensive intent data.

The Prompt:

I sell a compliance automation tool for Fintechs.

Scan the career pages of these 50 target fintech accounts. Look for job postings for “Head of Compliance” or “Risk Officer” that have been open for more than 60 days.

If you find one, assume they are understaffed and stressed. Output a list of these companies, the specific job title, the link to the job post, and a one-sentence hypothesis about the specific compliance risk they might be facing based on the job description requirements (e.g. mention of “SOC2” or “GDPR”).

The Output:

• Company: NeoBank X

• Role: Chief Risk Officer (Open 74 days)

• Hypothesis: The JD specifically mentions “preparing for upcoming SEC audit,” suggesting they are currently managing this audit with insufficient leadership (signal).

Now you have a loaded gun. You aren’t sending a generic “checking in” email. You are sending a note to the CEO saying:

“Looks like you’ve been trying to fill the CRO role for 2 months right before your SEC audit. We can automate the SOC2 prep so you pass even if that seat stays empty.”

That is how you use AI for GTM.

Takeaway

Tools come and go. But you are stuck with your own mind forever.

Providing the AI with the right ingredients, such as your unique POV and specific customer insight, keeps the output valuable.

Do not buy the Magic Button. The magic exists in the strategy you feed it.

Jasper

The prompt in this article is a starting point. The thinking behind it is what the Claude Code GTM Starter Guide packages into a full system.

Inside the guide:

• The CLAUDE.md constitution: a 50-line file that forces Claude to think like a senior GTM partner, not a generic assistant. It knows your ICP, your voice, and what to ignore before you type a single word.

• The MCP setup: a 2-minute connection that gives Claude live access to the web. LinkedIn, Crunchbase, job boards, company news. No more working from stale data.

• The Researcher Skill: visits company career pages, spots hiring trends, and categorises every prospect by buying signal. The same logic as the compliance example above, built into a reusable command.

• The Deep Dive Dossier: takes your top 5 accounts and produces a one-page briefing for each. Recent news, tech stack, and a Why Now hypothesis sharp enough to open a sales call with verbatim.

Get access here.

Two companies proved that the answer might be “later than you think.”

Warmly.ai recently finalised its Series A ($17M total funding). They used founder-led insights to drive a recent 5 times ARR surge with a new AI sales platform.

Ocean.io is sitting around $6M ARR, yet their founder still sends outbound messages. The result? Doubled demo bookings and a 25% lift in close rates.

Different stages. Same result.

The founders stayed close to the conversation long enough to encode market insights into the product.

We often talk about Founder-Led Sales as a bridge to hiring a sales leader. But looking at these two, it’s clear that it’s actually an engine for clarity.

Here is the pattern I’m seeing:

The silent killer: signal decay

Many founders have the same instinct: validate the product, hire sales leadership, and focus on strategy.

But there’s a hidden cost to delegating the funnel too early. Signal Decay.

When you step out …

Read more

Welcome back to the Founder’s GTM with lean strategies to land clients predictably.

Four founders running outreach on LinkedIn sent me their numbers. Different offers, different markets, but same pattern.

The only thing they had in common was this: before they used a system, their DMs were getting ignored. After the system, conversations started happening.

This edition breaks down the friction that kills conversations and shows the 3-step DM system that fixes it.

You’ll see exactly how to open a real business dialogue without sounding like every other seller on the platform.

Let’s get into it.

Read more

Welcome back to the Founder’s GTM, with lean strategies to land new clients predictably.

We’ve spent the last few weeks deep in the weeds of email and LinkedIn: building signal lists, messaging and connection systems that actually go somewhere. Before we continue, I want to take a short break from tactics today.

Because there’s one topic we haven’t really touched yet: AI. That’s intentional. I don’t build AI-powered systems. I build AI-ready ones.

AI hype

Every founder right now feels pressure to do something with AI. Tools are everywhere, promising to write your messages, qualify your leads, book your calls.

But here’s what I keep seeing: people trying to automate parts of their GTM they don’t even understand yet.

The result looks smart from a distance: dashboards, agents, sequences. But up close, it’s just noise on autopilot.

The problem comes from poor judgment, not missing tools.

Automation without understanding is chaos

A recent study from Carnegie Mellon University found that leading AI agents failed nearly 70% of standard business tasks in simulation.

Another MIT and Fortune report claimed that roughly 95% of enterprise AI pilots fail to deliver expected value.

That stat could have been written about GTM. Teams collect tools like trading cards, then wonder why nothing compounds. They automate before they clarify, and chase the next integration while their positioning is still mushy.

That’s like hiring a salesperson before you know what they’re supposed to say.

Automation belongs to what’s already working, not what’s still unproven.

The AI-obsessed founder

I recently worked with a fintech founder who was in love with his AI sales system. He’d automated everything from data enrichment and message writing to follow ups. After 8,000 sends, he had booked only two meetings.

We tore it apart. His targeting was vague, chasing “enterprise finance teams,” his message packed with AI buzzwords, and none of it had been tested manually.

We rebuilt from the ground up. Focused on one clear ICP: operations leads at mid-market fintechs using Stripe. Sent 120 emails, tracked every reply, action, and refined the copy each day.

His meeting rate jumped from 0.03% to 1.3%. After scaling that working system, he landed 4 deals worth €48K ARR in six weeks.

He didn’t need more AI. Jus clarity.

What AI-ready actually means

Being AI-ready starts with structure: systems so clear they can be automated later without breaking.

When your process already works, AI becomes leverage. When it doesn’t, AI just multiplies the chaos.

AI won’t build your strategy for you. It can only build with you. It doesn’t know why buyers pick one product over another or what makes your story resonate.

Founders love to argue with that, mostly because it’s easier to believe AI can skip the hard parts. But AI doesn’t infer context it hasn’t lived. It can’t tell you why a deal slipped or why one message landed while another died.

Until you can explain your own logic clearly, giving it to AI just outsources confusion. It won’t invent clarity and only scales whatever thinking you’ve already done, good or bad.

That’s why I treat automation like delegation: never hand off what you don’t already understand.

Real business intelligence is still human

Here’s how I think about it:

Let automation handle what’s predictable: cleaning data, enriching leads, following up on inbound leads. Let humans handle what’s unpredictable: timing, testing, nuance, and real conversation.

Even major enterprise surveys show that most AI deployments struggle because teams lack structure and clarity, not because the tech is weak.

AI doesn’t replace human intelligence. It just reveals who has structure and who’s winging it.

Founders who get this use tech to multiply good judgment. The rest use dashboards to measure how efficiently they’re getting nowhere.

Takeaway

Before you build an AI-powered GTM, make sure it’s AI-ready. Prove your message, test your motion, and understand your buyers before you ever start automating.

That’s exactly why I started The Founder’s GTM: it’s a blueprint for building founder-led GTM systems that actually work before you scale them.

If your system is broken, AI won’t fix it. It’ll only make the cracks show faster: in your messaging, your targeting, and your process. It’s a race to the bottom.

This Substack exists to help you build judgment first, then efficiency.

Next week, we’ll return to the practical series and talk about how to turn your new connections into business partners.

Until then: keep it simple, stay human, and don’t let automation steal your judgment.

Jasper

Want help making your GTM system AI-ready? Book your outbound optimiser session. I’ll review your current setup and show you what to automate (and what to keep human).

If you’ve been executing the previous edition’s LinkedIn playbook, you’ve built a steady stream of leads by pulling live signals like post engagement, “Actively Hiring” badges, and competitor connections.

Your list is now full of intent leads already showing interest in a solution like yours.

Now comes the part many founders fumble: turning those leads into connections.

The problem: overcomplicating connection

The default move on LinkedIn is to fire off connection notes that try way too hard to prove “personalisation.”

Every person writing “Saw you liked…” or “We’re both in…” is broadcasting the same subtext: I’m about to sell you something.

The more fluff they add, the less authentic the message feels. And when 70% of requests go ignored, they have no idea why.

Worse, most think they can compensate for a weak profile with clever copy. But no line of text can fix weak positioning.

If your h…

Read more

Welcome back to Founder’s GTM⚡ with lean strategies for business owners to win clients predictably.

Last time we dove into account-level intent using LinkedIn job alerts. This week, we scale.

Instead of reacting to accounts already in motion, you’re going to build a persona-level radar that auto-collects high-signal leads from LinkedIn.

More volume, but without losing context.

The problem: scaling the wrong thing

When pipeline slows, the instinct is usually the same: just send more messages.

But more outreach ≠ more pipeline when it’s based on flat criteria (industry, headcount, title). You’re scaling noise, not signal.

Worse: most “scalable” outreach ignores relevance and timing: the two things that actually get you replies, meetings, and eventually revenue.

Persona signals are the goldmine

Funding rounds and job posts are useful signals, but they’re table stakes. Everyone sees them.

What people often miss are the LinkedIn-native persona triggers that reveal actual buyer behavior:

• Who RSVP’d to…

Read more