The most useful Claude Code workflows for GTM teams are not the long demos. They are the 30-second prompt patterns you run every week to ship campaigns, score leads, clean the CRM, and run signal-based outbound. This is a copy-paste recipe collection covering 30 specific Claude Code workflows for GTM across six categories. Each recipe is a working prompt pattern plus the tools it depends on.
Treat this as a reference. Skim to the category that matches your motion. Steal the prompts that fit. Adapt the rest to your CLAUDE.md.
Key takeaways:
30 recipes across six categories: ICP research, enrichment, segmentation, sequencing, CRM hygiene, and signal-based outbound.
Every recipe assumes a stack of Claude Code plus a data layer (e.g. Databar) plus your CRM. Most also assume a sending tool MCP.
The recipes work because the prompts are short, the tools are explicit, and the output format is structured (table, JSON, or Markdown brief).
Most GTM teams running Claude Code workflows hit productivity gains in week two, once they have a CLAUDE.md tuned to their motion.
What Claude Code Workflows for GTM Look Like
A Claude Code workflow for GTM is a short, repeatable prompt pattern that calls one or more tools (data layer, CRM, sending platform) and produces structured output you can review. Not a chat session. Not a long demo. A 30-to-90-second pattern you run multiple times a week.
Three things make a workflow work:
A CLAUDE.md tuned to your motion. ICP definition, voice rules, closed-won patterns, forbidden phrases. The agent reads this before every prompt.
Tools wired through MCP. Databar for enrichment, Smartlead for sending, Attio or HubSpot for CRM. The agent calls them as part of the workflow.
A structured output format. Table, JSON, or Markdown brief. Not a wall of text. Structure makes the output reviewable and reusable.
The 30 recipes below all assume this baseline setup. The agentic GTM stack five-layer architecture covers the broader pattern.
ICP Research and Segmentation Workflows (Recipes 1-5)
1. Build a data-backed ICP from closed-won deals.
Prompt: "Read closed_won.csv from the project folder. Identify the top three patterns across firmographics, signals, and tech stack. Output a structured ICP definition I can paste into our CLAUDE.md."
Tools: file read, structured output. No external API calls needed.
2. Find lookalikes for a single best-fit account.
Prompt: "Take this company URL [URL]. Use Databar's company search waterfall to find 50 lookalikes by industry, employee count, and tech stack. Score each by similarity and output as a table."
Tools: Databar MCP (company search waterfall).
3. Segment a CRM list by ICP fit.
Prompt: "Read the candidate list. For each company, score against the ICP definition in CLAUDE.md. Output a table with company, fit_score, and the top three reasons for the score."
Tools: file read, ICP context from CLAUDE.md.
4. Identify ICP gaps not covered by current outbound.
Prompt: "Compare the CRM contact list against our ICP definition. Identify segments where ICP fit is high but contact density is low. Output as a Markdown brief."
Tools: CRM MCP (read), ICP context.
5. Build a regional ICP variation.
Prompt: "Take the base ICP from CLAUDE.md. Adapt it for the EMEA market based on the closed-won deals from EMEA in closed_won_emea.csv. Output the regional ICP variation."
Tools: file read, CLAUDE.md context.
Enrichment and Contact Discovery Workflows (Recipes 6-10)
6. Enrich a list of company URLs with full firmographics.
Prompt: "Read company_urls.csv. For each URL, run Databar's company-data waterfall and write the results to a Databar table named enriched_companies."
Tools: Databar MCP (company-data waterfall, table write).
7. Find decision-maker emails for a company list.
Prompt: "From the enriched_companies table, find decision-makers matching titles in CLAUDE.md (VP Sales, Head of Growth, CRO). Run Databar's email-finding waterfall for each. Write verified emails back to the table."
Tools: Databar MCP (contact-finding waterfall, email waterfall).
8. Enrich LinkedIn URLs with verified emails.
Prompt: "Read linkedin_urls.csv. For each URL, run Databar's LinkedIn-to-email waterfall. Output a table with linkedin_url, name, verified_email, confidence."
Tools: Databar MCP (LinkedIn-to-email waterfall).
9. Find mobile phone numbers for a high-priority list.
Prompt: "From the priority_accounts table, find mobile phone numbers for the named decision-makers. Use Databar's phone waterfall. Flag any results with confidence below 80%."
Tools: Databar MCP (phone waterfall).
10. Enrich tech stack data for ABM targeting.
Prompt: "For each company in target_accounts.csv, run tech stack enrichment via Databar. Flag companies running [target tech]. Output a filtered list."
Tools: Databar MCP (tech stack enrichment).
Sequencing and Copywriting Workflows (Recipes 11-15)
11. Draft a three-email sequence for a single segment.
Prompt: "For the [segment name] from our ICP, draft a three-email outbound sequence. Use the voice rules from CLAUDE.md. First email: pain-led. Second: value-led. Third: case-study-led. Output as Markdown."
Tools: CLAUDE.md voice rules.
12. Personalize first lines for 50 contacts.
Prompt: "Read enriched_contacts.csv. For each contact, write a personalized first line referencing their company's recent funding, hiring, or product launch. Use only verified data, no inferred details. Output as a table with contact_id, first_line."
Tools: file read, enriched data.
13. Write subject line variations for A/B testing.
Prompt: "For the email body in body.md, write five subject line variations. Vary by curiosity, specificity, value-first, contrarian, and direct. Output as a numbered list."
Tools: file read.
14. Translate an email sequence to a new region.
Prompt: "Take the US email sequence in sequence_us.md. Adapt for EMEA: localize the references, adjust tone for the region, and update the social proof. Output as sequence_emea.md."
Tools: file read and write.
15. Generate a follow-up sequence triggered by reply intent.
Prompt: "Read smartlead_replies.csv. Classify each reply as positive, neutral, objection, or negative. For positive and neutral, draft a personalized follow-up. Output as a table."
Tools: Smartlead MCP, classification logic.
Signal-Based Outbound Workflows (Recipes 16-20)
16. Build an outbound campaign triggered by funding announcements.
Prompt: "Find companies that announced funding in the last 14 days matching our ICP. Run Databar's funding signal endpoint, filter by ICP fit, then enrich contacts. Output as a campaign-ready table."
Tools: Databar MCP (funding signals, contact enrichment).
17. Trigger outbound on hiring signals.
Prompt: "Find companies actively hiring [target role] in the last 30 days. Filter by ICP fit. Find hiring-manager emails. Output a campaign list."
Tools: Databar MCP (job posting signals, contact enrichment).
18. Spin up a campaign on tech stack changes.
Prompt: "Find companies that added [target tech] to their stack in the last 60 days. Cross-reference with our CRM to skip existing customers. Output a fresh prospect list."
Tools: Databar MCP (tech stack signals), CRM MCP.
19. Build a campaign on news mentions.
Prompt: "Find companies in our ICP mentioned in the last 7 days for [trigger event]. Filter by company size. Draft personalized outreach referencing the news event."
Tools: Databar MCP (news signals), contact enrichment, copywriting.
20. Trigger on inbound website visits.
Prompt: "Read website_visits.csv (from RB2B or similar). For each company, score by ICP fit, find the right contact via Databar, and draft a relevant first email referencing the visited page."
Tools: file read, Databar MCP, copywriting.
CRM Hygiene and Pipeline Workflows (Recipes 21-25)
21. Identify stale CRM records and refresh.
Prompt: "Pull CRM records older than 90 days with missing or stale fields. Run Databar enrichment to fill gaps. Output a proposed-update table for human review."
Tools: CRM MCP, Databar MCP. The full clean your CRM with an AI agent workflow walks through this in depth.
22. Dedupe contacts across the CRM.
Prompt: "Run fuzzy matching across CRM contacts by email, LinkedIn URL, and name+company. Output merge proposals with confidence scores. Anything above 95% can auto-merge."
Tools: CRM MCP, fuzzy matching logic.
23. Score open opportunities by win likelihood.
Prompt: "Read open_opportunities.csv. Score each by ICP fit, engagement signals (last touch), and stage progression speed. Output a prioritized list with reasoning."
Tools: file read, scoring logic.
24. Surface accounts likely to churn.
Prompt: "Read customer_health_signals.csv. Flag accounts with low engagement, support tickets, or contract end dates within 90 days. Output a save-priority list."
Tools: file read, signal analysis.
25. Update CRM industry tags consistently.
Prompt: "Pull all CRM accounts with no industry tag. For each, look up the canonical industry via Databar's company-data waterfall. Propose industry updates as a table."
Tools: CRM MCP (read), Databar MCP, write proposals.
Reporting and Analytics Workflows (Recipes 26-30)
26. Generate a weekly outbound performance report.
Prompt: "Pull last week's Smartlead campaign data. Calculate open rate, reply rate, meeting rate per segment. Compare to the prior week. Output as a Markdown brief."
Tools: Smartlead MCP.
27. Find which segments converted best last quarter.
Prompt: "Read closed_won_q1.csv. Group by segment and calculate conversion rate, average deal size, and sales cycle length. Output a ranked table."
Tools: file read, aggregation.
28. Identify which messaging angle drove the most replies.
Prompt: "Read sent_emails.csv with reply flags. Group by subject line theme. Calculate reply rate per theme. Output a table ranked by reply rate."
Tools: file read, classification.
29. Build a quarterly business review brief from CRM data.
Prompt: "Read q1_pipeline.csv. Generate a QBR brief covering pipeline by stage, won deals by segment, lost-deal patterns, and three recommendations for Q2. Output as Markdown."
Tools: file read, analysis.
30. Surface where the team is underweighting an ICP segment.
Prompt: "Compare actual outbound volume per segment against ICP-fit-weighted target volume. Flag segments more than 20% below their fit-weighted target. Output as a brief with three suggested rebalances."
Tools: aggregation, ICP context.
How to Use These Claude Code Workflows for GTM
Pick three to five recipes that match your motion. Wire them into your CLAUDE.md and stack. Most teams running Claude Code workflows for GTM see real productivity gains by week two, once they have:
A CLAUDE.md with ICP, voice rules, closed-won patterns, and forbidden phrases
The data layer (Databar) connected via MCP
The CRM (Attio, HubSpot, Salesforce) connected via MCP
The sending tool (Smartlead, Instantly) connected via MCP if outbound is the focus
The recipes scale with the stack. With just file reads and the data layer, you can run 15-20 of these. Add CRM and sending MCPs and you can run all 30. The headless GTM piece walks through the broader pattern of running outbound from a single context window.
FAQ
What are the best Claude Code workflows for GTM teams?
The most useful workflows are short, repeatable prompt patterns that call tools (data layer, CRM, sending tool) and produce structured output. The 30 recipes above cover ICP research, enrichment, sequencing, CRM hygiene, signal-based outbound, and reporting. Most teams pick three to five recipes that match their motion and run them weekly.
What stack do I need for these workflows?
Claude Code as the agent runtime. Databar as the data layer (100+ providers behind one MCP). Your CRM (Attio, HubSpot, or Salesforce) connected via its MCP. Your sending tool (Smartlead, Instantly) connected via MCP if outbound is the focus. Most workflows assume this baseline. The agentic GTM stack five-layer architecture covers the broader pattern.
Do I need to be technical to run these Claude Code workflows for GTM?
You need comfort with the terminal or IDE for the initial setup (writing CLAUDE.md, connecting MCPs). Once running, day-to-day operation is closer to writing prompts than writing code. Many GTM operators with no engineering background run these workflows productively.
How long until I see productivity gains from these workflows?
Most teams hit real gains in week two. Week one is setup (CLAUDE.md, MCP connections, first runs on small batches). Week two onwards is shipping campaigns, hygiene runs, and reporting at a fraction of the manual time. Productivity compounds as your CLAUDE.md gets sharper over time.
Why do these recipes use Databar specifically?
Databar covers 100+ providers behind one MCP, so the recipes work without listing five separate provider tools. Native MCP, SDK, and REST surfaces work with Claude Code out of the box. Outcome-based billing (you only pay when data is successfully returned) fits agent workflows that retry calls. Other data providers work too, but the recipes assume the aggregator pattern.
Can I run these recipes on Cursor or Windsurf instead of Claude Code?
Yes, with minor adjustments. The recipes use MCP-style tool calling, which Cursor and Windsurf both support. The CLAUDE.md file maps to the equivalent context configuration in those tools. The recipes themselves (prompt patterns, tool calls, output formats) work in any MCP-compatible agent runtime.
What's the most underrated Claude Code workflow for GTM?
The CRM hygiene workflow (recipe 21). Most teams treat CRM hygiene as a once-a-quarter project. Running it weekly via an agent keeps the dataset usable for outbound and reporting. Contact data decays at roughly 30% per year, so quarterly hygiene leaves significant staleness in between runs. Weekly hygiene catches changes faster.
Start Running Claude Code Workflows for GTM
Pick three recipes that match your motion. Wire your CLAUDE.md, connect your data layer and CRM, and start.
The data layer is the place to start if you do not have one yet. Databar covers 100+ providers, native MCP and SDK, outcome-based billing where you only pay when data is returned. 14-day free trial at build.databar.ai.
Also interesting
Recent articles
See all
