How to Get Warm Slack Leads (Using Claude Code)

TL;DR

• A warm Slack lead is a member of your own community who posts often and fits your buyer profile. Claude Code finds the first group, LeadOcean checks the second.
• It takes about 30 minutes. It spends up to 55 of your 1,000 free records: 40 to enrich the top members and up to 15 to find their colleagues.
• Everything runs on files you already control: a member list and a public-channel export from a workspace you own.

Prerequisites

• A Slack workspace you own or admin. A Workspace Owner can export public-channel messages on any plan, September 2026.
• A Slack app token with the users:read and users:read.email scopes, in your shell as $SLACK_TOKEN. Slack's docs say the email field needs users:read.email, September 2026.
• A LeadOcean API key from app.leadocean.io, as $LEADOCEAN_API_KEY. The free plan has no card and 1,000 records, one-off.
• Claude Code, plus curl and jq.

export SLACK_TOKEN="paste-your-slack-token-here"
export LEADOCEAN_API_KEY="paste-your-key-here"

Step 1: Pull the member list with emails

Slack's users.list returns every member, including invited and deactivated users, September 2026. Drop deleted accounts and bots, and keep the email.

curl -s "https://slack.com/api/users.list?limit=200" \
  -H "Authorization: Bearer $SLACK_TOKEN" |
jq -r '.members[]
  | select(.deleted == false and .is_bot == false and .profile.email != null)
  | [.id, .profile.real_name, .profile.email] | @csv' > members.csv

Slack pages the list. If response_metadata.next_cursor is not empty, repeat the call with &cursor=<that value> and append to members.csv.

Guests have is_restricted or is_ultra_restricted set to true. Decide now whether guests count as leads. Most teams drop them.

Step 2: Export public-channel messages

In Slack, open Admin, then Workspace settings, then Security, then Import & export data. Pick a date range on the Export tab and start the export. Slack emails a ZIP when it is ready.

Slack says each channel gets its own folder with one JSON file per date, September 2026. Every message carries type, user, text and ts. Unzip it next to members.csv.

unzip slack-export.zip -d export
ls export | head

Use public channels only. Do not export direct messages to score leads. Slack also notes that a retention policy can delete old messages, so a date range older than your retention window returns no channel folders.

Step 3: Score activity with Claude Code

Count messages per member over the last 90 days. This is a job for a script, not a chat, so tell Claude Code to write one and show it to you first.

In ./export each public channel is a folder of date files. Each file holds messages
with the fields type, user, text and ts. Write score.sh that counts messages per
user ID from the last 90 days, ignoring messages without a user. Join the counts to
members.csv on user ID and write warm.csv with columns id, name, email, messages.
Drop anyone whose email domain is mine (acme.com). Keep the top 40 by messages.
Show me the script before you run it, then run it.

The core of what it should write looks like this. Check one date file with jq 'type' first, so you know whether it is a list.

find export -mindepth 2 -name '*.json' -exec cat {} + |
jq -r '.[] | select(.type == "message" and .user != null)
  | select((.ts | tonumber) > (now - 90 * 86400)) | .user' |
sort | uniq -c | sort -rn > counts.txt

Messages are a crude signal. A member who asks three questions about your product is warmer than one who posts ten jokes. Ask Claude Code to print the three most recent message texts per member, then judge for yourself.

Step 4: Enrich the top members by email

You already hold the email, so enrich to learn the title and company. POST /v1/people/enrich takes exactly one of linkedin_url, email, phone or person_id. It costs one record per match. A member we hold nothing for is a 404 and costs nothing.

Test on five rows before the full run. The script writes one CSV row per member and skips any member that returns a 404.

echo "id,email,messages,name,job_title,company,domain,employees" > enriched.csv
tail -n +2 warm.csv | head -5 | tr -d '"' | while IFS=, read -r id name email messages; do
  curl -s -X POST https://api.leadocean.io/v1/people/enrich \
    -H "x-api-key: $LEADOCEAN_API_KEY" -H "Content-Type: application/json" \
    -d "{\"email\":\"$email\"}" |
  jq -r --arg id "$id" --arg email "$email" --arg m "$messages" '
    select(.data != null) | .data as $d | $d.contact_data.contact_current_experiences[0] as $e
    | [$id, $email, $m, $d.profile_data.profile_full_name, $e.job_title, $e.company_name,
       $e.company_domain, $e.company_employees.number_of_employees] | @csv' >> enriched.csv
done

warm.csv has a header row, so tail -n +2 skips it. A member name with a comma breaks the split. If you have any, ask Claude Code to parse the CSV with Python instead.

Look at the five rows in enriched.csv. If the titles match the people you know, run the other 35. Replace head -5 with sed -n '6,40p' and drop the echo line, so the header is not rewritten. Leave reveal_email off: you do not need an address you already have.

Step 5: Keep the members who match your buyer

Ask Claude Code to filter enriched.csv against your buyer, and to explain every drop.

From enriched.csv keep rows where job_title is a VP, Head, Director or founder in sales,
marketing or revenue, and employees is 51 to 500. Write buyers.csv sorted by
messages. List each dropped row with the reason. Do not guess a title that is empty.

The kept rows are your warm leads. Expect a short list. Many active members are peers, students and vendors, not buyers. Open each member's last messages before you write to them, and mention what they asked. A reply to a real question is the only reason this list is warm.

Step 6: Find colleagues at the warm accounts

One engaged person is a signal about their company. GET /v1/companies/{domain}/people returns the people we hold there, one record per person returned. Cap it with limit.

curl -s "https://api.leadocean.io/v1/companies/acme.com/people?limit=3" \
  -H "x-api-key: $LEADOCEAN_API_KEY" | jq '.data | length'

Run it for your top five accounts at limit=3: at most 15 records. The rows are thin, with no addresses. Enrich a person by person_id only after you pick them.

To see how many more buyers like yours exist outside your community, size them for free. Add the MCP server, sign in with /mcp, then ask for a count.

claude mcp add --transport http leadocean https://api.leadocean.io/mcp

Use leadocean_count_leads with jobLevel ["VP","Director"], jobFunction
["Sales & Business Development"], industry ["Software Development"], employeeRange
["51-200"], country ["US"], facets false. Report the total and say which email
statuses it counted. Do not call any tool that spends records.

What you get

The run ends with buyers.csv: one row per warm member, with their message count, title and company. This is the shape, with placeholder values.

name,email,messages,job_title,company,domain,employees
Jane Doe,jane@acme.com,31,VP Sales,Acme,acme.com,120

The enrich call returns this shape. It is the response shape from the API reference, with placeholder values, and carries no email addresses because reveal_email is off.

{
  "success": true,
  "data": {
    "profile_data": { "profile_full_name": "Jane Doe" },
    "contact_data": {
      "has_email": true,
      "email_status": "verified",
      "contact_current_experiences": [{
        "job_title": "VP Sales",
        "company_name": "Acme",
        "company_domain": "acme.com",
        "company_employees": { "number_of_employees": 120 }
      }]
    }
  }
}

The free count in Step 6 returned this on 1 October 2026:

{"total":4665,"totalIsExact":true,"credits":0,"defaults":"Counted only addresses worth sending to (emailStatus: verified or catch_all_valid or catch_all)"}

That is 4,665 mailable VPs and Directors in sales at US software companies with 51 to 200 staff. It is a mailable count, not everyone matching the other filters. Your number will differ as the data changes.

Troubleshooting

FAQ