B2B SDR Agent

What does this agent do, and when should you use it?

B2B SDR Agent is an OpenClaw-based template for export businesses handling prospects through WhatsApp, Telegram, and email. Its operating context is split across IDENTITY.md, SOUL.md, AGENTS.md, USER.md, HEARTBEAT.md, MEMORY.md, and TOOLS.md, with a ten-stage workflow spanning lead capture through multi-channel orchestration. It creates CRM records, conducts BANT qualification, enriches companies, produces multilingual PDF quotations, tracks negotiations, and schedules pipeline reporting and nurture activity. Memory is divided among MemOS extraction, proactive summarization, customer-isolated ChromaDB turn storage, and daily CRM snapshots. It can be installed as an OpenClaw skill through ClawHub or deployed as a service on a Linux server, where OpenClaw Gateway connects the channels and exposes a token-protected control dashboard. The code is MIT-licensed, while a separate managed PulseAgent offering is also promoted.

On inbound WhatsApp, Telegram, CTWA-ad, or email activity, the workflow identifies a lead and builds a CRM record containing fields such as name, company, country, product interest, and ICP score. It conducts BANT qualification in conversation and can run a three-layer enrichment process using Jina AI web search and company-site analysis; high-scoring leads are flagged for the owner. quotation-generator reads product specifications, MOQ, and lead times from the product knowledge base to create multilingual PDF proforma invoices for owner approval. delivery-queue and sdr-humanizer govern message splitting, 3–90 second delays, timezone handling, and cultural adaptation, while Day 1/3/7/14 email sequences manage follow-up. MemOS extracts BANT facts, commitments, and objections; ChromaDB stores each turn under customer_id; proactive summarization starts at 65% context use; and a daily CRM snapshot supplies disaster recovery. Scheduled checks also cover Gmail replies, pipeline reports, lead discovery, stalled opportunities, competitor intelligence, memory health, Teams mentions, and switching to Telegram when a WhatsApp 72-hour window expires.

1. An exporter receiving substantial WhatsApp inquiry volume wants automatic lead records, qualification, and rapid owner alerts.
2. A sales team needs multilingual PDF proforma invoices generated from governed specifications, MOQ, and lead-time data, with human approval.
3. A company selling through WhatsApp, Telegram, and email wants coordinated follow-ups and channel switching after the WhatsApp window expires.
4. A long-cycle B2B team needs quotations, objections, promises, and customer state preserved across sessions and system restarts.
5. A service operator hosting agents for several customers needs separate workspaces, configurable remote paths, and per-tenant WhatsApp exit IPs.

How do you install or deploy this agent?

For an existing OpenClaw installation, install the skill from ClawHub:

clawhub install b2b-sdr-agent

Then edit the company identity and customer profile, optionally copying all seven context files into the main workspace:

vim ~/.openclaw/workspace/skills/b2b-sdr-agent/references/IDENTITY.md
vim ~/.openclaw/workspace/skills/b2b-sdr-agent/references/USER.md
cp ~/.openclaw/workspace/skills/b2b-sdr-agent/references/*.md ~/.openclaw/workspace/

A full deployment requires a Linux server, Node.js 18+, and an API key for at least one supported model provider; a WhatsApp Business account is optional but recommended. Clone the repository and replace the workspace placeholders with business information:

git clone https://github.com/iPythoning/b2b-sdr-agent-template.git
cd b2b-sdr-agent-template
vim workspace/IDENTITY.md
vim workspace/USER.md
vim workspace/SOUL.md

Create the deployment configuration and supply the server address, PRIMARY_API_KEY, channel credentials, Google Sheets CRM identifier, and administrator phone details as applicable:

cd deploy
cp config.sh.example config.sh
vim config.sh

Validate the template before deployment:

cd ..
npm test

Deploy an instance named my-company and run the remote health checks:

cd deploy
./deploy.sh my-company
./doctor.sh

How do you use this agent?

Populate workspace/IDENTITY.md, USER.md, and SOUL.md with the company, catalog, ICP, competitors, and conversation policies. Maintain product specifications, MOQ, and lead times in product-kb/catalog.json and the product info.json files. Once the model and channel credentials are configured and OpenClaw Gateway is running, inbound WhatsApp, Telegram, and email conversations enter the ten-stage workflow. Open the dashboard using the address and token printed by deployment:

http://YOUR_SERVER_IP:18789/?token=YOUR_GATEWAY_TOKEN

The dashboard exposes bot state, WhatsApp connectivity, conversation threads, cron execution, and channel health, so its URL and token must remain private. To prevent network access to it, configure deploy/config.sh as follows:

GATEWAY_BIND="loopback"

WhatsApp direct messages can use open, allowlist, or pairing policy. The documented default accepts every contact:

{
  "channels": {
    "whatsapp": {
      "dmPolicy": "open",
      "allowFrom": ["*"]
    }
  }
}

Restart the gateway after changing the live configuration:

systemctl --user restart openclaw-gateway

For a multi-tenant server, assign an instance a dedicated WhatsApp exit IP:

./deploy/ip-isolate.sh acme-corp

What are this agent's strengths and limitations?

How does this agent compare with similar options?

Key facts side by side with the most closely related agents.

How does FollowAgents rate this agent?

FAQ