Use Cases
Sales Development
goClaw as an AI SDR — CRM-integrated lead qualification, follow-up sequencing, and context-aware engagement.
A goClaw SDR (Sales Development Representative) agent handles the top of your sales funnel autonomously. It takes leads from multiple sources, qualifies them against your ICP, runs personalized outreach sequences, and hands off warm leads to your human sales team.
What makes goClaw different from traditional SDR tools
| Traditional SDR tools | goClaw |
|---|---|
| You build email sequences in a drag-and-drop editor | Agent writes each message based on accumulated context |
| Fixed templates with merge fields | Genuine personalization from prospect research |
| Rules-based branching | Agent makes judgment calls based on conversation |
| Static knowledge base | Self-evolving: files and fills its own knowledge gaps |
| You maintain the tool | Agent maintains its own skills and procedures |
Lead sources
goClaw accepts leads from multiple sources:
Webhook inbound
Your lead generation tools send a webhook when a new lead arrives:
// From Clay, Apollo, LinkedIn, your website form, etc.
POST /api/webhooks/new_lead
{
"email": "alex@startup.io",
"name": "Alex Kim",
"company": "Startup.io",
"title": "Head of Growth",
"source": "apollo_export",
"icp_score": 0.78
}
The agent picks up the lead, creates a contact record, and kicks off the outreach sequence.
CSV import
npx @clawrm/cli import --file leads.csv --group outbound_sales --start-sequence
Admin dashboard
Paste a list of emails or LinkedIn URLs directly into the dashboard. The agent creates contacts and starts research.
The SDR sequence
A standard SDR sequence with goClaw:
Stage 1 — Research (Day 0, before first touch)
Agent research run:
1. web_search: "{company} SaaS product 2026"
2. web_search: "{name} {title} {company} background"
3. web_search: "{company} funding news recent"
4. knowledge_search: "ICP signals {industry}"
5. Evaluate: does this prospect match ICP?
→ ICP score 0.8+: proceed with personalized outreach
→ ICP score 0.5–0.8: proceed with lighter touch
→ ICP score < 0.5: archive contact, tag "not_qualified"
6. knowledge_file_curiosity: "What products does {company} sell?"
(queued for nightly research pipeline)
Stage 2 — First touch (Day 0)
The agent writes a personalized email that demonstrates research:
"Hi Alex,
Noticed Startup.io raised a seed round last quarter — congrats. The growth from 12 to 45 employees you mentioned on the podcast usually means the point when manual GTM processes start breaking.
We help Head of Growth teams at companies your size replace their outbound stack with a single agent that researches, reaches out, and follows up autonomously.
Worth a 20-minute call to see if it fits?
Jordan (AI SDR, Acme)"
Stage 3 — Follow-up sequence (Days 3, 7, 14)
Each follow-up is written fresh — not a template. The agent references:
- Previous messages in the thread
- Any new research filed since last touch
- Seasonal or news-based angles
Stage 4 — Response handling
| Response type | Agent action |
|---|---|
| Interested ("let's talk") | Flag as hot lead, notify human team, send calendar link |
| Objection ("we already have X") | Use objection knowledge, respond with reframe |
| Timing ("reach out in Q3") | Schedule reactivation task, archive for now |
| Opt-out ("remove me") | Unsubscribe, tag contact "opted_out", stop all outreach |
| No response after 3 touches | Archive, schedule reactivation check in 45 days |
Stage 5 — Warm lead handoff
When a lead shows interest:
- Agent sends calendar link or asks for availability
- Agent creates a briefing note in the CRM:
# Lead Briefing: Alex Kim, Startup.io
## Qualification
- Company: Startup.io, 45 employees, Series Seed
- Role: Head of Growth — direct decision maker for outbound stack
- Pain: Manual GTM process post-fundraise
- ICP fit: 0.87 (high)
## Conversation summary
Responded to follow-up #1 (Day 3). Interested in understanding how we handle
multi-channel. Asked specifically about Telegram support.
## Recommended talking points
1. Multi-channel: email + SMS + Telegram out of the box
2. Self-learning knowledge base — mention 60 files in 30 days stat
3. Competitive: vs. Clay (they're probably already using it)
## Notes
Alex mentioned they're currently using Apollo for prospecting. Potential
expansion play: replace both Apollo + outreach tooling with goClaw.
- Agent notifies human sales rep via Slack with briefing summary
Performance benchmarks (internal dogfood)
Based on Indigo's internal deployment:
| Metric | Value |
|---|---|
| Research time per prospect | ~45 seconds |
| Email open rate (personalized) | 38% |
| Reply rate (personalized) | 12% |
| Positive reply rate | 5% |
| Leads qualified per day | 50–200 (depends on tier) |
| Average follow-up touches before response | 2.1 |
These benchmarks reflect goClaw's own internal GTM operation. Results will vary by industry, persona, and ICP precision.
Configuration tips
Be specific in your ICP. The more precise your ICP knowledge file, the better the agent filters. A vague ICP leads to low-quality outreach at scale.
Seed the objection knowledge. The most common objections (price, timing, competitor) should be in the knowledge base before you start. The agent will add more from live conversations.
Set appropriate batch sizes. On the Starter tier (5M tokens/month), you can research ~300 prospects and send ~1,000 emails/month comfortably within baseline. Scale up with the Growth or Scale tier.
Monitor the first 50 leads manually. Review the first month's outreach in the admin inbox before going fully autonomous. Adjust persona, ICP, and objection knowledge based on what you see.