Cold Emails That Feel Warm and Personal

AI researches each prospect and crafts individually tailored messages. You review every email before it sends to protect your reputation.


Why this matters

Generic cold emails get less than 1% response rates and damage sender reputation. Personalized outreach works dramatically better but takes 15-20 minutes per email to research the prospect, find a relevant hook, and compose a natural-sounding message. This makes quality outreach at scale impossible for most teams.


How MultiMail solves this

MultiMail's AI agent researches each prospect, finds personalization angles, and drafts individually crafted cold emails that feel hand-written. With gated_all oversight, sales reviews every element — recipient, subject, body — before any email sends, preventing wrong-person or wrong-company errors that damage sender reputation.

1

Research the Prospect

Your AI agent gathers context about each prospect — company, role, recent activity, public content, and potential needs — to find genuine personalization angles.

2

Craft Individual Messages

Using send_email, the agent drafts a unique email for each prospect that references specific, relevant details. No templates — each message feels individually written.

3

Human Reviews Everything

With gated_all oversight, every element is reviewed: Is this the right person? Is the company research accurate? Does the message tone match our brand? Sales approves before any email sends.

4

Track Engagement

The agent monitors check_inbox for replies and tracks which personalization approaches generate the highest response rates.


Implementation

Send a Personalized Cold Email
python
import requests

API = "https://api.multimail.dev/v1"
HEADERS = {"Authorization": "Bearer mm_live_xxx"}

response = requests.post(
    f"{API}/send",
    headers=HEADERS,
    json={
        "from": "[email protected]",
        "to": ["[email protected]"],
        "subject": "Saw your talk on AI agents at PyCon - quick question",
        "text_body": (
            "Hi Sarah,\n\n"
            "I caught your PyCon talk on building production AI agents. "
            "The part about email handling challenges really resonated — "
            "we see the same issues with our customers.\n\n"
            "We built MultiMail specifically for that problem: giving "
            "AI agents email capabilities with human oversight controls. "
            "Your mention of needing &"cm">#039;guardrails on outbound sends' is "
            "exactly what our gated_send mode does.\n\n"
            "Would you be open to a 15-minute call to see if it could "
            "help your agent project?\n\n"
            "Best,\nYour Name"
        ),
        "html_body": "<p>Hi Sarah,</p><p>I caught your PyCon talk...</p>"
    }
)
print(f"Cold email queued for review: {response.json()[&"cm">#039;id']}")

Draft a deeply personalized cold outreach email queued for review.

Manage Prospect Contacts
python
import requests

API = "https://api.multimail.dev/v1"
HEADERS = {"Authorization": "Bearer mm_live_xxx"}

"cm"># Add prospect with research context
response = requests.post(
    f"{API}/contacts",
    headers=HEADERS,
    json={
        "email": "[email protected]",
        "name": "Sarah Chen",
        "metadata": {
            "company": "Acme Corp",
            "role": "Head of AI Platform",
            "source": "PyCon 2026 speaker",
            "hook": "Talked about email handling challenges in agents",
            "status": "cold-prospect"
        }
    }
)

"cm"># Search for prospects by source
prospects = requests.get(
    f"{API}/contacts/search",
    headers=HEADERS,
    params={"query": "PyCon cold-prospect"}
).json()
print(f"PyCon prospects: {len(prospects[&"cm">#039;results'])}")

Add prospects with research data for personalized outreach.

MCP Tool: Cold Outreach Campaign
typescript
"cm">// Add prospect with research data
await mcp.add_contact({
  email: "[email protected]",
  name: "Sarah Chen",
  metadata: {
    company: "Acme Corp",
    role: "Head of AI Platform",
    source: "PyCon 2026",
    hook: "Email handling challenges in agents talk"
  }
});

"cm">// Send personalized outreach (gated_all: full review)
const result = await mcp.send_email({
  to: "[email protected]",
  subject: "Saw your PyCon talk on AI agents - quick question",
  text_body: "Hi Sarah, I caught your talk on building production agents..."
});

await mcp.tag_email({
  email_id: result.id,
  tags: ["cold-outreach", "pycon-campaign", "first-touch"]
});

"cm">// Check for replies
const inbox = await mcp.check_inbox({
  mailbox_id: "sales-mailbox-id",
  status: "unread"
});

Use MCP tools to manage prospect research and send personalized outreach.


What you get

5-10x Higher Response Rates

Deeply personalized emails that reference specific, relevant details consistently outperform template-based outreach by 5-10x.

Reputation Protection

Gated all oversight means every email is reviewed before sending. Wrong-person errors, factual mistakes, and off-brand messaging are caught before they damage your sender reputation.

Scale Without Sacrificing Quality

AI handles the time-intensive research and drafting. Sales reps review in seconds what would take 15 minutes to create from scratch.

Compliance Built In

Every cold email includes required CAN-SPAM elements. The AI agent handles unsubscribe links, physical address requirements, and opt-out processing.


Recommended oversight mode

Recommended
gated_all
Cold outreach carries the highest reputation risk of any email type. Wrong-person errors, inaccurate company research, or tone-deaf messaging can permanently damage your sender reputation and brand. Gated all ensures human review of every element before delivery.

Common questions

Why gated_all instead of gated_send?
Cold outreach requires reviewing the recipient as much as the content. Gated all lets sales verify that the AI targeted the right person at the right company before sending. A great email sent to the wrong person is worse than no email at all.
How does the AI find personalization angles?
Your AI agent researches prospects using public data — conference talks, blog posts, LinkedIn activity, company announcements. It identifies genuine connection points that make the email feel hand-written rather than templated.
How do we handle GDPR and CAN-SPAM compliance?
MultiMail includes unsubscribe links and honors opt-out requests automatically. For GDPR compliance, your AI agent should verify legitimate interest basis before outreach. For CASL (Canada), explicit consent may be required. The compliance tags in your contact records help the agent enforce these rules.
What response rate should we expect?
Well-researched, personalized cold emails typically see 8-15% response rates compared to 1-2% for template-based outreach. The key is genuine personalization — not just inserting a company name into a template, but referencing specific, relevant details about the prospect's work.

Explore more use cases

The only agent email with a verifiable sender

Email infrastructure built for AI agents. Verifiable identity, graduated oversight, and a 38-tool MCP server. Formally verified in Lean 4.