AI Agent for Sales Outreach: Automate Prospecting and Personalization

TL;DR: A three-agent sales pipeline — Prospector (Gemini, free), Outreach Writer (Claude Sonnet, $0.08), Lead Scorer (Claude Haiku, $0.02) — researches prospects, writes personalized emails, and scores leads for $0.10 per prospect. Here's the setup and real examples.

Sales outreach is repetitive: research the prospect, find their pain points, write a personalized email, follow up, track responses. Most reps spend 4-6 hours daily on research and writing. AI agents can do it in 60 seconds per prospect.

This guide shows you how to set up an AI-powered sales outreach pipeline.

Related: AI for Sales Teams Guide · AI Agent Workflows: 10 Examples · AI Agent Task Board · BYOK AI Platform Comparison · AI Agent Content Writing Workflow

The Three-Agent Sales Pipeline

Agent Roles

Agent	Model	Role	Cost
Prospector	Gemini 2.5 Pro	Research prospect: company, role, recent news, pain points	Free
Writer	Claude Sonnet	Write personalized outreach email based on research	~$0.08
Scorer	Claude Haiku	Score lead quality and suggest follow-up timing	~$0.02

Total per prospect: ~$0.10. Time: 60-90 seconds.

Setup (5 Minutes)

Step 1: Create an Account

Go to ivern.ai/signup. Free, no credit card.

Step 2: Add Your API Key

In Settings, add your Anthropic API key ($5 at console.anthropic.com). BYOK pricing — no markup.

Step 3: Create a Sales Squad

Click Create Squad → name it "Sales Outreach". Add three agents:

Prospector (Gemini 2.5 Pro):

"You are a sales research analyst. Given a prospect's name and company, research and provide: 1) Company overview (size, industry, recent news), 2) Prospect's likely role and responsibilities, 3) Probable pain points based on their industry and role, 4) Recent company changes (funding, hires, product launches), 5) Competitors and market position. Structure findings clearly for an outreach writer."

Writer (Claude Sonnet):

"You are a B2B sales outreach writer. Using the prospect research, write a personalized cold email. Requirements: reference specific company context, lead with their pain point (not your product), keep under 150 words, include one clear CTA, avoid generic buzzwords. Subject line should be 5-7 words and specific to them."

Scorer (Claude Haiku):

"You are a sales manager reviewing outreach quality. Score the email: 1) Personalization (1-10): does it reference specific prospect details? 2) Pain point alignment (1-10): does it address a real problem? 3) CTA clarity (1-10): is the next step obvious? 4) Spam score (1-10): will it land in inbox or spam? Also suggest: best follow-up timing and next touchpoint."

Step 4: Connect and Run

npx @ivern-ai/agent install --key YOUR_IVERN_KEY --provider gemini
npx @ivern-ai/agent install --key YOUR_IVERN_KEY --provider claude

Create a task for each prospect:

"Research and write outreach for: Jane Smith, VP Engineering at TechCorp (Series B, 200 employees, just raised $15M)."

The pipeline runs: Prospector researches → Writer drafts email → Scorer evaluates quality.

Real Output Example

Input

"Prospect: Marcus Chen, Head of Data Science at FinPay (fintech, 150 employees, recently expanded to European markets). Our product: Ivern Squads — AI agent orchestration platform."

Phase 1: Prospector Output (Gemini, free)

Company: FinPay — B2B fintech, ~150 employees, Series B ($28M raised)
Recent: Expanded to EU markets Q1 2026, hiring 12 data engineers
Prospect: Marcus Chen, Head of Data Science — manages team of 8
Likely pain points:
  - Scaling data pipelines across multiple markets
  - EU compliance (GDPR) adding complexity to data workflows
  - Team overloaded with manual data processing
  - Need to automate without hiring 12 engineers immediately
Competitors: Stripe (larger), Adyen (EU-native), Checkout.com
Trigger event: EU expansion means new regulatory data requirements

Phase 2: Writer Output (Claude Sonnet, $0.08)

Subject: EU compliance slowing down your data team?

Hi Marcus,

Scaling FinPay's data pipelines across European markets while keeping 
GDPR compliance — that's a lot of overhead for an 8-person team hiring 
12 more engineers.

Teams in your position use Ivern Squads to set up automated data 
workflows in minutes. Your data team defines the pipeline, AI agents 
execute the repetitive work — extraction, validation, compliance 
checks. You keep full control with your own API keys, no data leaves 
your infrastructure.

Happy to show you a 5-minute demo of an automated compliance pipeline.

Best,
[Name]

Phase 3: Scorer Output (Claude Haiku, $0.02)

Personalization: 9/10 — References FinPay, EU expansion, team size, GDPR
Pain point alignment: 9/10 — Directly addresses scaling + compliance tension
CTA clarity: 8/10 — Clear demo offer, could be more specific ("Tuesday at 2pm?")
Spam score: 8/10 — Should land in inbox. Avoids ALL CAPS and exclamation marks.

Follow-up: Send in 3 business days. Reference the hiring — "Saw FinPay is hiring 
12 data engineers. We've helped teams like yours defer 3-4 of those hires with 
automation."

Score: 8.5/10 — Send.

Total: 45 seconds, $0.10 for a fully personalized, scored outreach email.

Cost Comparison

Per Prospect Cost

Approach	Cost/Prospect	Personalization	Time
Manual research + writing	$25-50 (rep time)	Good	30-60 min
Apollo/ZoomInfo template	$0.05-0.50 (subscription)	Low	5 min
AI agent pipeline	$0.10	High	1 min

Monthly Cost by Volume

Prospects/Week	Monthly Cost (AI Pipeline)	Compare to Rep Time
25	~$10	~$5,000
50	~$20	~$10,000
100	~$40	~$20,000
200	~$80	~$40,000

The AI pipeline produces personalized outreach at 1/200th the cost of manual research.

Batch Processing: 50 Prospects at Once

For batch outreach, create a CSV with prospect details:

Name,Company,Role,Industry,Employees,Trigger
Marcus Chen,FinPay,Head of Data Science,Fintech,150,EU expansion
Sarah Kim,CloudScale,VP Engineering,Cloud,300,Series C
David Park,MedAI,CTO,Healthcare,80,FDA approval
...

Upload as a batch task to the pipeline. Each prospect gets individualized research, writing, and scoring. The Prospector runs free research for all 50, then the Writer and Scorer process each one.

Batch of 50: ~$4 total, 30-40 minutes.

Workflow Integration

CRM Integration

After the pipeline produces emails, add them to your CRM:

1. Export scored emails from Ivern dashboard
2. Import to HubSpot/Salesforce/Pipedrive
3. Schedule sends based on Scorer's timing recommendation

Follow-Up Automation

Create a second pipeline task for follow-ups:

"The prospect [name] at [company] hasn't responded to our initial email about [topic]. Write a follow-up that references the initial email and adds one new piece of value. Keep under 100 words."

Response Handling

When a prospect replies, create a task:

"Prospect [name] replied: '[their response]'. Based on their reply, determine: interest level (hot/warm/cold), objections raised, next step to propose. Draft a response."

What This Workflow Does NOT Do

Be clear about limitations:

• Does not send emails automatically — the pipeline produces drafts for human review
• Does not guarantee responses — personalization improves open rates but deliverability depends on your email infrastructure
• Does not replace relationship building — use AI for initial outreach, humans for closing
• Does not access private data — research uses publicly available information only

Frequently Asked Questions

Is AI-generated outreach effective?

Personalized outreach gets 2-3x higher response rates than templates. AI agents produce genuine personalization by researching each prospect individually. The key is the research phase — that's what separates this from mail merge.

How many prospects can I process daily?

Limited by your API rate limits. With standard Anthropic limits: ~200-500 prospects per day. With higher limits: 1000+.

What about spam filters?

The Scorer agent checks for spam triggers (ALL CAPS, excessive exclamation marks, generic subject lines). Emails scoring below 7/10 on spam score get flagged for revision before sending.

Can I customize the outreach style?

Yes. Modify the Writer's system prompt to match your brand voice. Include examples of emails that have worked for your team. The more specific the prompt, the more the output matches your style.

Does this work for B2C?

The pipeline is designed for B2B outreach where personalization matters. For B2C, you'd typically use broader campaigns rather than individual personalization.

What if the prospect info is wrong?

The Prospector uses publicly available data, which can be outdated. Always review the research before sending. The human review step (before clicking "send") catches most inaccuracies.

Get Started

1. Sign up free at ivern.ai/signup
2. Add your Anthropic API key ($5 minimum)
3. Create a Sales Outreach squad (Prospector + Writer + Scorer)
4. Enter your first prospect — name, company, role
5. Get a personalized, scored outreach email in 60 seconds

Stop spending hours on prospect research. Start automating the grunt work.

Set up your AI sales outreach pipeline →