How to Automate Client Onboarding with AI (Complete Python Implementation)

#ai #python #saas #automation

Client onboarding is one of the biggest time sinks in consulting. Here's how I automated 80% of it.

The Problem

Every new client required:

Welcome email + access provisioning (30 min)
Kickoff questionnaire review (45 min)
Project setup in PM tools (20 min)
Initial research + briefing document (2 hours)
Calendar scheduling for kickoff call (15 min back-and-forth)

Total: ~4 hours per new client.

With 3-4 new clients per month, that's 12-16 hours of repetitive onboarding.

The Automated Pipeline

import anthropic
from datetime import datetime

client = anthropic.Anthropic()

def onboard_client(client_info: dict) -> dict:
    """Complete client onboarding pipeline."""

    # Step 1: Generate personalized welcome package
    welcome = client.messages.create(
        model="claude-sonnet-4-20250514",
        max_tokens=2048,
        messages=[{
            "role": "user",
            "content": f"""Create a personalized client welcome package:

Client: {client_info['company']}
Contact: {client_info['name']}
Industry: {client_info['industry']}
Project: {client_info['project_type']}
Budget: {client_info['budget']}

Include:
1. Personalized welcome email (reference their industry)
2. What to expect (timeline, milestones, communication)
3. Pre-kickoff questionnaire (10 questions specific to their project)
4. Access checklist (what we need from them)
5. Quick-start guide for our PM tool"""
        }]
    )

    # Step 2: Research the company
    research = client.messages.create(
        model="claude-sonnet-4-20250514",
        max_tokens=2048,
        messages=[{
            "role": "user",
            "content": f"""Research brief for {client_info['company']}:
- Company overview (size, revenue, products)
- Industry trends relevant to their AI project
- 3 specific automation opportunities based on their industry
- Competitive landscape
- Key metrics to track"""
        }]
    )

    # Step 3: Generate project plan
    plan = client.messages.create(
        model="claude-sonnet-4-20250514",
        max_tokens=2048,
        messages=[{
            "role": "user",
            "content": f"""Create a project plan for:
Company: {client_info['company']}
Project: {client_info['project_type']}
Budget: {client_info['budget']}
Timeline: {client_info.get('timeline', '4 weeks')}

Include:
- Week-by-week breakdown
- Deliverables per phase
- Resource requirements
- Risk factors
- Success metrics"""
        }]
    )

    return {
        "welcome_package": welcome.content[0].text,
        "research_brief": research.content[0].text,
        "project_plan": plan.content[0].text,
        "created_at": datetime.now().isoformat()
    }

Results

Step	Before	After
Welcome email	30 min	2 min (review)
Questionnaire	45 min	5 min (customize)
Project setup	20 min	3 min
Research brief	2 hours	8 min
Scheduling	15 min	0 (Calendly)
Total	4 hours	18 minutes

92% time reduction per client onboarding.

Implementation Tips

Template the prompts, customize the inputs — same pipeline, different client data
Always add a human review step — AI generates, you approve
Save outputs to your CRM — builds institutional knowledge
Iterate the prompts weekly — each client teaches you what works better

This is one of 30 automation blueprints with complete code at wedgemethod.gumroad.com/l/ai-automation-playbook-smb.