Success OS
Your AI-powered customer success team.
Success OS isn't a tool you operate — it's a proactive CS team that tells you what to do. Like having a Head of Customer Success who wakes up, checks your users, and tells you what needs attention.
Price: $99 (one-time)
Quick Start
npx @aiorg/cli init success-os ~/my-project
cd ~/my-project
claude
> /setup
Overview
| Traditional CS Tools | Success OS |
|---|---|
| "156 users inactive" | "Here's who needs attention and why" |
| Generic re-engagement emails | Personalized, context-aware outreach |
| Same playbook for everyone | Adapts to your stage and data |
| You figure out the strategy | Claude recommends, you decide |
Important distinction:
- If users leave before activating → PMF OS
- If users leave after activating → Success OS
Success OS assumes users have experienced value at least once.
When to Use Success OS
Use Success OS if:
- Users activated but are churning later
- You want to reduce churn
- You need to identify at-risk users
- You want to build engagement campaigns
Don't use Success OS if:
- Users leave before activating → Use PMF OS
- You need more users → Use Marketing OS
- You haven't launched yet → Use Idea OS
What's Included
Commands
Setup & Discovery
| Command | Description |
|---|---|
/setup | Connect to project |
/discover | Re-analyze project |
/export | Generate export script |
/import | Import contacts |
/learn | Answer questions to help Claude learn |
Analysis & Recommendations
| Command | Description |
|---|---|
/briefing | Today's CS briefing (default on session start) |
/health | Detailed user health analysis |
/signals | What signals is Claude seeing? |
/cohorts | Cohort analysis |
Actions
| Command | Description |
|---|---|
/draft [user/segment] | Draft email for specific user or segment |
/send | Send approved emails |
/ask [question] | Ask users a question (for learning) |
Configuration
| Command | Description |
|---|---|
/strategy | View/update strategy for your business |
/focus [goal] | Change current focus |
/stage [stage] | Update company stage |
Proactive Briefings
Every session, Success OS gives you a briefing — like a morning standup with your CS team:
Good morning! Here's your customer success briefing:
🚨 NEEDS ATTENTION
├── 5 trial users expire tomorrow
└── Power user John hasn't logged in for 10 days (unusual)
📊 THIS WEEK
├── 23 users went inactive (up from 15 last week)
├── 8 new signups
└── 2 users upgraded
💡 MY RECOMMENDATIONS
1. [URGENT] Re-activate 5 expiring trials
→ I've drafted personalized emails. Want to review?
2. [HIGH] Check on John (john@company.com)
→ This is unusual for him. Personal email?
3. [MEDIUM] Ask the 2 upgraded users for testimonials
→ Perfect timing - they just upgraded
What should we focus on?
Stage-Aware Strategy
Success OS adapts to your stage:
< 100 users (Early Stage)
- Focus: Manual, high-touch, LEARNING
- "With 50 users, let me help you write personal emails"
- "Let's understand WHY users churn before automating"
100-1000 users (Growing)
- Focus: Finding patterns, light automation
- "I see a pattern: users who don't use X churn within 7 days"
- "Let's create your first automated segment"
> 1000 users (Scale)
- Focus: Automation, playbooks, efficiency
- "Time to set up automated triggers"
- "Let's build formal playbooks for your team"
User Health Scoring
For each user, Success OS calculates a health score:
Health Score = weighted average of:
├── Recency (when last active) - 40%
├── Frequency (how often active) - 30%
├── Feature usage (core feature used?) - 20%
└── Plan/stage (trial, free, paid) - 10%
This lets Claude say: "These 23 users have health score < 30 and need attention."
Signal Detection
Success OS watches for:
Negative Signals:
- Usage drop (was daily, now weekly)
- Feature abandonment
- Support tickets (frustration)
- Login gap (unusual inactivity)
Positive Signals:
- Usage spike (exploring features)
- Invite sent (expansion signal)
- Settings changed (investment)
- Support praise (testimonial candidate)
Data Structure
Success OS stores data in .successos/ in your project:
.successos/
├── project.json # Connected project
├── strategy.json # Strategy for your business
├── product-context.json # What Claude knows about your product
├── learnings.json # What Claude learned from feedback
├── contacts.db # Your users
├── health-scores.json # Calculated health scores
├── signals.json # Detected signals
└── sent-log.json # What Claude sent
PMF-Aware Routing
Success OS checks your PMF status on startup:
- If
pmf.status = "searching"→ Suggests PMF OS - If
pmf.activationRate < 30%→ Warns about activation problem - If
pmf.status = "achieved"→ Full Success OS features
vs. Customer.io / Intercom
They're automation tools. You set up flows, they execute.
Success OS is a thinking CS team:
- Tells you what to do (doesn't wait for configuration)
- Adapts to your stage
- Learns from results
- Asks questions to get smarter
vs. Hiring a CS Person
Success OS is:
- Always available (24/7)
- Instantly knowledgeable (reads your code)
- Consistent quality every day
- $99 one-time, not $5k/month
But it can't:
- Have real conversations (you handle calls)
- Make judgment calls (you approve everything)
- Build relationships (you're the face)
Success OS is your CS team's brain, you're the face.
Related Kits
Before:
- PMF OS — Make sure users are activating first
- Marketing OS — Get users to retain
After:
- Continue with Success OS for ongoing retention
- Marketing OS for expansion/upsell
FAQ
Q: What's the difference between Success OS and PMF OS?
PMF OS: Why do users leave before activating? (Product problem) Success OS: Why do users leave after activating? (Retention problem)
Q: Do I need PMF before using Success OS?
Yes, ideally. If users aren't activating, fixing retention won't help. Success OS will detect this and suggest PMF OS.
Q: Can Success OS actually send emails?
Yes, if you connect an email provider (Resend, SendGrid, etc.). Otherwise, it drafts emails for you to send manually.
Q: How does Success OS learn my business?
During /setup, it conducts an interview about your stage, focus, trial length, and activation metric. It also reads your codebase and landing page.