Get the Full Stack
01. System 02. Stack 03. Source 04. Answers Get the Full Stack
User-onboarding protocols

The onboarding system your product deserves.

Built from designing products used by millions.

Get the onboarding protocol stack. Activation. Retention. Expansion.

Get the Full Stack

Created from two decades of work across teams at

Google Microsoft Intel Couchbase Heatmap Blue Ocean
00. The problem

Your most expensive problem is the weeks after signup.

Not pricing. Not features. It’s the gap between “signed up” and “I need this,” and that’s where the money quietly leaves.

You’ve bolted tools on before. They don’t hold.

Tooltips, tours, checklists. They sit on top of the product and point at it. They narrate the thing instead of changing it, so the user is still just as lost, now with captions. That treats the symptom, a confused user, and leaves the cause alone: an onboarding nobody actually designed.

And that layer is running out of road. As products get more AI-led, there’s less fixed screen left to stick a tooltip onto. The fix has to live inside the product, not float above it.

01. The system

User-onboarding is not a funnel, it’s a game.

Game levels are built on loops, consisting of cue, action, reward until the user unlocks new tools and upgrades.

Games teach you while you play, a guide always at your side. That’s agent-led onboarding.

Good onboarding works like a game, not a manual.

Most people picture onboarding as the sign-up wizard. It’s not. It’s everything from the first win, to the habit that keeps them coming back, to the day they bring their team with them.

Great games turn strangers into experts without a tutorial anyone reads. They use a loop: a cue, one small action, a real reward, then the next cue.

The same loop carries a user past the first win, into the habit that brings them back, and on to the upgrade that feels like the next natural move. Curious user to power user, with no nudge that feels like nagging.

Agent-led means the guide does the work, not you.

It’s not a separate layer bolted on top with tooltips. It’s a guide that sets the product up for you, so you never have to go learn the dashboard, hunt for where things live, or figure out the UX on your own.

It does the work and hands you the win. One place to go for everything, instead of a map you have to memorize. The product onboards you to the product.

Cue Action Reward Habit loop

This is what agent-led looks like.

The product onboards you to the product.

The agent does the work.

Step 1:
Completed Triggered Pending

Ask setup intent

Question

In-app question: 'What kind of work is this?'

Preparing question…
Client work
Internal team
Just exploring
Answered
Step 2:
Completed Triggered Pending

Build the workspace

Action

Runs when the answer comes in.

Processing answer…

Building project0% complete
Built
Step 3:
Completed Triggered Pending

Ask to add team

Question

In-app question: 'Want to add a teammate?'

Awaiting response…

The user gets the win.

JaimieAI Assistant

Hey Olivia, let's set up your first project.

Olivia, you're crushing it

You've got your first project set up and you're ready to add teammates.
JaimieAI Assistant
Thinking...
What kind of work is this? I'll build the workspace to match.
AClient work
BInternal team
CJust exploring
JaimieAI Assistant
Thinking...
Building your workspace0%
JaimieAI Assistant
Your client workspace is ready. Want to add a teammate?
Add teammate
View project
Ask me anything...
02. The stack
The Full Protocol Stack Three protocols. One toolkit. Yours to keep. Feed them to your favorite LLM and build the onboarding your product should’ve shipped with.
The Activation Protocol

The Activation Protocol

Get a new user to a real win on day one.

  • The Level Design model: why activation is Level 1, and how the agent runs it

  • How to define the one first-session win that matters

  • The path, the first quest, and the win moment, built for the agent to walk the user through

  • The Activation Protocol worksheet, plus how to run it with your favorite LLM

The Retention Protocol

The Retention Protocol

Turn the first win into a daily habit.

  • The Level Design model: why retention is Level 2, and how the loop keeps users coming back

  • The one repeated action a habit is built on, and how the agent rewards and calibrates it

  • Why friction is a feature, and when the agent hands over the controls

  • The Retention Protocol worksheet, plus how to run it with your favorite LLM

The Expansion Protocol

The Expansion Protocol

Turn mastery into more seats and advocates.

  • The Level Design model: why expansion is Level 3, and why your best users are the ones most likely to leave

  • How to tell when a user has outgrown their plan, so the agent offers more at the right moment instead of nagging everyone

  • How mastery turns into more seats and advocates, with the agent inviting the right person by name

  • The Expansion Protocol worksheet, plus how to run it with your favorite LLM

Bonus toolkit
Level Design Patterns

Level Design Patterns

17 agent-led patterns behind the protocols.

  • 17 named patterns, grouped by activation, retention, and expansion

  • Each with its mechanism, a real example, what works, and what goes sideways

  • The level-design lens that ties them together

  • A swipe file to pull from, or hand to your favorite LLM when you're building

How it works

  1. 1Download the four PDFs
  2. 2Upload to your favorite LLM
  3. 3Build your onboarding blueprint

Built specifically to collaborate with Claude, ChatGPT, Gemini, or whatever LLM you use.

$196$129Save $67

Lifetime updates included.

Get the Full Stack
03. The source

The system every founder deserves.

Hey, I'm Bal.

I've spent twenty years across marketing and product, working with teams at companies like Google, Microsoft, and Intel. The same thing happened over and over. A team builds something genuinely good, launches it, and watches people sign up, look around, and leave before they ever get to the part that would have hooked them.

The product was rarely the problem.

The onboarding was the problem. So I spent years working out how to get a stranger to a real win and keep them coming back, until it became the one thing I do.

This is the same system I've used on enterprise B2B products that I charge 24K+ for. I packaged it so an early-stage founder can run it in an afternoon, without a team or a big budget. It's the closest I can get to a do-it-yourself version.

"Bal turns complex workflows into onboarding that feels effortless."

Corey Leger

Corey Leger, Director of CRO, Heatmap

From 8.2% to 4.9% churn 250% growth

04. The answers

FAQs

Everything you need to know.

What exactly do I get?
The three protocols, Activation, Retention, and Expansion, plus the Level Design Patterns. Four PDFs covering the whole Onboarding Loop, from a stranger's first session to your power users bringing in their teams. After checkout you land on a page with the downloads and a short walkthrough for using them with your favorite LLM. Yours to keep, no app, no login.
Why PDFs instead of skills, GPTs, or an app?

Anyone selling you prompts, skills, or prebuilt agents is selling to noobs. You don’t need to be an AI expert to use these, but if you’re getting real value out of them, you’re probably already the type who writes your own prompts, builds your own skills, customizes your own agents, maybe ships your own apps.

And every product is different, so there’s no prompt or agent that works out of the box for yours. What you need are the protocols that guide your LLM into building the right onboarding for your product, not someone else’s.

They’re PDFs instead of .md files in case you feel like sitting down and reading them yourself.

Does this only work for desktop SaaS?
No. The Onboarding Loop is about how any product turns first-time users into committed ones, so the model holds whether you're mobile, B2B, PLG, or a marketplace. The examples lean SaaS because that's where I've worked most, but the patterns are the same wherever people adopt a product over time: activation, retention, and expansion. Your LLM handles translating them to your context.
Do I need an LLM to use these?
No, but it's how I'd use them. You can read straight through and do the work yourself. What I actually do, and what I built these for, is hand them to my favorite LLM and let it apply the thinking to my product. A workout plan works as written; onboarding is specific to each product, so instead of generic steps these give your LLM the expert context to design onboarding for yours. There's no magic prompt to paste. The protocols are the brief, you tell it what you're building, it does the rest.
What if I don't have an agent in my product yet?
Then this is where you start. The protocols don't assume you already have one. They show you what agent-led onboarding does, so you can build your first version at whatever level fits your product. It doesn't take a heavy behavior-tracking setup to be useful; even a simple guided agent beats a tour of tooltips. The point is to design the onboarding your agent should run, whether that agent exists yet or not.
How long until I see results?
That's up to you, your product, and how you implement. These give you the model and the moves; how fast it shows up depends on what you build with them and where your onboarding is weakest. Some of it you could ship this week.

Get the complete system.

Three protocols. One toolkit. Activation to expansion, designed to build from.

$196 $129 Save $67
Get the Full Stack

Hi, I’m Bal. These are the protocols I use to build agent-led onboarding that sells your product for you.

Get the Full Stack