Most startups think of a help center as an afterthought - a library that exists so support agents can copy-paste links.
We approached it differently.
Our docs became the backbone of our PLG motion, powering onboarding, automation, and real-time guidance.
Here’s how we built it in a single day (for <$25/month), and how you can too.
The Case for a Knowledge Base (Even if Nobody Reads It)
We’ll admit it: very few users actually read documentation. We estimate doc engagement at ~5%. That’s why Autoplay focuses on real-time, in-context guidance.
So why did we still build one?
To fuel automation. Docs power Autoplay Chat in Zendesk.
To create reusable assets. Tutorials work across docs, onboarding, and tooltips.
To set the foundation for proactive support. The same content will soon drive in-app nudges.
Think of it less as documentation, more as a content backbone for growth.
The Workflow We Wanted to Enable
User signs up → cohort defined in PostHog
User hesitates → chatbot serves the right doc from Zendesk
Future state → contextual popups triggered in-app, powered by the same content
Our knowledge base wasn’t just for browsing; it was the infrastructure for guided onboarding.
The Tools That Made It Possible
Arcade: Interactive Product Tutorials
Why it worked:
Record step-by-step product walkthroughs in 30-40 minutes
Look and feel just like the real product
Reusable across docs, emails, onboarding flows, and tooltips
Zero developer bottlenecks
Impact: Within one week, we had 12 tutorials live, covering ~80% of new-user support volume.
Notion Desk: Simple, Familiar, Cheap
Why it worked:
Native to a tool we already use daily
Easy setup, minimal context switching
<$25/month
Caveat: Adoption wasn’t instant - we only got there after a live demo with the founder. Ironically, a reminder that even PLG products often need human onboarding.
Lessons Learned (Case Study Mode)
Speed matters more than polish. 12 scrappy tutorials eliminated the bulk of repetitive support work.
Docs should be a growth engine, not a graveyard. Ours are designed to be surfaced automatically, not hunted down manually.
Your “help center” is also a dataset. Every doc you write feeds into automation, contextual triggers, and future AI guidance.
How to Replicate This (Our Playbook)
Pick your content backbone. Choose a simple help desk (we used Notion Desk). Don’t over-engineer it.
Create interactive tutorials. Use a tool like Arcade to produce tutorials quickly and make them reusable.
Map to user journeys. Define cohorts and hesitations in PostHog (or your analytics tool).
Automate delivery. Pipe content into your support bot, onboarding flows, and in-app nudges.
Iterate weekly. Add 2–3 tutorials at a time, guided by the top support questions.
Product Genie: The AI That Turns Docs Into Guidance
Behind Autoplay sits Genie, our AI engine that actually understands a product’s UI. Genie takes raw material - like the tutorials and docs we just built - and uses them to power:
Session Replays: Genie overlays intent, hesitation, and friction signals on top of raw replay data. Instead of just watching clicks, you see why users struggled.
Chat: Genie plugs into support bots, serving the right doc or tutorial automatically based on what a user is doing in the product, not just what they typed.
Agents: Genie is the brain that will eventually drive fully autonomous product guides - AI agents that can navigate the UI alongside users in real time.
The help center is just the starting layer. Genie is what transforms it from static content into a living guidance system across replays, chat, and in-app experiences.
The Takeaway
A help center doesn’t have to be a static library.
With the right tools and mindset, it can be your first growth engine: powering automation, reducing support load, and making onboarding feel proactive instead of reactive.