Creating a complete set of help articles for a SaaS feature launch — from cross-functional planning and content architecture through writing, review, and AI-ready structuring.
A SaaS platform is launching Automations — a major new feature that lets users build custom workflows with triggers, conditions, and actions. There's no existing documentation. The content team needs to go from zero to a complete, published article suite before launch day, coordinating across Product, Engineering, and Support teams under a tight timeline. The help center also runs an AI agent, so every article needs to be structured for accurate AI retrieval from day one.
Before writing a single word, I map out every article needed for the launch — organized by topic, assigned a priority, and tracked through production. This gives all stakeholders a clear view of progress and prevents gaps from slipping through.
| Article | Category | Priority | Status | Audience | Reviewer |
|---|---|---|---|---|---|
| What Are Automations? | Getting Started | 🔴 P0 | Published | All Users | Product |
| Creating Your First Automation | Getting Started | 🔴 P0 | Published | All Users | Product |
| Automations Pricing & Plan Limits | Getting Started | 🔴 P0 | Review Ready | All Users | Product + Legal |
| Available Triggers and Actions | Configuration | 🔴 P0 | Published | All Users | Engineering |
| Using Conditions and Filters | Configuration | 🟡 P1 | Published | All Users | Product |
| Connecting Third-Party Apps | Integrations | 🟡 P1 | Content Drafted | All Users | Engineering |
| Testing and Debugging Automations | Configuration | 🟡 P1 | Content Drafted | All Users | Engineering |
| Managing Team Permissions for Automations | Admin | 🟡 P1 | In Progress | Admins | Product |
| Automation Run History & Logs | Configuration | 🟢 P2 | Content Drafted | All Users | Engineering |
| Automations API Reference | Developer Docs | 🟢 P2 | In Progress | Developers | Engineering |
| Common Automation Errors | Troubleshooting | 🟡 P1 | Review Ready | All Users | Support + Eng |
| Automation Rate Limits & Quotas | Troubleshooting | 🟢 P2 | Content Drafted | All Users | Engineering |
Here's an example of how I write help center articles — clear headings, scannable structure, specific details, and a tone that respects the reader's time. This article is structured to be AI-ready: each section answers a distinct question so AI agents can surface precise answers.
Automations let you connect triggers, conditions, and actions to automate repetitive tasks in your workspace. This guide walks you through building your first automation from scratch.
To create automations, you need:
Navigate to Settings → Automations from your workspace sidebar. Click New Automation to open the builder.
A trigger is the event that starts your automation. Select one from the trigger menu:
Conditions filter which records the automation applies to. For example, you can set the automation to run only when the "Priority" field equals "High." Click Add Condition to set up filters.
An action is what happens when the automation runs. Common actions include:
Give your automation a descriptive name (e.g., "Notify team when high-priority task is created"). Toggle the automation to Active. The automation will begin running the next time the trigger fires.
Once active, your automation runs in the background. You can view run history, check for errors, and adjust settings at any time from the Automations dashboard.
Modern help centers are increasingly powered by AI agents (like Intercom's Fin or Zendesk AI). These tools work best when articles follow specific structural patterns. Here's how I structure content so AI can surface accurate, relevant answers — not hallucinate them.
| Principle | What It Means | Example |
|---|---|---|
| One topic per article | Each article answers one clear question or covers one workflow | "Creating Your First Automation" is separate from "Automation Pricing" |
| Descriptive headings | H2/H3 headings state the answer topic, not just labels | "Step 2: Choose a Trigger" not "Triggers" |
| Front-load key facts | Put the most important information in the first sentence of each section | "A trigger is the event that starts your automation" before listing trigger types |
| Structured data patterns | Use lists, tables, and consistent formatting that AI can parse | Trigger types in a list with bold labels, not buried in a paragraph |
| No ambiguous pronouns | Repeat the noun instead of using "it," "this," "that" | "The automation runs when…" not "It runs when…" |
| Audience-appropriate language | Match vocabulary to the reader's expertise level | End-user articles say "connect your app"; developer docs say "configure the OAuth flow" |
Product launch documentation doesn't happen in isolation. Here's the workflow I use to coordinate with Product, Engineering, and Support teams — especially under tight timelines.
| Phase | Activities | Teams Involved | Output |
|---|---|---|---|
| 1. Discovery | Review PRDs and specs, attend feature demos, interview PMs and engineers on edge cases | Content + Product + Eng | Article list, content plan, audience mapping |
| 2. Drafting | Write first drafts, capture screenshots from staging, create step-by-step flows | Content (lead) | Draft articles in CMS |
| 3. Review | Route drafts to PMs for accuracy, Engineering for technical correctness, Support for common questions | Content + Product + Eng + Support | Reviewed and approved articles |
| 4. Polish | Final edit pass, add media, set audience visibility, optimize structure for AI agents | Content (lead) | Publish-ready articles |
| 5. Launch | Publish articles, add to help center collections, enable for AI agent, notify Support team | Content + Support | Live documentation suite |
| 6. Post-Launch | Monitor search queries, track ticket deflection, iterate on gaps from real user behavior | Content + Support + Data | Updated articles, gap analysis |
Launching documentation alongside a product feature requires speed without sacrificing quality. Here's what drives the process:
I can build your documentation suite from scratch — on time, on strategy, and AI-ready from day one.
Start a Conversation