Building Your Own Playbook
Most marketers can prompt an AI, but few can build a repeatable system that produces consistent, on-brand content without starting from scratch every time. In this lesson, Aaron walks you through the full process of designing, building, and scaling your first AirOps Playbook, from defining your challenge statement to running batch content across a Grid.
What You'll Learn
Start With the Challenge
Every Playbook in AirOps begins with a clearly defined challenge. Aaron frames this as a fill-in-the-blank statement: "I need to [outcome] by [method]." This keeps your Playbook focused from the start.
- Before you build anything, articulate what you're solving. The content refresh scout Playbook, for example, started with: "Which pages are losing visibility, and which ones matter most?"
- A good challenge statement connects a specific outcome to a specific method. Aaron's example: "I need to transform a blog post into a LinkedIn post by extracting the most relevant insights for my target audience."
- This framing applies to any use case, whether you're refreshing existing content, repurposing across channels, or building a research workflow.
Four Design Questions for Any Playbook
Once you have your challenge, Aaron walks through four questions that shape your Playbook before you touch the builder.
- What type of Playbook is this? A Scout finds and surfaces information (like scanning for underperforming pages). An Action Playbook transforms an input into an output (like turning a blog post into a LinkedIn post). Some Playbooks do both.
- What does a good outcome look like? Define success in terms of content quality and structure. For the blog-to-LinkedIn example, a good outcome means a strong hook, a clear CTA, concise copy, and a match to your brand's voice.
- What inputs do you need? Identify every input and how it contributes to the output. The LinkedIn Playbook needs a blog URL (the content foundation) and a Brand Kit (for tone, voice, positioning, and audience details).
- What steps get you from input to output? Map the sections your Playbook will run through. Aaron's example uses three: analyze the blog, create a post outline, and draft the post.
Building Your Playbook in AirOps
Aaron demonstrates the build process inside the AirOps platform, showing how inputs, sections, and outputs connect.
You add inputs through the control panel. In the LinkedIn example, Aaron adds a blog URL (as a text input) and a Brand Kit. Both are marked as required.
Each Playbook section has three parts:
- Objective: the north star that tells the agent what this section is for.
- Instructions: step-by-step guidance that references your inputs, tools, and artifacts. You use slash commands to pull in specific inputs or tools.
- Output: a detailed description of what the section should produce. Outputs can be artifacts like articles, reports, charts, or even platform changes.
You can also feed your challenge statement and design answers into Quill to auto-generate a first draft of your Playbook, then refine from there.
Three Fixes That Improve Playbook Quality
Aaron shares three common adjustments that make a significant difference in output consistency.
- Be explicit with output formats. Define a schema using markdown so your outputs are predictable and consistent. When you leave formatting to agent interpretation, results vary.
- Be explicit with references. Call out specific Brand Kit components (voice, tone, audience, writing rules) and explain how agents should use them. This applies to any large context source, including Knowledge Bases and research documents.
- Be explicit with decision criteria. When your Playbook asks agents to prioritize, filter, or select, define the criteria. Aaron suggests specifics like novelty, relevance, and actionability instead of vague instructions like "pick the best ones."
Connecting Your Playbook to a Grid for Scale
The final section covers how to move from a single Playbook run to batch execution using Grids.
- Create a blank Grid and add columns for each input your Playbook requires (blog URL, Brand Kit, etc.).
- Add a Playbook column and map its inputs to the corresponding Grid columns.
- Paste in your blog URLs to build a full content calendar that runs your Playbook across every row.
Aaron suggests a next-level project: create a Scout Playbook that monitors your blog for new publications and automatically sends URLs to the Grid, closing the loop between content creation and distribution.
Key takeaways
- Your Challenge Statement Is Your SpecFrame every Playbook with "I need to \[outcome\] by \[method\]." This one sentence prevents scope creep and keeps your agent focused on a single, measurable goal.
- Vague Instructions Create Inconsistent OutputsAaron shows that telling an agent to "pick the best idea" produces different results every run. Defining criteria like novelty, audience relevance, and actionability turns subjective judgment into repeatable decisions.
- Markdown Schemas Lock In Your FormatWithout a defined output schema, an agent might use paragraphs one run and tables the next. Specifying a markdown structure for each section ensures every output matches the format you expect.
- Name the Brand Kit Fields You Actually NeedReferencing "the Brand Kit" as a whole gives the agent too much room to interpret. Calling out specific components like voice, tone, audience, and writing rules tells the agent exactly what to apply and how.
- One Grid Column Turns a Playbook Into a Content CalendarAdding a Playbook column to a Grid lets you paste in dozens of URLs and run them all at once, converting a single-use workflow into a scalable content operation.
FAQs
A Scout Playbook finds and surfaces information, like scanning a blog for underperforming pages or identifying content gaps. An Action Playbook transforms an input into an output, such as converting a blog post into a LinkedIn post. Some workflows combine both: a Scout identifies opportunities, then an Action Playbook executes on them.