How do you build your first AirOps workflow?

Quick Answer: AirOps is an AI content platform that lets you build automated workflows for SEO and content production without writing code. This guide walks you through your first workflow from a blank canvas to a working output, covering steps, agents, models, and how to map your results.

Getting started with AirOps feels overwhelming at first. The interface is flexible, the terminology is new, and the documentation assumes you already know what a workflow is supposed to do.

This guide does not. It starts from zero and walks you through building a real, working AirOps workflow, step by step. By the end, you will know how inputs and outputs work, how to choose a model, how to add an agent step, and how to map your results so they are actually usable.

What Is AirOps, and Why Does It Matter for Content Teams?

AirOps is an AI workflow platform built specifically for content and SEO teams. Instead of prompting ChatGPT one query at a time, AirOps lets you build repeatable, structured workflows that run AI tasks at scale, consistently, with your brand voice and content logic baked in.

A workflow in AirOps is a sequence of steps. Each step takes an input, runs a task (usually an AI model or a data lookup), and passes an output to the next step. You can build workflows for keyword research, content briefs, article drafts, internal linking, and more.

Why this matters: manual AI prompting does not scale. AirOps turns a one-off prompt into a repeatable process your whole team can run.

Before You Build: Three Concepts You Need to Understand

1. Inputs

An input is the variable information you feed into a workflow each time you run it. Think of it like a form field. You might have an input called keyword or target_url. Every time you run the workflow, you fill in that field, and the workflow uses it throughout.

2. Steps

Steps are the building blocks of a workflow. Each step does one job: call an AI model, run a search, transform text, or pass data forward. Steps run in sequence, and each one can reference the output of any previous step.

3. Outputs

An output is what a step produces. You name it, and then you can reference it in later steps using a variable like {{step_name.output}}. Good output mapping is what makes a workflow actually useful rather than a black box.

How to Build Your First AirOps Workflow: Step by Step

This walkthrough builds a simple Keyword Content Brief Generator. You give it a keyword, and it returns a structured content brief. It is simple enough to finish in 30 minutes and practical enough to use immediately.

Step 1: Create a New Workflow

1. Log in to AirOps and go to Workflows in the left sidebar
2. Click New Workflow
3. Give it a name: "Content Brief Generator"
4. Click Create

You now have a blank canvas with a start node on the left.

Step 2: Add Your Input Variable

1. Click the Start node
2. Select Add Input
3. Name the input keyword (lowercase, no spaces)
4. Set the type to Text
5. Add a description: "The target keyword for this content brief"

This input is now available to every step in the workflow as {{keyword}}.

Step 3: Add Your First LLM Step

1. Click the + button to add a new step after Start
2. Choose LLM from the step type menu
3. Name the step brief_generator

Now configure it:

Model selection: AirOps gives you access to multiple models including GPT-4o, Claude 3.5 Sonnet, and Gemini. For content briefs, Claude 3.5 Sonnet performs well on structured, long-form outputs. Select it from the model dropdown.

Write your prompt:

You are an expert SEO content strategist.

Create a detailed content brief for the following keyword: {{keyword}}

Include:
- Recommended H1
- Search intent summary (2-3 sentences)
- 5 suggested H2 headings
- 3 competitor angles to address
- Recommended word count
- Key entities and terms to include

Format the output as structured text with clear labels for each section.

4. Click Save Step

Step 4: Name and Map Your Output

After saving, click on the step and find the Output field. Name it content_brief.

This means later steps (or your final export) can reference {{brief_generator.content_brief}} to pull this result through.

Output mapping tip: always name outputs by what they contain, not what step produced them. content_brief is more useful than step_2_output when you are building more complex workflows later.

Step 5: Add an Agent Step (Optional but Powerful)

An agent step is different from a standard LLM step. Instead of following a fixed prompt, an agent can make decisions, use tools, and run sub-tasks to reach a goal.

To add one:

1. Click + after your LLM step
2. Choose Agent from the step type menu
3. Name it serp_researcher

Configure the agent:

• Goal: "Research the top 5 ranking pages for {{keyword}} and summarise the key topics they cover"
• Tools: Enable Web Search from the tools panel
• Output name: serp_summary

The agent will browse live search results and return a structured summary. You can then reference {{serp_researcher.serp_summary}} in your brief prompt to ground it in real SERP data.

When to use an agent vs. an LLM step: use an LLM step when you know exactly what you want and can write a precise prompt. Use an agent when the task requires gathering information, making decisions, or handling variable inputs.

Step 6: Connect Your Steps and Test

1. Make sure your steps are connected in sequence: Start → brief_generator → serp_researcher (or whatever order fits your logic)
2. Click Run Test in the top right
3. Enter a test keyword, for example: "AI workflow automation for content teams"
4. Click Run

Watch each step execute in real time. Green means it passed. If a step fails, click it to see the error log.

Step 7: Review and Export Your Output

Once the workflow runs successfully:

1. Click on the final step to see its output
2. Check that the content brief is structured correctly
3. Use the Output Panel on the right to see all named outputs from every step

You can export results to a Google Sheet, copy them directly, or connect AirOps to your CMS via the integrations panel.

Choosing the Right Model for Your Workflow

Model choice affects quality, speed, and cost. Here is a practical guide for beginners:

You can switch models per step, so a single workflow can use different models at different stages depending on what each step needs.

Common Mistakes Beginners Make in AirOps

Not naming outputs clearly. If you leave outputs as default names, referencing them later becomes confusing. Name every output before you move to the next step.

Writing one giant prompt. A single step trying to do five things produces inconsistent results. Break complex tasks into multiple steps, each with one clear job.

Skipping the test run. Always test with a real input before sharing a workflow with your team. Edge cases break prompts in ways that are not obvious until you run them.

Using an agent when an LLM step is enough. Agents are slower and use more credits. If you know exactly what output you need, a well-written LLM prompt is faster and cheaper.

What Can You Build After Your First Workflow?

Once you are comfortable with the basics, AirOps workflows can handle significantly more complex content operations:

• Content brief pipelines that pull SERP data, analyse competitors, and output a ready-to-write brief
• Programmatic SEO workflows that generate hundreds of location or category pages from a data input
• Internal linking assistants that scan your existing content and suggest links for new articles, often alongside a broader B2B SaaS content marketing workflow
• Brand voice checkers that review drafts against a style guide before publication
• AI search optimisation workflows that reformat existing content for citation by tools like ChatGPT and Perplexity, which is increasingly relevant for teams exploring GEO and AEO support

At SaaS Hackers, we use AirOps workflows as part of a broader AI content system. The platform is most valuable when workflows connect to each other, with the output of one feeding the input of the next. For teams building that wider engine, it often sits alongside SEO specialists, SEO agencies, or a more complete digital strategy function.

FAQs

What is AirOps and what is it used for? AirOps is an AI workflow platform built for content and SEO teams. It lets you build multi-step automated workflows that run AI tasks at scale, including content briefs, article drafts, keyword research, and internal linking. It connects to multiple AI models and can pull live data from the web via agent steps.

Is AirOps suitable for beginners with no coding experience? Yes. AirOps uses a visual, no-code workflow builder. You add steps, write prompts, and connect outputs without writing any code. The learning curve comes from understanding workflow logic and prompt writing, not from technical setup.

How long does it take to build a first workflow in AirOps? A simple workflow with one or two LLM steps takes around 20 to 30 minutes to build and test for the first time. More complex workflows with agent steps, branching logic, or external integrations take longer, typically a few hours.

What is the difference between an LLM step and an agent step in AirOps? An LLM step runs a fixed prompt against a model and returns an output. An agent step can use tools (like web search), make decisions, and complete multi-step sub-tasks to reach a goal. Use LLM steps for predictable, structured tasks. Use agent steps when the workflow needs to gather or evaluate information dynamically.

How much does AirOps cost for beginners? AirOps pricing starts at $200 per month for team plans. There is no permanent free tier, though AirOps has offered trial access. For teams running content at scale, the cost is typically justified by the reduction in manual production time across writers and strategists.