Over the past year or so, I’ve used a variety of LLM workflow tools to build out content automation processes for our marketing clients. Let’s look at the three that I like the most.
AirOps, Gumloop, and n8n all have different strengths. Each is suited for different budgets and skill levels.
At ércule we use n8n because we want that technical flexibility, but we have a lot of clients for whom AirOps or Gumloop are better fits.
Overview: AirOps is for content folks, n8n is for product developers
I think of these three products on a spectrum, from least technical to most technical.
- AirOps is the best fit for marketers who don’t want to tinker with the back-end. The subscription-based cost model makes it the most expensive of the three, but you’re paying for a lot of ease and support.
- Gumloop is the best fit for marketers who want to create more custom workflows. It’s got a steeper learning curve than AirOps but also a lower monthly cost. No subscription required.
- n8n is the best fit for product developers and marketers who really want to get in the weeds. It’s also probably the most scalable option (as long as you know how to manage your server costs).
👇 Here’s a video walk-through of all these products if that’s more your speed.
Before we get into the details, here’s a little primer for my less technically inclined marketing friends.
For beginners: how do these LLM marketing flows work?
If you’re new to this whole LLM workflow concept, here’s a basic rundown…
- Create a series of nodes, which are (basically) small AI automations.
- Connect these nodes in a sequence.
In the example below, my first node is collecting a list of topics that I want to write about. It will then run each topic through the subsequent nodes (or tasks) in the workflow.
The workflow is going to…
- Review the Google SERP results for each topic phrase.
- Scrape the websites that appear at the top of the SERP.
- Ask AI to take the stuff that I just scraped and extract a list of 10 human-readable questions from this data.
- Generate answers to those questions.
- Export the questions and answers to a spreadsheet.
You might use workflows like this for any number of content production and optimization tasks…
- Analyzing and updating your existing library
- Conducting organic content research
- Drafting new content to fill any gaps in your library
All of the platforms I’m reviewing here are built with that basic design concept in mind.
AirOps
AirOps is designed for content engineers who want to create complex workflows without building anything from scratch.
Usability and pre-baked features are quite awesome
The UX is led by AI agents, which is super helpful, especially if you’re new to content automation. It provides pre-built nodes so you can drag-and-drop tasks like…
- Deep Research
- Create content from keyword
- Add internal links
- Content scorecard
This pre-bake factor is probably the biggest differentiator from Gumloop and n8n. (You’ll probably want to build your own at some point in AirOps but for getting started the pre-baked agents save a ton of time.)
AirOps overall: extremely convenient (but not so flexible)
AirOps is a little bit less flexible in terms of what it can do. It's a little harder to build things that you want. There's less of a community marketplace. Gumloop and n8n have those things. (Maybe it exists for Gumloop and I'm just not aware of it.)
Pricing is on a monthly subscription, which means you’re paying more every month than you would with n8n or Gumloop. But in exchange, you’re getting a lot of convenience. For example: n8n and Gumloop require you to manage data in spreadsheets but AirOps has a native “grid” feature that saves you the trouble.
Gumloop
Gumloop is more flexible than AirOps but it also requires more hands-on building and management.
Content folks can definitely use Gumloop! But it's designed for people who have a little bit of sophistication in terms of what they know how to develop.
The UX requires more hands-on building
Gumloop uses the drag-and-drop format like AirOps, but there are no pre-baked agents.
The nodes are designated for types of tasks…
- Ask AI
- Extract data
- Scrape a website
- Generate text
Each of those nodes needs to be engineered a bit. For example: you’ll have to enter specific prompts for each AI node. This can be a strength if you’d like to customize every detail. It can also be a drawback, if you’re trying to build basic flows as fast as possible.
Gumloop overall: dynamic (but more labor intensive)
The workflows you create with Gumloop can be much more dynamic than the pre-baked AirOps options. But it requires knowing what you want and learning how to build it.
Building a workflow also means maintaining and fixing it on the fly. For my workflows in Gumloop (and n8n) I store data in spreadsheets. That’s the kind of outside maintenance task that AirOps does not require.
In terms of pricing, there’s no monthly subscription. Gumloop works on a credit system. You have to run everything through their infrastructure. This can get expensive pretty fast if you’re not careful.
n8n
n8n is the most technical and flexible of the bunch but it also has the steepest learning curve. It’s designed more for software developers than content engineers.
Our dev team at ércule loves n8n because we spend a lot of time building pretty robust workflows and we need something very efficient to let us do that.
This interface is not for the faint of heart
The UX can be a little rough compared to the other platforms but the testing tools in n8n are way better.
On the flipside: troubleshooting and debugging functions are much better in n8n than Gumloop. (And if you’re not building anything in AirOps then you don’t need to debug much in AirOps.)
n8n overall: the building potential is huge (but so is the learning curve)
If you do enjoy getting into the nitty gritty side of development then n8n is pretty awesome.
And, compared to Gumloop, n8n makes it much easier to share and archive your workflows. An n8n workflow can be output in a format JSON, for example, so you can store it on Github.
As for pricing, n8n tends to be less expensive to run because you can opt to run your workflows locally, in which case you’re not paying extra to have your workflows run in the cloud. But optimizing your costs will require knowing how to optimize your server usage.
Automation makes content systems even more crucial
Which of these is the best fit for you? It’s really a matter of your priorities.
- AirOps is geared toward people who want to learn content engineering and want a lot less hassle around building and maintaining the workflow side.
- Gumloop is for marketers who want more custom workflows and aren’t afraid to do a bit of building and maintenance.
- n8n is for more development-oriented teams that are looking to build the content systems of their dreams from the ground up.
No matter which of these platforms you choose, you’ll wind up with a level of production power that we’ve never seen before in marketing.
That power comes from a new mode of content marketing. While we used to treat our practice like studio artists we now approach it as engineers. The shift is a bit jarring at first, I know. But on the other side of that transition is a much more creative and scalable way to help customers out.
🙂
