Product Discovery: Scenario Outputs and Synchronous Subscenarios? Let's talk!

Features
1

Hey all! Daniel Born, PM @ Make here :wave:

I’m now doing product discovery :mag_right: on how to better chain scenarios.
Are you chaining your scenarios? Do you use Subscenarios? Are you still relying on HTTP+webhook+response combo? Let’s talk!
Reach out or schedule us a meeting here (no longer accepting call bookings)

5 Likes
8

I’m using them extensively, removing http/webhook scenario chaining with subscenarios.

11

I’m using subscenarios because I have variables in them that I want to control through several different runs and I want to reuse them across different scenarios.

I’m finding the flat list of scenarios and subscenarios a bit difficult to navigate - it would be nice if there was some kind of a folder structure to it that could just show the parent scenarios with an indented display of child scenarios.

Are there other improvements you’re thinking about?

13

Hi, I would like to replace http/webhook scenario chaining with subscenarios but it seems to be impossible for a sub scenario to return any data.
How did you manage to do that ?

14

Hi Mikael, and welcome to the community!

Currently the only way for scenarios to send data back to another scenarios is using a webhook trigger and a webhook response module.

We’ll however be introducing a capability that is native to the platform and works on top of the existing Subscenarios capability

3 Likes
15

Hello there, this is Vojtech, a product designer here at Make. I’m excited to announce that we’ve just released scenario outputs and synchronous subscenarios.

What’s new?

  1. You can define scenario outputs in the ‘Scenario inputs and outputs’ panel
  2. New ‘Return output’ module that allows you to return data from a scenario
  3. New ‘Call a subscenario’ module with ‘Wait for subscenario output’ settings
  4. New ‘Start a subscenario’ module for better scenario inputs mapping

Why is it important?

  1. Synchronous subscenarios: you can break down large and complex scenarios into small and reusable pieces that return data back to the parent scenario
  2. MCP tools: you can build scenarios that function as MCP tools along with the Make MCP server, allowing you to call Make scenarios from clients like Claude
  3. AI Agent tools: you can build tools for the Make AI Agent that return structured data to help the agent understand context

If you have any feedback or questions, feel free to book some time with me though this scheduling link.

2 Likes
16

Hey guys, for some reason I don’t see the “Scenario output” tab.

And the button (on hover) also says “Scenario inputs” instead of “Scenario inputs and outputs” … am I missing something?

image
image711×462 36.7 KB

17

Hello Mudit, scenario outputs should now be available in your Make account. Could you please double-check it and let me know if you can access it?

18

Hey everyone! I love this new feature and currently migrate all my workflows to use this structure as it’s very powerful for conditionals and routes. I’m only having one important concern: I can’t store failed executions and because of that I can’t use error handling, I really hope you find a way to fix this as it’s a very big downside…

19

I just stumbled across this thread and want to share my 2cents:

i) I love that you can add collections inside the scenario inputs. This enables stuff like

{
"company":  {
  "name":"..",
  "email":".."
  },
"projectManager":  {
  "name":"..",
  "email":"..."
  }
}

And you don’t need to use those looong variable names such as projectManagerName & companyName. However, when you use the collection variables in the scenario you can only choose ‘company’ and then you have to use

get(company,'name')

which is is okay for a few attributes but definitely more complicated then it needs to be!

ii) Scenario input queues
There have been a number of situations where I have a few hundred e.g. PDFs that need to be analyzed but doing so in parallel would be too much in terms of API rate limits. Now with scenario outputs you can actually wait until one is over to start the next but that becomes complicated if the total time is longer than 40min (maximum scenario execution). This is where I revert back to using webhooks so I can simply dump them into the queue which is being executed one after the other…

iii) +1 for failed executions. In webhooks they stayed in the queue to retry but in subscenarios they are gone.