Join list items (smushing) and Filtering
You’re getting the hang of Gumloop now—reading from Google Sheets, looping on lists, triggering flows. But there’s one more superpower to unlock when working with lists: knowing how to use them depending on your goal.
When you’ve got a list (whether it’s contacts, attendees, tickets, or anything else), there are usually two things you might want to do:
- Handle each item individually – Like sending an email to each attendee. This is where loop mode comes in (we covered that earlier).
- Do something with all the items together – Like writing a report from all tickets received this week, or—like we’re doing—having AI recommend talks from the full list of available sessions.
When you’re in that second camp, you’ll need to convert your list into one big chunk of text. That’s where our favorite helper comes in:
Let’s bring it back to our conference workflow.
Originally, we hardcoded the full list of talks into the AI prompt that recommends sessions for attendees. But now that those talks live in a Google Sheet, we want to pull them in dynamically—and make sure any new talks are included automatically.
Here’s the situation:
- We’ve got a list of talks from Google Sheets ✅
- We want to send all those talks into the Ask AI node as part of the prompt ✅
- But we can’t just pass in a list to a text field (that would trigger loop mode) ❌
So, we need to smush.
Once smushed, we pass that text into the prompt field of the Ask AI node. And now, the model sees the full list of talks and can recommend the most relevant ones based on the attendee’s background.
Enter: Filtering
So far, we’ve assumed we want to use all the data in our lists—but in reality, that’s rarely the case.
What if some talks are still being confirmed? Or some are cancelled? We only want to recommend talks that are marked Active.
This is where the Filter node comes in.
The filter node lets you specify:
- The list to filter by (e.g. the “Status” column from Google Sheets),
- The condition (e.g. contains “Active”),
- And which list(s) to output if the condition is met (e.g. the talk titles).
It works like this:
Once the filter runs, you’re left with a clean list of only the talks that are live and ready to recommend.
You can then smush that filtered list using Join List Items—and send only the most up-to-date, relevant talks into your Ask AI prompt.
So now, the magic comes together:
- We’re only including active talks (thanks to filtering).
- We’re passing them all as one block of text (thanks to smushing).
- And AI is making recommendations from the most current list—no manual updates needed.
That’s automation done right.
Now that we’re dynamically pulling, filtering, and smushing our data, our flow is looking pretty smart—and super scalable. But as your workflows grow more complex, you’ll inevitably hit bumps in the road: missing data, unexpected inputs, or logic that needs a fork in the road.
That’s what we’re tackling next.
In the next lesson, we’ll explore how to make your flows more resilient and adaptive using Error Shields and If-Else logic—so even when things go sideways, your automation keeps moving forward. Let’s dive in.
