Building a workflow from scratch can be surprisingly intimidating. The logic sounds simple when you say it out loud: “Remind a customer they’ve abandoned items in their cart, but don’t if they buy”. And yet, when turning that sentence into connected blocks, it’s a familiar story to wonder — where to start? And what’s the difference between a Delivered, Clicked, and Opened block again?

And even when you finally piece the blocks together, you’re not done. You go through every branch by hand, send yourself a test message, and… something is still off. The recommendation block shows empty spaces, or you somehow get three messages in a row. But that part of the job just got much easier: Yespo has built an AI Assistant right into the workflow editor. You describe your goal in plain language, and it maps the sequence together for you.

Not long ago, we brought AI into the message editor to help you create emails faster, and now we’re going further. Let’s dig into how we streamline your work.

What Trigger Workflows Are And Why They Matter

Before we get into the assistant, let’s quickly align on what we’re actually building.

A trigger workflow responds to customer actions with relevant messages instantly, or after a period the marketer sets, while a promo campaign talks to everyone at once — the same message whether a person bought yesterday or hasn’t opened email since signing up for the newsletter. That’s why 43% of consumers unsubscribe when brands send too many emails. 

This matters because most people don’t leave for the reason marketers assume. They can be distracted at the moment, and nobody followed up. And with CAC rising day by day, it becomes too expensive to lose customers you’ve paid for. 

Set up properly once, triggers rarely need further enhancements and can account for up to 20% of total revenue. Our client Liki24.com saw a +295% CTR and a +33.2% CR compared to promo emails. 

The workflow itself is a map that leads the customer. A set of blocks decides: 

• What happens
• With whom
• In what order
• Under which conditions

It branches along the way to check things — did the customer click the push notification? If not, they get an email; if they did, they don't. Each branch sends them down a different path. One contact might get a single reminder and convert; another might move through email, then Web Push, then Viber before they're back.

But all that branching logic is also what makes workflows fiddly to assemble manually, especially for brands who strive for seamless omnichannel marketing. This is where the assistant comes in handy. 

Building And Editing Workflows with AI 

Workflows pile up over the years until nobody dares touch them — the person who built the logic is long gone, and every change feels like a risk. We spent years making the editor simpler: added a search that understands synonyms and made block settings openable in preview mode. And now you don’t need to learn the editor at all — just toggle the AI Chat button in the top-right.

From there, the editor works like a conversation — you describe the logic, and the assistant figures out which blocks and conditions it requires, building them automatically instead of making you drag and configure each one by hand. If something's unclear, it asks follow-up questions, and it replies in whatever language you write in.

You don't even have to be building anything to use the chat. If you're unsure how a block works or which one you need, just ask and get an answer right there in the editor. It's the quickest way to get unstuck without leaving your canvas to dig through the docs and is especially helpful if you’re new to the workflow editor.

Creating a workflow

You can create a workflow in two ways: from your own description or from one of 20 ready-made ecommerce templates covering the usual suspects — Abandoned Cart, Abandoned Browse, Back-in-Stock, Reactivation, Wishlist reminders, and more. If you're not sure where to start, we’ve rounded up the 10 most high-performing triggered workflows for ecommerce.

Starting from a description? The more detail you give, the more accurate the result. A short prompt works as a starting point, but the assistant will likely follow up with questions about which messages to use, when to send them, and when to stop — and you can ask it for help if you need to confirm a nuance.

Editing

The same chat handles edits. Describe the change in plain language, and the assistant applies it in the editor — no digging through block settings. It can add or remove blocks, rewire connections, adjust timers, change conditions, and tidy up start/stop rules. 

Review Your Workflow Before Launch

Building the workflow is only half the job — the other half is making sure it actually does what it was supposed to do. Usually, that means tracing every branch by hand and sending test messages to yourself until you catch what's off, and now the assistant can do that review for you. Point it at a finished workflow, and it reads the structure back to you — what it does, how contacts move between branches, and which messages go out — then flags what needs attention, like an Email block with no recipient set.

Not everything it flags is an error, either. It also points out where the structure could be tidier — two separate End blocks that could share one ending — and explains the tradeoff instead of quietly rearranging things on you.

The report isn't the end of the conversation. If a flagged issue isn't clear, you can ask about it directly in the chat — for example, why a particular branch was marked as a logic gap — and the assistant explains its reasoning. Once you approve the fixes, they're automatically applied.

The point isn't only catching mistakes but launching a workflow knowing what it does, instead of hoping.

Things to Keep in Mind

• The assistant is still in beta, so expect it to keep improving — and don't be surprised if it occasionally needs a second try.
• Be extra careful when editing an active workflow. Changes can affect contacts who are already moving through it, and the outcome depends on where each contact is and what exactly you changed — a block, a connection, a condition, or a timer. Review the updated logic before you apply it.
• If the result isn't quite what you wanted, refine it in the same chat rather than starting over. The assistant keeps the context of your conversation, so it's usually faster to clarify — name the block, branch, timing, or condition that needs to change — than to rebuild from scratch.
• Test the workflow before launch. The assistant does a lot of the heavy lifting, but the final check is still yours.

Final Thoughts

Putting a workflow together used to mean starting at a blank canvas, assembling blocks one by one, and hoping the logic held when you finally traced it. Now, it starts with a sentence as you describe what you want, and the assistant builds it, reviews it, and helps you fix what's off — so a workflow that once took an afternoon comes together in minutes.

The blocks and branches haven't gone anywhere; you've just got a faster way through them. So the next time an idea for a campaign pops into your head, you can go straight from "here's what I want" to a workflow that's ready to test.

Ready to build? Turn on the AI Chat and describe your first workflow — or, if you'd like to see it in action first, book a demo and our team will walk you through it.

Get professional expertise