Executive Series
by Jesse Greenberg
Posted on September 9, 2026
Recently I met with the VP of Customer Care at a retail company. They wanted to automate their return order call drivers but IT told them it was far down their list of priorities.
The answer they got wasn’t “we can’t automate that.” It was “we can’t automate that yet.” The AI was capable, but they were waiting on a connection into the system that hadn’t yet been built.
This Isn’t a One-Off. It’s a Pattern.
Every client I work with runs a different stack, and there’s almost always at least one system in the mix that nobody wants to touch. Sometimes it’s a legacy tool with no documented API. Sometimes it’s homegrown software built a decade ago by someone who’s long since moved on. Sometimes it’s a big vendor platform where connecting anything new means a lengthy procurement cycle before you write a single line of code.
Whatever the reason, the effect is the same: good ideas die in the integration conversation before they get a chance to prove themselves. Nobody’s going to greenlight a six-figure integration project to test a use case for a two-week seasonal spike. Nobody’s going to fight for budget to automate a workflow that only three people run. So the idea just doesn’t happen, no matter how good, because it isn’t worth the cost of connecting it.
A New Automation Strategy: Computer Use Agents (CUA)
This is where computer use agents (CUAs) change the math. Instead of plugging into a system’s back end, a CUA works the front end the way a person would. It sees what’s on the screen, decides what to do, clicks, types, reads the results, and moves on to the next step.
Most automation runs somewhere you can’t see it. But a CUA operates out in the open, on the same screens your team already uses.
Why This Is Becoming A New Strategy For CX Organizations
Think about it: no API, no integration project, no waiting on a vendor to build you a connector before you can start.
If your team switches tools next year, you don’t rebuild the automation from scratch. You show it the new screen and the new workflow, and it adapts.
And when it comes to security, a CUA is scoped the same way you’d scope a new hire. Access on an as-needed basis, approved ahead of time, and nothing more. The CUA doesn’t get to wander into systems it wasn’t trained and cleared to touch.
What This Looks Like in Practice
Here’s a straightforward example of a CUA in action, using Laivly’s Sidd Solo. A return request comes in through voice, chat, or email, doesn’t matter which. Sidd talks to the customer directly, pulls up the order, checks the return policy, issues the label, updates the record, and closes it out. And none of it requires an integration project, an API, or a months-long buildout before the first return gets processed.
Because Sidd runs in the cloud, it’s not limited to one instance handling one conversation at a time. Multiple copies can work through a queue in parallel, which means the same automation that handles one return can handle a hundred without you needing to add infrastructure.
And this isn’t only about automating full processes or reducing headcount. The same underlying automation mechanism also shows up in Sidd Duo agent assist, running the lookups and the busywork alongside a live associate in real time. It’s automation that shows up wherever it’s needed. Sometimes that means running the whole task, and sometimes it’s just clearing the friction out of a human agent’s way.
So Here’s My Question for You
What in your stack doesn’t have an API? And what is that actually costing you? Not in dollars necessarily, but in ideas you’ve stopped bringing up because everyone already knows how that conversation ends?
Well, let’s change how that conversation ends. Reach out and chat with me about how Sidd can make all those good ideas happen after all.
Jesse Greenberg is Head of Strategic Accounts at Laivly.
