Agentic-Aware Testing: Winning in a World of AI-Powered Users

“We have some new capabilities… agentic workers that are coming in, essentially to take all of the alerts that we generate… and do all the assessments, the reviews, the writing of the reports, the submission of those reports — takes down the workload by 80 percent.” — Adena Friedman, President and CEO of Nasdaq, Inc.

As I was finishing my toast and coffee, this was the conversation Adena Friedman was having this morning with Becky Quick and Andrew Ross Sorkin on CNBC’s Squawk Box. She was talking about how Nasdaq Verafin is deploying agentic AI workers inside financial services. Not for one-off use cases. Not in theory. These digital workers are already doing real jobs across customer environments.

They review alerts. They draft reports. They complete due diligence tasks that used to take entire teams. And they’re doing all of it autonomously. She said this shift is cutting workloads by more than 80 percent.

That number stuck with me. It wasn't the first time I had heard it from an executive.

I've read several articles in McKinsey, American Banker, and AML Intelligence that were circling around the same idea — that agentic AI is scaling faster than most of us are ready for. Podcasts I listen to each week are telling me the same thing. Reading and listening to those helped reinforce the sense I already had. The volume of work being done by agents on the web is rising at a crushing pace.

The web is about to change

Here’s what’s coming. Over the next five years, a growing percentage of traffic hitting websites and applications won’t be people. It’ll be agents. AI agents working on behalf of people. They won’t be browsing. They’ll be executing.

They’ll arrive with intent. With instructions. With a goal to complete. That goal might be reviewing a report, booking travel, submitting a form, scraping a data table, checking an invoice, verifying an account, placing an order.

And they won’t just show up once. They’ll come by the millions. Quietly doing work that used to require a team of humans.

We’ve only tested for human behavior

The challenge is this: for decades, the testing industry has operated under a single assumption. That humans are the users.

Humans were the ones clicking around, submitting forms, and navigating flows. So we designed tests based on human journeys. We imagined what a human might do. Then we wrote test cases to match that journey. If we were lucky, we automated a few of them.

But that model is starting to break.

Because agents don’t behave like humans. They move differently. They optimize for speed, not usability. They follow the shortest path, not the one your designer intended. They navigate the application in ways you didn’t predict — and frankly, couldn’t have predicted.

Which means this: humans creating test cases for agentic behavior won’t work. You’ll miss things. And those misses will create friction and failure for a growing portion of your [agentic] user base.

Agentic-ready apps will have the advantage

Here’s where it gets interesting. Agents will start to show preferences. They'll work faster on apps that are designed with their workflows in mind. The less friction they encounter, the more likely they'll be to complete their work and return.

That means the apps that work better for agents will become the default. They’ll win the agentic traffic. Which means they’ll win the businesses and customers behind that traffic.

Companies will want their agents to work with apps that get the job done right the first time. They won’t tolerate failed flows, broken buttons, or misaligned fields. The standard will go up...and fast.

So what do we do about it?

We stop guessing. We stop imagining what agents might do. And we start observing what they’re already doing.

This is the case for a new kind of testing. One that doesn’t rely on human assumptions. One that doesn’t limit itself to human flows. One that sees both.

This is Agent-Aware Testing.

And right now, Katalon is the only company doing it.

With TrueTest, Katalon can see what’s happening in production. It watches real users. It watches real agents. And it captures both types of behavior — turning them into tests automatically.

It sees the human flows and the agentic flows. It doesn't guess at what those journeys might be. Then it creates tests for both. You do it without writing scripts or guessing at behavior.

That’s how you keep up. That’s how you stay competitive. And that’s how you make sure your application is ready for what’s already happening — not just what used to happen.

This shift is already underway

If you’re building a modern SaaS app, an enterprise workflow, a customer portal, a B2B integration layer, or anything else on the public or private web, this shift is coming for you.

Like so many innovation waves we've lived through, it’s not a matter of if. It’s a matter of when the agentic traffic shows up. And whether your app performs when it does.

If you’re not testing for agent behavior now, you’re falling behind.

But the good news? You can catch up. TrueTest can help.

Look into it. Understand what Agent-Aware Testing means for your business. The sooner you start, the faster you get ahead.

And make no mistake. The agents are already knocking. Will your door be open?