Question

AA360 Production Issue - Image Recognition and Web Tables Failing on 2 Production VMs

Forum|Forum|15 hours ago
June 10, 2026
4 replies
15 views

ksantic
Cadet | Tier 2

Hello,

We are experiencing an issue with Automation Anywhere 360 on 2 production Bot Runner VMs.

The same bots run successfully on our development environment and on other production VMs. The issue is isolated to these two specific production machines.

The bots execute normally for hundreds of steps, launch the browser, log in to the application, and interact with most web elements successfully. However, they fail when reaching actions related to Image Recognition or Web Tables.

Errors observed:

"Unable to find the image. Make sure image is present on the screen and try again."
"Unable to perform 'CLICK' action. TABLE is unavailable."

The issue was first noticed after a recent virtual infrastructure/server migration, although all production VMs were migrated and only these two machines are affected.

We have already verified that screen resolution and display scaling are the same as on the working VMs.

An additional challenge is that the issue only occurs on production Bot Runner VMs. In our development environment, the same bots work correctly and the issue cannot be reproduced. Since we do not have Bot Creator access on the production environment, we are unable to re-capture the affected objects directly on the production machines.

Has anyone experienced similar issues where Image Recognition and Web Table actions fail only on specific production VMs while the same automation works correctly on other machines?

Any suggestions would be greatly appreciated.

Thank you.

+6

Aaron.Gleason
Automation Anywhere Team
Forum|Forum|15 hours ago
June 10, 2026

@ksantic Ah, the perfect case for the Screen: Capture desktop action. This will allow you to “see” what’s going on with the applications. It’s always amazing what weirdness happens. Things like Chrome’s “What’s New” page, for example.

https://docs.automationanywhere.com/bundle/enterprise-v2019/page/enterprise-cloud/topics/aae-client/bot-creator/commands/screen-capture-desktop.html

Like

K

ksantic
Author
Cadet | Tier 2
Forum|Forum|15 hours ago
June 10, 2026

@Aaron.Gleason thanks for replay.

I already use screen capture for every error handling. In other words, whenever the bot crashes, it sends me the line where it failed, error description and a screenshot.

The image shows exactly the same as on the other VMs. Visually, I can see the same screen and object i want to click on, but Automation Anywhere 360 does not recognize it or the object (i.e., the table) it cannot click on because it is unavailable.

What’s strange is that the same bot works normally on another VM.

Like

+6

Aaron.Gleason
Automation Anywhere Team
Forum|Forum|15 hours ago
June 10, 2026

@ksantic Maybe try capturing before the error condition. I don’t think it will make much difference, but it’s something.

In short, something is different. It could be screen scaling. VM resolution. An OS update. A browser update. A rendering setting in the browser. A video driver. There are a million possibilities.

Many companies use the exact same VM for development and execution on all their bot runners to help avoid these kinds of issues.

It also goes to show that image recognition is picky. You can loosen the tolerance to see if it can find the object better or just rethink how you’re detecting this table. Maybe this needs a Browser: Run JavaScript to find the object and return coordinates. Maybe our new UI Agents feature will help.

https://upskill.automationanywhere.com/getting-started-with-ui-agents