Getting the fastest support (and where AI gets in the way)

Updated 2 minutes ago by Tristan Payne

AI tools can be incredibly useful, but they have one major limitation: they've never seen your account, your server, or your logs. Despite that, they'll often present a confident diagnosis as if they've inspected your environment directly.

A guess presented as a diagnosis can send an investigation down the wrong path. By the time we've untangled the assumptions and worked back to the actual symptoms, we've often spent longer than if the ticket had simply described what happened in the first place.

Why we ask this

Much of what we run across our hosting platform, email services, and our own products, such as Onyx (which powers Managed WordPress), is built in-house or configured in ways that are unique to Krystal.

AI models are trained on public information and common platform configurations. They haven't been trained on your Krystal account or our internal infrastructure. When you ask an AI "what's likely causing this?", it's usually describing how a completely different platform might behave, then presenting that explanation with confidence.

If that diagnosis is pasted into a support ticket, our engineers first need to verify or rule out the AI's assumptions before they can begin investigating the actual issue.

In almost every case, a blank page is more helpful than a confident but incorrect theory.

Please don't ask an AI to diagnose the problem for you

We'd much rather receive the raw symptoms than someone, or something's, best guess at the cause.

An incorrect diagnosis doesn't just add extra reading. It actively influences the direction of the investigation, which can slow down getting to the real answer.

Please don't paste long AI-generated troubleshooting guides

These often arrive as several paragraphs of plausible-sounding explanations, with the genuinely useful information buried somewhere in the middle.

The details we actually need are usually simple:

• The exact error message
• The affected URL
• When the problem started
• What changed, if anything
• Steps to reproduce the issue

Sending those directly is far more useful than pages of AI-generated analysis.

Please do tell us exactly what you saw

You're not expected to diagnose your own hosting environment.

A simple description of what happened is exactly what our engineers need to begin investigating.

For example:

Less helpful

"This is likely a server-side caching issue combined with a PHP memory limit or possible database connection cap. I recommend checking configuration files and object cache settings..."

More helpful

"Since 2pm today, mysite.com/shop shows 'Error establishing a database connection'. Other pages are fine. Nothing has changed recently."

The second example gives us concrete information we can investigate immediately. The first requires us to separate facts from speculation before we can begin.

In short

You know what you saw better than any AI does.

Tell us what happened, what you saw, and when it started. Leave the diagnosis to us. If you're unsure what information to include, please see the How to contact Krystal Sales and Support guide on the Do's and Don'ts to help us help you.

How did we do?