Skip to main content

Why your prompt matters

This section is about getting consistently good results from K-AI. By asking it the "right" questions. K-AI is smart, but it can be brilliant if you prompt it well.

Good answers start with good questions​

K-AI can see your data, but it can't read your mind. A vague request like "What's wrong, check it and just fix it" gives a poor result more often than not, because K-AI has to guess what you actually care about. Ask in a clear, specific way and you get a far better answer, far more reliably.

The good news is that prompting K-AI well is a simple, repeatable skill. You don't need special syntax or technical terms. You mostly need to tell K-AI four things.

The anatomy of a good prompt​

Strong prompts tend to cover four things. You won't always need all four, but they are worth keeping in mind:

  • Goal: what you're trying to achieve or decide.
  • Context: which assets, process, or part of your model you mean.
  • Scope: the slice that matters, like a time range, a segment, or a single process.
  • Output: the form you want the answer in, like a number, a list, a chart, or an explanation.

Compare:

Weak: "Why are my orders slow?"

Strong: "In the order process, which step has the longest average wait over the last quarter? I'm trying to find where to focus on cutting cycle time."

The strong version names the process (context), sets a time range (scope), and says what you're really after (goal), so K-AI can go straight to a useful answer instead of having to ask what you mean.

Brief K-AI like a sharp colleague​

Here's a useful way to picture it: you're briefing a sharp colleague who knows Konekti and can see your workspace, but is new to your specific question. You don't need to be polite or formal. You do need to be clear about what you want and why.

Refine, don't restart​

You rarely have to get it perfect in one go. K-AI remembers the conversation, so if the first answer isn't quite right, tell it what to adjust: "now only completed orders" or "break that down by region." Refining in the same conversation is faster than starting over, and it builds on what K-AI already worked out.

In this section​

Hit next to learn more about the best practices.