daBongo LMS AI Training Courses

Lesson Content

Common prompt issues and their causes.

Wrong style: art style descriptor missing or overridden by stronger style cues. Wrong mood/atmosphere: lighting or mood descriptor absent or too weak. Subject not matching intent: subject description too vague or conflicts with style. Unwanted elements: common associations of a word or phrase adding content not intended. Composition unexpected: no camera direction or framing descriptor to guide composition.

Diagnosing before iterating.

Before changing the prompt, look at the generation result and identify specifically what is wrong: is the style wrong, the subject wrong, the atmosphere wrong, or the composition wrong? The diagnosis should match the fix – changing the art style descriptor when the problem is the lighting direction wastes an iteration.

Iteration targeting.

Match the change to the diagnosis: wrong style -> change or add the art style descriptor; wrong mood -> change lighting and mood descriptors; unwanted element -> add specific exclusion language or be more explicit about what should NOT be present; wrong composition -> add framing and camera direction language.

When to abandon vs. iterate.

Sometimes a prompt direction fundamentally conflicts with the intended image – the combination of style, subject, and setting doesn't produce results in the right direction after 3-4 iterations. In these cases, rebuilding the prompt from a different conceptual starting point is more efficient than continuing to iterate on a direction that isn't working.