Bad prompt or bad generator?

[EMC1972]

Hi, I'm new to firefly but have been experimenting with video AI for a few months now. One of the features I'm looking for is to take an existing graphic look and copy that look with some simple animation design. This has worked really well in Chatgpt for still images but need a video solution. I tried another program with little success, so now it's Firefly's turn. However, I meeting with even lesser results and am now wondering if its the prompt or the limitation of the firefly platform to create the prompt. I was able to generate a still image in chatgpt based on the prompt but as I would like to add some animation like type-on text I need a motion generator. Here's the prompt I used and I will attach the file to show what it produced. If anyone knows a better way to go about this please post. 
using the text in this image, create a new graphic. replace the world "ARRIVES" with the word "WORLDS".

Each letter should type on. The text should have a subtle drift which scales up to make the image 10% larger. Change the text to say "WORLDS". Each letter should appear to animate on with a type on style effect

[Kevin Stohlmeyer]

@EMC1972 Have you tired the third-party models like Ray3 (SD) or Veo? They tend to handle text and replacing words better than Firefly base model. 

[Oh.N8]

Hey @EMC1972 

Just to add a bit more to what @Kevin Stohlmeyer  mentioned. Consider using Adobe Express to animate your text.

Benefits
1) Can change text font and resize the text later (non distructive)

2) You can animate the text in various ways (non distructive)
3) You essentially have flexibility

Short Vid how to add text in Adobe Express

Cheers

Nate