Hi, Libby - as Kathy said, the "what's working" was pretty uncanny. Or was it easily discovered? Hard to tell. It does seem like it repeats the same adjectives. For example, I noticed 'intriguing' on this thread, and also in mine.

The characters are colorful, quirky, and compelling. Their dynamics and relationships are complex and intriguing.

- The dialogue is sharp, humorous, and reveals the distinct personalities.

- The time period details of the late 80s/early 90s are fun and nostalgic.

- The scavenger hunt plot device is a clever way to propel the characters into funny and surprising situations.

- The story explores meaningful themes about friendship, loss, life choices, and personal growth.

- The ending montage provides satisfying resolution for the main characters.

Question for Jack from Greenlight - I'm not so much interested in Greenlight's opinion on how to improve the script, as I've had attachments before and received a ton of notes and have rewritten it for certain people, etc. - as I am about how much industry cache a good score has.

You have a quote from Scott Mednick who is a pretty big-time producer. Does that mean if I get a good score, it's a marketing tool with which to approach someone like Scott?

Thanks,

AJR

Hi all. Since AI seems to be the flavor of the week, thought I’d post my own findings to maybe quell or stoke anyone’s interest should they be thinking of using an AI service for script feedback.

I recently used the Greenlight AI Feedback service, and to date… it is the only feedback service I’ve ever paid for regarding a screenplay. Human or AI. Considering the script is about 7 years old and I’ve never had feedback on the script in its entirety, I thought what the heck, let’s see what AI does with 130pgs of surreal dreamscape.

Anyway, the end result for what I paid is more than satisfactory for what I was looking for. Alot of groovy feedback for the price, and it (most importantly) showed me where the script is and is not working, err… optimally.

As well, it allowed 3 additional questions after feedback regarding areas of the script, areas that I was really curious to have detailed with another’s opinion to see if it would give me additional ideas for those scenes, or information that would allow me to build on or reconstruct similar scenes elsewhere in the script.

It even spit out a succinct logline that pretty much resonates with the script.

Original logline (that it never prompted for):

“Inspired by the surreal works of Lewis Carroll and L. Frank Baum: A small town girl who dreams of traveling to distant lands, suddenly finds herself on an epic quest to release her lost love from the prison of an underworld garden matrix.”

Greenlight AI Logline:

“After a young woman enters a magical realm to find her lost love, she must defeat the evil overlord who has imprisoned him and restore balance to the mystical gardens.”

So, yeah, anyway that feedback is here if anyone is curious what Greenlight AI Feedback looks like, as well as the original script if anyone’s curious as to what the AI is actually giving feedback on…

Feedback: https://www.dropbox.com/scl/fi.....dl5ja26vwcj&dl=0

Script ‘The Arpeggio Garden’: https://www.dropbox.com/scl/fi.....909wt35e0o2&dl=0

Hopefully this helps, best of luck…

Thanks, AJR!
All of that just seems really generic to me.


I'm interested in the answer to this question too.

I'm curious to know if this is going to become the way of the future e.g. getting a minimum score from AI before you can progress with certain things.

I still think the test of this would be to have AI access a really badly written script.

Thanks LC! I have answered AJR's question in another reply, but we will be sharing our industry connections very soon =)

Personally I have been in the industry for about a decade, worked on 30+ produced movies ranging from super indie ones <$100k budget to Nicolas Cage movies budgeted over $20m (I was responsible for his casting decision on that one too). Evaluating screenplays and getting feedback have always been a difficult problem for me.  I certainly hope we can make this process better and help good screenplays get discovered and produced, as well as helping existing scripts to be improved.



We have done a few bad movies for sure, I will link a blog post that has "The Room" coverage report here if you want to check it out: https://glcoverage.com/2024/01/27/what-is-script-coverage-script-coverage-examples-and-insights/

Cheers!

I'll say this...the AI LOVES them some Timothee Chalamet. That, or, maybe we should all be on the alert...Timothee Chalamet may actually BE an AI bot. Which, when you think about it, makes a lot of sense.



Question for Jack: I didn't see the typical PASS, CONSIDER, RECOMMEND. Is that included in the coverage?

LOL re: Timothee Chalamet.

As for the Pass Consider Recommend, I also heard this from many users, so we are working on a section right now to include that in the future.  May take us a few weeks, but it will be there for sure.

Thanks Jack for coming in and having this conversation.

The casting recommendations sound like a value-added feature that could help screenwriters like me who have an interest in writing, but no interest in investing the time and energy in following A-, B- and C-tier celebrities.

Other possibilities:

• Generating comp films if it can be done with decent confidence. Probably only relevant for features.
  
  
• Expandable explanations for why certain feedback is present. I expect this would be an extra fee, but could go a long way in building trust in the model. (AKA "Explainable AI")
  
  
• Recommend/Consider/Pass but with an explanation that this is not some specific "industry pro" rendering an opinion. Making this expandable would be valuable as well, especially if it's not a straightforward point system.
  
  
• Summaries about characters or pseudo-characters (narrators, MacGuffins, etc.), again expandable if that's feasible.
  
  
• Comments on the script's structure. "Your script appears to be structured according to the Hero's Journey, but...". Again, probably only for features. The most you can say about a short is that it has a beginning, middle, and end.
  

I understand these are all things that are way easier to say than do... take them as idle suggestions from someone who knows a little about AI but not enough to help implement any of those ideas.