Time Required to Create a Claygent Prompt: Poll Options

·Nov 19, 2025 10:09 PM·

How much time does it take for you to create a good Claygent prompt? 📊 Please vote below! . 1️⃣ Less than 30 seconds 2️⃣ Around 1 minute 3️⃣ Around 5 minutes 4️⃣ 10+ minutes .

4️⃣8

1️⃣2

15 comments

· Sorted by Oldest

Abdul S.
·
·
Less than 1 minute 🙂
1
🔥2
Marina H.
·
·
Depends on the complexity of the task, and whether by good you mean the first version or the tested and revised one that takes into account the outliers :)
👀1
Bharat D.
·
·
Marina H. - Right! I'm trying to get a sense of working version of the prompt... 🙂
🪄1
Laz G.
·
·
Easy prompt: Around 5 min. Even the simple ones need to be tweaked. For high-critical prompts? 10+ minutes for sure.
🎯1
Bilal R.
·
·
Around 1 minute if your internet is good
👍1
Ken M.
·
·
I’ve spent hours or days on critical scraping prompts, compared 100 results, manually checked, then looked again etc
🙌2
Arpit C.
·
·
Have you folks tried using the metaprompter as a starting point?
🙋‍♂️1
Laz G.
·
·
Yes, but it’s mid. Claude is way better for complex prompts. (No offense).
Arpit C.
·
·
Haha none taken!
Arpit C.
·
·
It’s good feedback
Iveta A.
·
·
Oh, boy! Recently I had a couple of days to create not even perfect but quite good for around 1500 contacts. What worked for one account didn't work for others. Any tips to improve this? Or is this normal - spending days to create a prompt? (Will check out metaprompter).
Bharat D.
·
·
Iveta A. - interesting. What information are you trying gather and which model are you using?
Iveta A.
·
·
Bharat D. I am extracting (starting point I have email addresses and comany domain and company LI for some), I need to extract full name: first: the name to provided email, second: if the email is generic, then search domain and LI for decision maker, then serach online to find the best match. I have a narrow niche, hence not all companies have an official online presence and I am saving costs and using GPT 5 Nano (I proposed to my company that I could help with research and all costs are my own atm). The return results vary when I manually checked them on a 20 row batch and for days I was improving my prompt. 😏
Bharat D.
·
·
Iveta A. - 4o mini (not O4) works great for this. What other data points do you have for the company? It's always good to feed as many data points as possible for AI to get accurate information. For example - Company name, Location, industry etc.
Iveta A.
·
·
Thanks for 4o mini! And yes, I have other data points, like company address, city, state, project name and project type. But I guess I need to improve my prompting skills 🙂
1

Abdul S.
·
·
Less than 1 minute 🙂
1
🔥2
Marina H.
·
·
Depends on the complexity of the task, and whether by good you mean the first version or the tested and revised one that takes into account the outliers :)
👀1
Bharat D.
·
·
Marina H. - Right! I'm trying to get a sense of working version of the prompt... 🙂
🪄1
Laz G.
·
·
Easy prompt: Around 5 min. Even the simple ones need to be tweaked. For high-critical prompts? 10+ minutes for sure.
🎯1
Bilal R.
·
·
Around 1 minute if your internet is good
👍1
Ken M.
·
·
I’ve spent hours or days on critical scraping prompts, compared 100 results, manually checked, then looked again etc
🙌2
Arpit C.
·
·
Have you folks tried using the metaprompter as a starting point?
🙋‍♂️1
Laz G.
·
·
Yes, but it’s mid. Claude is way better for complex prompts. (No offense).
Arpit C.
·
·
Haha none taken!
Arpit C.
·
·
It’s good feedback
Iveta A.
·
·
Oh, boy! Recently I had a couple of days to create not even perfect but quite good for around 1500 contacts. What worked for one account didn't work for others. Any tips to improve this? Or is this normal - spending days to create a prompt? (Will check out metaprompter).
Bharat D.
·
·
Iveta A. - interesting. What information are you trying gather and which model are you using?
Iveta A.
·
·
Bharat D. I am extracting (starting point I have email addresses and comany domain and company LI for some), I need to extract full name: first: the name to provided email, second: if the email is generic, then search domain and LI for decision maker, then serach online to find the best match. I have a narrow niche, hence not all companies have an official online presence and I am saving costs and using GPT 5 Nano (I proposed to my company that I could help with research and all costs are my own atm). The return results vary when I manually checked them on a 20 row batch and for days I was improving my prompt. 😏
Bharat D.
·
·
Iveta A. - 4o mini (not O4) works great for this. What other data points do you have for the company? It's always good to feed as many data points as possible for AI to get accurate information. For example - Company name, Location, industry etc.
Iveta A.
·
·
Thanks for 4o mini! And yes, I have other data points, like company address, city, state, project name and project type. But I guess I need to improve my prompting skills 🙂
1