𝗦𝗮𝗺𝗲 𝗣𝗿𝗼𝗺𝗽𝘁, 𝗙𝗼𝘂𝗿 𝗔𝗜 𝗧𝗼𝗼𝗹𝘀, 𝗢𝗻𝗲 𝗖𝗿𝗶𝗰𝗸𝗲𝘁 𝗕𝗮𝗻𝗻𝗲𝗿
I tested four AI tools on a new task.
A cricket coach needed a banner for his academy. He wanted a simple graphic to post in WhatsApp groups. No code. No websites. Just a flyer.
I used the same four tools: ChatGPT, Claude, Gemini, and Grok.
The results were different for every tool.
𝗖𝗵𝗮𝘁𝗚𝗣𝗧 𝘄𝗼𝗻 𝘁𝗵𝗲 𝗶𝗺𝗮𝗴𝗲 𝘁𝗲𝘀𝘁. It understood the brief immediately. It produced clean, professional banners in two prompts. It followed instructions and kept the design simple. It gave me something I could use right away.
𝗖𝗹𝗮𝘂𝗱𝗲 𝘁𝗿𝗶𝗲𝗱 𝘁𝗼 𝘄𝗿𝗶𝘁𝗲 𝗰𝗼𝗱𝗲 𝗮𝗴𝗮𝗶𝗻. Claude is great at building websites. But when I asked for a banner, it gave me an HTML file. It built a nice web section, but you cannot post a web page in a WhatsApp group. It tried to turn a design job into an engineering job.
𝗚𝗿𝗼𝗸 𝘄𝗼𝗻 𝘁𝗵𝗲 𝘃𝗶𝗱𝗲𝗼 𝘁𝗲𝘀𝘁. Grok failed at the static image. It made a cluttered design with too many fonts and photos. However, when I asked it to make the banner move, it excelled. It turned a still image into a high-quality video with a voiceover. It looked professional, though I had to check the text for errors.
𝗚𝗲𝗺𝗶𝗻𝗶 𝗳𝗮𝗶𝗹𝗲𝗱 𝗯𝗼𝘁𝗵. Gemini refused to make the image due to guardrails. Later, it made a video that had nothing to do with my banner. The text was gibberish.
𝗧𝗵𝗲 𝗹𝗲𝘀𝘀𝗼𝗻𝘀:
• The best coding tool is not the best design tool. Use the right tool for the specific medium.
• Guardrails matter. If a tool refuses simple creative tasks, it loses utility.
• Done is better than impressive. I prefer a tool that gives me a finished product over a tool that shows a flashy demo but requires heavy editing.
Do not look for one model to rule them all. Use ChatGPT for stills and Grok for motion.
Which tool do you use for graphics? Has your favorite coding agent ever tried to turn a design task into a coding task?
Optional learning community: https://t.me/GyaanSetuAi