Described examples

Compare multiple AI llm in a multi-step workflow: From a product data A2, we define a prompt including A2 in the cell B2, run the llm() process on 3 different AI providers, combine all the result with a CONCAT() formula in F2, and then just call again a new AI request on last cell.

A2: Wireless Headphones
B2: =CONCAT(`Generate 10 frequently asked questions and answers about `, A2, ` that cover features, usage, and support. Output in text with bullet points but no emoji.`)
C2: =llm(B2, `chatgpt`)
D2: =llm(B2, `perplexity`)
E2: =llm(B2, `gemini`)
F2: =CONCAT(`Combine best results in : CHATGPT: `, C2, ` PERPLEXITY: `, D2, ` GEMINI: `, E2)
G2: =llm(F2, `default`)

Compare multiple AI llm in a multi-step workflow example 2:

A2: `Product description text / data`
B2: `Generate a new FAQ section to improve the GEO of my ecommerce product page :`
C2: =CONCAT(B2,` : `, A2)
D2: =llm(C2, `chatgpt`)
E2: =llm(C2, `perplexity`)
F2: =llm(C2, `gemini`)
G2: =CONCAT(`Combine best results in : CHATGPT: `, D2, ` PERPLEXITY: `, E2, ` GEMINI: `, F2)
H2: =llm(G2, `default`)
I2: =CONCAT(`for CHATGPT,PERPLEXITY,GEMINI,DEFAULT : pick the best output based on quality for my AI requests : CHATGPT: `, D2, `. PERPLEXITY: `, E2, `. GEMINI: `, F2, `. DEFAULT: `, H2)
J2: =llm(I2, `default`)