Workflow Examples

Real-world formula patterns to get the most out of Batcher.ai. Each example shows the exact formulas to use in your spreadsheet.

Simple Workflow — Product descriptions

Generate a product description from a product name.

A2: Wireless Noise-Canceling Headphones
B2: =LLM(CONCAT("Write a compelling product description in 50 words for: ", A2), "default")

Advanced Workflow — Full SEO package

Generate multiple SEO assets from a single product description.

A2: Wireless Noise-Canceling Headphones with 30h battery life
B2: =LLM(CONCAT("Write an SEO title (max 60 chars) for: ", A2), "default")
C2: =LLM(CONCAT("Write a meta description (max 160 chars) for: ", A2), "default")
D2: =LLM(CONCAT("List 10 SEO keywords for: ", A2), "default")
E2: =LLM(CONCAT("Write a FAQ section with 5 questions and answers about: ", A2), "default")
F2: =LLM(CONCAT("Write a compelling product description in 100 words for: ", A2), "default")

Compare AI — Multi-provider comparison

Run the same prompt through multiple AI providers and combine the best results.

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.")
C2: =LLM(B2, "chatgpt")
D2: =LLM(B2, "perplexity")
E2: =LLM(B2, "gemini")
F2: =CONCAT("Combine the best elements from these three outputs: CHATGPT: ", C2, " PERPLEXITY: ", D2, " GEMINI: ", E2)
G2: =LLM(F2, "default")

Advanced variant — Pick the best output

Extend the comparison by adding a final judge step:

A2: Product description text
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: CHATGPT: ", D2, " PERPLEXITY: ", E2, " GEMINI: ", F2)
H2: =LLM(G2, "default")
I2: =CONCAT("Pick the best output based on quality: CHATGPT: ", D2, ". PERPLEXITY: ", E2, ". GEMINI: ", F2, ". DEFAULT: ", H2)
J2: =LLM(I2, "default")

Multi-step Workflow — Data enrichment pipeline

Sequential processing where each step builds on the previous result.

A2: Raw product name
B2: =CONCAT("Enrich this product name with category and brand: ", A2)
C2: =LLM(B2, "default")
D2: =CONCAT("Verify and standardize this enriched product data: ", C2)
E2: =LLM(D2, "perplexity")
F2: =CONCAT("Generate SEO optimized description based on verified data: ", E2)
G2: =LLM(F2, "default")
H2: =CONCAT("Add FAQ section and GEO optimization to: ", G2)
I2: =LLM(H2, "default")

Web Scraping — Extract content from URLs

Use the Jina integration to scrape web pages and process the content.

A2: https://example.com/products/wireless-headphones
B2: =LLM(A2, "jina")
C2: =LLM(CONCAT("Summarize the key product features from this page content: ", B2), "default")

Note: The "jina" configuration converts a URL into markdown content. You can then pass that content to another AI model for analysis.

SEO Gap Analysis — Competitor research

Combine Haloscan (SEO data), Jina (web scraping), and AI processing for comprehensive SEO analysis.

A2: Your product description text
B2: =LLM(CONCAT("Extract the primary keyword (2-3 words max) from: ", A2), "default")
C2: =LLM(B2, "haloscan")
D2: =LLM(CONCAT("From this JSON data, extract the first 3 competitor URLs separated by | : ", C2), "default")
E2: =TRIM(SPLITSTR(D2, "|", 0))
F2: =TRIM(SPLITSTR(D2, "|", 1))
G2: =TRIM(SPLITSTR(D2, "|", 2))
H2: =LLM(E2, "jina")
I2: =LLM(F2, "jina")
J2: =LLM(G2, "jina")
K2: =LLM(CONCAT("Analyze SEO gaps between my product: ", A2, " and competitors: ", H2, " | ", I2, " | ", J2), "default")

Note: "haloscan" returns SEO data (search volumes, keyword rankings, SERP results) as JSON. Process it with another AI call to extract actionable insights.

Duplicate Content Detection

Check and rewrite product descriptions to avoid duplicate content across your catalog.

A2: Product Title
B2: Existing product description
C2: =LLM(CONCAT("Description to check: ", B2, ". Compare with known descriptions and identify duplicate phrases. If none found, return: NO_DUPLICATE"), "default")
D2: =IF(C2="NO_DUPLICATE", B2, LLM(CONCAT("Rewrite only the duplicate sections: ", C2, ". Original: ", B2), "default"))

Translation Workflow

Translate content into multiple languages with a single data row.

A2: Original product description in English
B2: =LLM(CONCAT("Translate to French: ", A2), "default")
C2: =LLM(CONCAT("Translate to Spanish: ", A2), "default")
D2: =LLM(CONCAT("Translate to German: ", A2), "default")

Tips for building workflows

Start with row 2 — Row 1 is always for column headers. Write formulas in row 2 and they will be duplicated to all other rows.
Test with a few rows first — Process 3-5 rows to verify your formulas before running the full batch.
Use CONCAT for prompts — Build prompts that reference your data dynamically instead of hardcoding values.
Chain formulas step by step — Do not try to do everything in one formula. Break complex tasks into multiple columns.
Choose the right model — Use "default" for simple tasks, and premium models (ChatGPT, Claude) for tasks requiring higher quality.