How to Use ContentBot.ai Bulk Mode for Product Copy

November 11, 2026

ContentBot.ai bulk mode takes a spreadsheet of inputs and runs one prompt across every row, producing a column of generated output. 200 product descriptions typically render in 5 to 15 minutes and consume around 20,000 to 40,000 words of allowance.

It’s the feature most people underuse. Single-prompt generation gets the marketing attention, but the bulk runner is what turns ContentBot.ai from a writing assistant into a catalog tool.

What does bulk mode actually accept?

A CSV or pasted table. Each column becomes a variable you can reference in the prompt — typically product name, features, material, price and a category.

Row count per run is capped by plan, and the word allowance caps total output long before that. Output lands in a new column you can export as CSV.

How do you structure the input table?

Five columns are enough for product copy.

Column Example Notes
name Merino Crew Sock Exact product name
features moisture-wicking, arch support, 3-pack Comma separated, 3-5 items
material 80% merino wool Matters for claims
price $18 Used in copy sometimes
category running accessories Sets vocabulary

Keep the feature column short and factual. Long feature lists produce descriptions that list, rather than describe.

How do you write the prompt template?

Template it like this: “Write a 60-80 word product description for {name} in the {category} category. Use these features: {features}. Mention the material {material} once. Do not use exclamation marks. Do not invent specifications not listed.”

Two clauses do the heavy lifting. Word range stops one row returning 40 words and the next returning 200. Don’t invent reduces the fabricated-spec problem that bulk runs otherwise amplify across hundreds of rows.

Run the template on 5 rows first. Read all five out loud. If any sound like a different brand, tighten the category vocabulary before running the full set.

How do you QA hundreds of rows?

Sampling, not full reading. Read the first 5, a random 10 from the middle, and the last 5. That’s 20 out of 200 and it catches most systematic problems.

Check for four failures:

  1. Invented specs — a weight, a certification or a material that wasn’t in the input.
  2. Word count drift — anything obviously over or under the range.
  3. Banned punctuation — exclamation marks creep in regardless of the instruction.
  4. Duplicated phrasing — if rows 40 through 60 all open with “Crafted for”, the prompt needs a variation instruction.

Failure four is the one that gets products rejected by marketplaces for thin or duplicate content.

What is the word allowance math?

Bulk runs consume allowance by output words. 200 products at 70 words each is 14,000 words, plus retries. Plans start around $19 a month with a modest allowance — enough for a few hundred products, not a full catalog.

For a 5,000-SKU catalog, you’re looking at 350,000 output words minimum, which means an upper tier or splitting across months. There’s no per-run pricing to soften that; it’s pure allowance consumption.

Where does bulk mode break?

On anything requiring judgment across rows. Variant products — the same shirt in six colors — get six near-identical descriptions, which is exactly what you don’t want for SEO or for a marketplace duplicate check.

On technical products, too. Specs that must be exact (battery capacity, dimensions, compatibility lists) should not pass through a generative step at all. Generate the prose around the specs, inject the specs from data.

Best fit: apparel, home goods, food and other categories where the copy is descriptive and the specs are few. Worst fit: electronics, industrial parts, anything with regulatory language.

Quick Summary

  • Bulk mode runs one prompt over a CSV and renders 200 product descriptions in 5-15 minutes.
  • Five input columns are enough; keep features to 3-5 factual items.
  • Set a word range and a no-inventing clause in the template, then test on 5 rows.
  • Sample 20 of 200 rows for invented specs, count drift and duplicated phrasing.
  • Allowance scales with output words — a 5,000-SKU catalog needs an upper tier.