01
Step 1
Gather the video's audio, the product terminology list with exact spellings, the required caption language, and the merchant's reading-time rule for how long each cue may stay.
Back to topic
Spark playbook scenarios · Spark: AI Store Assistant
Create accurate captions for product videos
Videos are published without captions, so muted viewers miss the explanation and accessibility requirements go unmet. In this case, the immediate task is to create accurate captions for product videos. The workflow should produce a caption file matched to the actual audio, with product names and technical terms flagged against the merchant's records.
Topic
Product Video Creation and Editing
Target keyword
add subtitles to product videos
Copyable AI prompt
A prompt you can drop directly into Spark
First replace every [bracketed] placeholder with real store scope and constraints, then add any actual records or data you have, and paste it into your preferred AI assistant or Spark agent. Spark will follow this page's workflow to produce structured output and verification steps.
ai-prompt.txt
Act as a Shopify operations specialist.
Help me create accurate captions for product videos.
My store context is [describe the business, products and target audience].
My scope is [specify the affected products, pages, customers or reporting period]. Use these inputs: [the video's audio, the product terminology list with exact spellings, the required caption language, and the reading-time rule per cue]. Transcribe the audio into cues aligned to the spoken timing, keeping measurements, safety notes and exclusions verbatim while dropping only filler, and flag every cue containing a product name or technical term for terminology review.
Return a caption file with flagged terminology listed for review for the specified task.
Show the relevant evidence or before-and-after examples and explain how I can verify the result.
If essential data is missing, ask for it and mark unsupported conclusions as unknown.
Do not invent metrics, product claims or customer facts.
For any publication, message, account or store change, use only connected tools within my explicitly authorized scope and report success only after verification; otherwise provide the draft or proposed steps.How to solve
A recommended resolution workflow
Split the work into input, judgment, and deliverable phases. Verify each step's output before moving to the next.
02
Step 2
Transcribe the audio into cues aligned to the spoken timing, keeping measurements, safety notes and exclusions verbatim while dropping only filler, and flag every cue containing a product name or technical term for terminology review.
03
Step 3
Deliver the caption file with the flagged-terms list, then verify cue-by-cue that timings match the audio, no cue overlaps its neighbour, and every flagged term is spelled exactly as in the records.
FAQ
Common questions when running this scenario
Cover decision rules, frequent misunderstandings, and required inputs so the team aligns on prerequisites before execution.
Why keep measurements verbatim in cues?
Measurements, safety notes and exclusions are the facts viewers act on, and rounding or simplifying them in captioning changes the product claim; filler words may be dropped because they carry no recordable fact.
How is cue timing verified?
Each cue's start and end are checked against the corresponding spoken passage, and the reading-time rule bounds the minimum display duration, so no cue is too brief to read or overlaps its neighbour.
What happens to inaudible passages?
They are marked as unintelligible in the file and listed for re-recording rather than guessed, because an invented line in captions looks as authoritative as the spoken audio it replaces.
Related scenarios
Continue with adjacent or same-topic scenarios
Start with scenarios under the same topic, then expand into cross-topic matches suggested by similarity.
Next step
Want these AI scenarios embedded directly into daily operations?
Ciwi Spark builds Shopify-native AI tools and agents around the workflows merchants actually run: research, ads, translation, and customer support. Stop rewriting the same prompts by hand.