A merchant has several analytics views that disagree because each report defines its metrics differently. In this case, the immediate task is to use AI to interpret store analytics evidence. The workflow should produce an interpretation with its evidence trail and the checks that would confirm each hypothesis, drawn from the merchant's own records.
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.
Act as a Shopify operations specialist.
Help me use AI to interpret store analytics evidence.
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 disputed metric's numerator and denominator per report, each source's deduplication and attribution rule, the timezone and currency used for timestamps and totals, the shared merchant-defined window, the segment breakdown, and the export or report link behind every figure]. Recompute every figure from its own definition inside one shared merchant-defined window, label as definitional any gap that disappears once both sides are rebuilt because the underlying exports agree, and drop any measure whose source export lacks the fields needed to reconstruct it.
Return the interpretation with its evidence trail and every figure verified against the merchant's own records 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.Split the work into input, judgment, and deliverable phases. Verify each step's output before moving to the next.
Pull, for every report that touches the disputed metric, its definition stated as numerator and denominator, the deduplication and attribution rule behind it, the timezone and currency used for timestamps and totals, the reporting window, the segment breakdown requested, and the export or report link each figure came from.
Recompute every figure from its own definition inside one shared merchant-defined window, label as definitional any gap that disappears once both sides are rebuilt because the underlying exports agree, and drop any measure whose source export lacks the fields needed to reconstruct it.
Deliver the interpretation as an evidence trail naming for each claim the metric, period, segment, source and definition behind it, list one check per hypothesis with the record that would confirm or kill it, and reconcile the totals against the order or session export before publishing.
Cover decision rules, frequent misunderstandings, and required inputs so the team aligns on prerequisites before execution.
It can only list hypotheses. Without a comparison window, a segment breakdown and the metric's own numerator and denominator there is nothing to recompute, so the output stays an opinion rather than a finding tied to the merchant's records.
Each figure is rebuilt from its own definition inside one shared window first, and any gap that then disappears is labelled definitional because the underlying exports already agree; only differences surviving that rebuild are treated as movement worth investigating.
Every claim carries the metric, period, segment, source and definition behind it plus one check another operator can rerun, and the totals reconcile with the merchant's own export; an interpretation missing either element goes back to collection.
Start with scenarios under the same topic, then expand into cross-topic matches suggested by similarity.
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.