PIING
From order history to a useful customer profile
How Piing: Intelligence analyses selected Shopify data, checks confidence and writes the fields you choose back to the customer.
In this article
Start with the evidence you actually have.
A run of knitwear purchases can be useful evidence of a preferred category. A single gift order is a weaker basis for deciding what someone will want next. Intelligence makes those distinctions reviewable instead of hiding them inside a segment.
The app reads selected Shopify order data, applies deterministic rules and AI analysis, and writes approved outputs to app-owned customer metafields. It can work independently of Piing's Wallet and store-recognition app.
A captured example from the app. Each value shows its source, with controls to keep, edit or remove it. The 19 fields shown belong to this example customer; the ready-made catalogue contains 22.
Different fields need different methods.
Some values can be calculated from orders, such as purchase frequency or average days between orders. Others, such as preferred style or likely next category, are inferences.
Intelligence includes 22 ready-made fields: nine AI-inferred, nine deterministic and four optional pixel-derived fields. Custom fields and prompts are available on Scale and above. Select the fields that answer a real question for your team.
Customer contact details are not needed in the inference prompt. The output is a structured field with its source and, where applicable, confidence, rather than a new marketing contact record.
Decide what is strong enough to write.
AI-inferred fields include a confidence score. You choose the threshold a result needs to meet before it is written. A suggestion below the threshold can remain out of the live customer profile.
Example write threshold: 80%. You choose the fields and the confidence required.
Review samples before building a campaign or staff workflow around a field. More complete order history may provide better evidence, but it does not remove the need to check that the labels are useful for your category.
Browsing data is optional.
The optional storefront pixel adds first-party browsing signals where consent and the configured identity match allow it. It uses Shopify's customer privacy controls.
A page view is not the same as a purchase, and an anonymous session is not automatically a known customer. Treat browsing-derived fields as another source of context, with their own limits.
Use the fields where the team already works.
Selected outputs are written to Shopify customer metafields. Shopify Flow, compatible CRM connections and metafield-aware staff tools can use them when configured to read those fields.
Start with one use: a reviewed category preference for a segment, or selected context for a store team. Choose which fields each destination should receive instead of sending the whole profile everywhere.
Keep the results reviewable.
Manual and scheduled runs can refresh selected fields as the available data changes. Run history shows progress and outcomes. Timing depends on the number of customers, their order history and the fields selected.
New plans count one successful customer profile refresh, regardless of enabled fields. Compare how often your customers need fresh profiles; failed refreshes, five-customer previews and metafield repairs do not use the allowance. Existing field-unit subscriptions retain their approved terms.
Read the setup documentation or explore Intelligence and its plans.