← All Case Studies
Case Study · Retail Data Platform

From months of mapping to a single session.

How a retail technology platform replaced a multi-cycle, spreadsheet-driven hierarchy build — with an AI-assisted tool the customer drives directly.

"If I could have had this in previous roles, it would have been game-changing."

— Customer Sales Lead
Engagement Product Build & Rollout Domain Catalog Onboarding Catalog Range 100–50,000 SKUs Estimated Impact 30–45 Days Saved

Cutting catalog onboarding from cycles to a single sitting.

Inconsistent catalogs, no shortcuts.

A retail technology platform required every new customer's product catalog to be mapped into a fixed 7-level product hierarchy — from Division down to SKU — before any downstream system could use the catalog.

Customer data arrived in inconsistent shapes: non-standard CSVs, multiple files needing merging, missing levels, and naming drift ("gray" vs. "grey"). The internal team produced a draft, sent it out, waited, reconciled comments, revised — two to four cycles per customer. The work scaled badly past 50,000 SKUs and consumed expert time on judgment the customer was best positioned to make themselves.

An AI-drafted hierarchy the customer drives.

We built an AI-assisted tool that ingests the customer's catalog — from a Shopify integration, uploaded files, or both — and proposes a complete mapping into the 7-level hierarchy as the starting point.

The customer reviews in a visual interface with SKU counts and product images at each node, accepts or edits suggestions in bulk, and approves. On approval, the structure is locked and writes directly into platform configuration — no manual handoff between customer sign-off and downstream availability.

Four capabilities, one guided workflow.

From messy catalog to locked hierarchy — the customer in the driver's seat the whole way.

Ingest

Shopify + CSV Merge

Pulls product titles, variants, tags, collections, and metafields; merges with uploaded CSVs to form a single normalized catalog.

Propose

7-Level AI Draft

Clusters SKUs by category and attribute patterns; produces a complete Division→SKU hierarchy with confidence-weighted suggestions.

Refine

Visual Review UI

Nested, expandable tree with product images and SKU counts at every node. Bulk accept, retag, or ignore branches in one click.

Override

Manual Edits Persist

Field-level edits available everywhere. Manual overrides supersede AI logic and survive re-proposals.

Built, in production, and measurable from day one.

Figures below are estimates, measured per customer as rollout continues.

Time Saved
30–45 days

Per customer onboarding vs. the prior draft-and-reconcile cycle.

Review Cycles
2–4 0–1

First one to two cycles eliminated; remaining rounds shortened.

Accuracy Target
≥ 90%

AI mappings accepted without modification, measured per customer.

Time to Approve
≤ 10 min

Target end-to-end, from upload to locked hierarchy.

Customer-driven, not internally drafted

The customer reviews and approves directly — no spreadsheets, no email cycles, no expert judgment standing in for theirs.

All segments, all sizes

Same workflow for Shopify-native and direct-format customers, across catalogs from ~100 to 50,000 SKUs.

Validation enforced before lock

Structural rules block incomplete or contradictory hierarchies from ever reaching production.

Clean handoff to the platform

Locked structures write directly into configuration. No manual internal step, no "we'll fix it later."

Four stages. One locked hierarchy.

01 / Ingest

Catalog In

Shopify pull, CSV upload, or both — merged into a single normalized view.

Ingestion
Shopify API CSV Catalog Merge
02 / Propose

AI Drafts

Complete 7-level hierarchy generated with confidence-weighted suggestions.

AI & Logic
LLM Proposal Confidence Scoring Attribute Clustering
03 / Refine

Customer Reviews

Visual tree, bulk actions, manual overrides — all in one interface.

Review UI
React Tree View Bulk Actions Role-Based Access
04 / Lock

Direct to Config

Approved hierarchy validates, locks, and writes straight to the platform.

Validation & Lock
Structural Rules Approval Workflow Direct Config Write

From spreadsheets
to production.

Jahnel Group builds AI-assisted product workflows for data-heavy businesses.