Building a Repeatable Supplier Import System for PIB Sales
Wippy
Runtime
NetSuite-Ready
Imports
Client
Independence
Solutions
Industries
Technologies
About the Project
PIB Sales, a US-based precision bearing distributor, needs every supplier part listed online, because customers search by part number and leave when a part is missing. Onboarding one supplier meant hand-translating manufacturer spreadsheets into NetSuite import files, which had taken months and depended on the client’s catalog lead. Spreadsheet edits and chat-based AI copy-paste broke down at scale and kept none of those decisions for the next import. Spiral Scout, which had previously rebuilt PIB’s eCommerce platform and ERP integration, shipped a supplier data preparation system on Wippy with reusable mappings, dedicated parsers, and NetSuite field sync. Several separate NetSuite imports per supplier file became one reviewed import file that the PIB Sales team prepares itself.
Objectives
- Standardize supplier inputs: Convert flat spreadsheets and separate product-and-attribute tables into records suitable for NetSuite export.
- Reuse operational knowledge: Save mappings, defaults, and supported description templates for subsequent imports.
- Reflect ERP fields: Pull NetSuite custom item fields into the system so supplier columns can be mapped to them.
- Apply explicit transformations: Handle supplier-specific units, part-number conventions, descriptions, and identifiers.
- Keep the client’s team in control: Provide editable mappings, sample validation, export previews, and configurable column order.
- Preserve the ERP handoff: Generate one consolidated CSV file for PIB’s established NetSuite import process.
Challenges
Solutions
Supplier files described products in incompatible ways
One manufacturer provided a conventional product table. Another separated product identity from thousands of attribute records. Others encoded specifications inside part numbers or supplied paired imperial and metric measurements.
One supplier workbook alone contained 12,306 products and 133,578 rows across two sheets. Renaming columns could not turn that structure into one export record per product.
Separate structural preparation from field mapping
We merged product and attribute tables, added supplier-specific unit selection, and dedicated part-number parsing before the final export.
The bearing parser extracts dimensions and specifications directly from part numbers. UPC values are converted to text with non-digit characters removed, so leading zeros survive the export.
This separation lets each supplier’s conventions be handled explicitly while retaining a shared mapping and CSV output process.
Supplier terminology had to match a changing NetSuite field catalog
PIB’s NetSuite account contained 300 custom item fields. A static mapping list would require maintenance whenever destination fields changed, while similar field names made manual matching difficult.
Bring NetSuite’s field list into the preparation system
We connected the system to NetSuite’s API to pull the ID and label of every custom item field. Wippy stores that list and shows the labels in the mapping screen, along with how many fields changed in each sync.
The integration has an explicit handoff: API reads bring in field names and selected records, and generated CSVs carry prepared supplier data into the existing NetSuite import process. Supplier product updates are not written back through the API.
Description generation needed both flexibility and repeatability
Supplier descriptions varied in structure and completeness. Each product category needed its descriptions split into the same fields in the same order, and copying rows through a chat assistant broke down at catalog scale.
Batch AI generation, with template rules where the data allows
AI generates descriptions for most suppliers. The system groups repeated source descriptions into batches and re-runs any rows the model skipped, so large files run without manual copy-paste. AI also proposes the initial field mappings for the team to review.
For one supplier whose catalog data is structured enough, descriptions come from adjustable template rules instead. A saved template can combine housing bore, shaft diameter, and seal width in the required order and unit format.
The PIB Sales team can edit the template rules and inspect both populated and missing fields before generating the file.
Repeat imports depended on specialist memory
Knowing which source column matched an ERP attribute was only part of the work. The team also needed the correct manufacturer defaults, description rules, category values, and export order. Repeating those decisions created a dependency on whoever had prepared the previous file.
Make recurring decisions configurable and reusable
We added named mapping templates, saved supplier settings, editable defaults, and drag-and-drop export ordering. Supported description templates can also be saved and reused. Export columns can be renamed and reordered, and one source field can feed several export columns, for example, when one image file name has to fill several image columns.
Sample validation identifies missing required fields and two source columns mapped to the same NetSuite field. The preview applies the export transformations to sample rows, so the team can check the result before download.
The PIB Sales team can change every configuration in the system interface without an engineer. Only a new supplier with a custom file structure requires engineering work.
Strategy
We treated repeat imports and client ownership as core design requirements. Shipping the first export mattered, but the architecture also had to preserve the decisions needed to prepare the next supplier update.
Phase 1: Establish the Data Structure
We separated product records, attribute records, identifiers, units, and destination fields. A column-by-column review of each source file made missing values and structural differences visible before mapping.
That foundation determined where the system needed a table merge, a unit rule, a parser, or a direct field assignment.
Phase 2: Encode Supplier Methodology
We translated repeatable decisions into supplier presets, mapping templates, description rules, and dedicated transformations.
This expertise automation approach gives operational knowledge a reusable representation. AI handles description writing and mapping suggestions, while explicit rules handle structure, units, and identifiers.
Phase 3: Put Routine Operation in the Client’s Hands
We exposed the controls needed for repeat work: selecting source data, refreshing destination attributes, applying saved mappings, editing defaults, checking samples, and ordering export columns.
The PIB Sales team can adapt configured imports without requesting a code change for each adjustment.
Project Results and Impact
PIB Sales now has a shared preparation workflow across multiple supplier formats. Supplier files that previously went through several separate NetSuite imports now produce one consolidated import file. Each new supplier with a supported file structure becomes a configuration the PIB Sales team owns, not a new engineering project. The lasting value is the combination of reusable rules, a synced list of NetSuite fields, and controls the PIB Sales team runs itself, which together reduce dependence on reconstructing supplier knowledge from previous spreadsheets.
Tangible Outputs:
– Supplier ingestion: XLSX and CSV upload, parsing, and source-data inspection.
– Reusable configurations: Supplier presets, named mappings, defaults, and supported description templates.
– Dedicated transformations: Merging product and attribute tables, unit handling, bearing part-number parsing, and UPC preservation.
– NetSuite connectivity: Read-only synchronization of custom item fields and selected source records.
– AI assistance: LLM integration for mapping suggestions and batched description generation.
– Review controls: Sample validation, editable fields, and transformed export previews.
– CSV generation: One consolidated import file with renamable, reorderable columns.
– A purpose-built operations interface: Custom software development focused on recurring catalog preparation.
Key Takeaways
- Encode the Decisions That Make the Data Usable: The engineering work extended beyond reading spreadsheets. Product identity, measurement conventions, attribute relationships, and destination-field requirements needed explicit treatment. Encoding those decisions makes the workflow repeatable and its behavior easier to inspect.
- Preserve Supplier Differences Within a Shared Workflow: The same data ingestion pattern applies wherever multiple external catalogs feed an established ERP: normalize each source’s structure, apply its specific rules, map to destination fields, and expose the output for review. Shared export mechanics can coexist with specialized parsers.
- Measure Independence Through Client Control: Client independence becomes concrete when the PIB Sales team can reuse mappings, refresh field catalogs, adjust defaults, edit supported templates, and inspect exports themselves. Those controls place routine preparation decisions with the people responsible for the catalog.
Worth a conversation if you’re building a supplier data preparation system and want a second set of eyes before the architecture hardens? Get in touch
Related Projects
-
Multi-Tenant Platform Architecture for a Restaurant Costing System
LinkOffichef went from a single-restaurant spreadsheet tool to a multi-tenant system built to run for many operators at once.
-
Agentic Data Controls for CarAdvise’s Automotive Marketplace
LinkEngineered a configuration system replacing tribal knowledge with enforced rules, deployed to distributors without IT drag.
-
Turning Complex Quoting into Controlled, Repeatable Systems
LinkImplemented CPQ that cut quote generation time by 88% and made quoting consistent and reliable.
-
Salesforce AI Integration for Fortress (Legal SaaS)
LinkAutomated 80% of deal management tasks in Salesforce, cutting manual work and boosting efficiency in legal workflows.
-
Secure Enterprise AI Automation for RFPs
LinkDeployed AI agents that cut RFP response time from 4 days to a few hours and pre-sales overhead by 30%.
-
Automating Supplier Data Mapping for Sourcing Value
LinkDeveloped automated data workflows that reduced manual effort by 90% and increased processing speed.
Have a Similar Data Workflow to Automate? Let’s Discuss
Meet the founders
Tell us your goals
Receive a proposal
Project kickoff
John Griffin
Co-Founder, CEO
Anton “JD” Titov
Co-Founder, CTO
“Anton is an exceptional technologist. I would feel comfortable having him work on any technical challenge.” – Ryland Goldstein, Head of Product, Temporal
