How a Glass Manufacturer Replaced Spreadsheets With AI

A glass-container plant offloaded its manual spreadsheets onto a governed warehouse, moving defect and downtime tracking from monthly retrospectives to shift-level visibility.

Shift-level

Replaced monthly plant reporting

Implementation Time

Unmukt Raizada

Founder & CEO

TrustEvals

Connect ↓

The Challenge

The facility relied on manual spreadsheets for daily plant data, with no shift-level visibility into operational efficiency. Placeholder and inconsistent job entries defeated naive ingestion, and analysis was limited to monthly retrospectives rather than actionable shift-level insight.

What They Built

TrustEvals built an end-to-end ingestion-to-dashboard platform: Python ingestion via the Sheets API, a PostgreSQL warehouse with row-level security, and web dashboards. Engineered KPIs include furnace capacity utilization and glass pack efficiency, plus shift-level defect and downtime tracking and a shift-approval application.

A custom dirty-data job-mapping engine resolves placeholder and inconsistent entries using a composite business key, so KPI calculations stay reliable. The stack moves plant operations off spreadsheets onto an automated, warehouse-driven decisioning model.

AI Role

The core intelligence is a custom dirty-data job-mapping engine that resolves placeholder and inconsistent entries using a composite business key, so KPI calculations stay reliable. Engineered KPIs — furnace capacity utilization, glass pack efficiency, and defect and downtime — are computed in a governed warehouse and surfaced at shift level. No generative model is named in the source.

AI Model

Python data ingestion via Google Sheets API, PostgreSQL warehouse with row-level security, web-based dashboards, custom composite-business-key data-cleaning engine, shift-approval application

Infrastructure

Google Sheets (source data via API) • PostgreSQL data warehouse with row-level security • Web-based operations dashboards • Shift-approval application

Integration Points

Python ingestion from the Google Sheets API into the warehouse • Composite-business-key cleaning engine normalizing job entries • Engineered KPIs feeding web dashboards

Impact

Shift-level tracking

Defect and downtime tracking now runs at shift granularity instead of monthly aggregation.

Dirty data resolved

A composite-business-key engine fixes placeholder and inconsistent job entries that defeated naive ingestion.

Governed warehouse replaces sheets

A PostgreSQL warehouse with row-level security and engineered efficiency KPIs replaces manual spreadsheets.

Technology Utilized

Implementation Complexity

Time Savings

Best Fit For

Best fit for manufacturers stuck on spreadsheet-based reporting whose messy operational data needs cleaning before it can drive shift-level decisions.

Unmukt Raizada

Founder & CEO

TrustEvals

Founder & CEO of TrustEvals. Builds AI evaluation and governance infrastructure for finance, real estate and regulated software — eval harnesses, semantic data dictionaries, and AI audits.

Get an intro

Industry:

Manufacturing & Industrial

Business Function:

Operations

Company Size:

Project Cost:

Cost not disclosed

Ownership:

Organization Type:

AI Pattern:

Data Synthesis & Reporting

Value Type:

Time Savings

AI Model:

Frequently Asked Questions

How can a glass manufacturer move from spreadsheets to shift-level plant analytics?

The manufacturer worked with the experts to build an ingestion-to-dashboard platform on Python, a PostgreSQL warehouse with row-level security, and web dashboards. A custom composite-business-key engine resolves dirty job data, so defect and downtime tracking now runs at shift granularity instead of monthly aggregation.

What tools and data approach powered the shift-level plant analytics?

The build centered on data synthesis and reporting: Python data ingestion via the Google Sheets API, a PostgreSQL warehouse with row-level security, engineered efficiency KPIs, web dashboards, a shift-approval application, and a composite-business-key cleaning engine. No specific AI model is named in the record.

What results did the glass manufacturer achieve?

Defect and downtime tracking now runs at shift level instead of monthly, a composite-business-key engine fixes placeholder and inconsistent job entries that defeated naive ingestion, and a governed PostgreSQL warehouse with engineered efficiency KPIs replaces manual spreadsheets.

How long did the plant analytics platform take to build?

The record does not specify a timeline. At the time of the case the solution was in user acceptance testing rather than confirmed full production.

Who is a shift-level plant analytics platform best for?

It is best suited to manufacturers stuck on spreadsheet-based reporting whose messy operational data needs cleaning before it can drive shift-level decisions.