Product descriptions:
Steelcase, a global leader in the world of work, faced mounting challenges in managing pricing across their expansive product portfolio. Legacy approval processes were slow and manual, stretching the sales teams thin. By unifying data and deploying AI models on the Databricks Data Intelligence Platform, Steelcase significantly increased their pricing workflow automation. Instant pricing recommendations replaced challenging 20-minute pricing submissions, boosting auto-approvals from 35% to nearly 50% and helping teams respond faster to complex, high-stakes commercial deals.
Turning pricing complexity into deal agility with data and AI
Steelcase is a global design and thought leader in the world of work. The organization designs and manufactures furnishings and solutions for the many places where work happens. With an extensive product portfolio and diverse customer base in 130+ countries, managing pricing across a vast number of projects presented a complex challenge. Traditionally, Steelcase relied on a manual, legacy pricing approval process characterized by time-consuming workflows and inconsistent decision-making. Sales teams spent about 20 minutes on average per pricing submission, grappling with a pricing ecosystem involving tens of thousands of project-customer-product combinations each year, intricate discount configurations and multiple stakeholder types, including dealers, architects and corporate real estate firms.
The complexity was not only in sheer volume, with over 70,000 pricing scenarios annually, but also in the nuanced interplay of discounts, product mix and participants involved in each deal. The manual approach introduced inefficiencies that hindered the ability to swiftly identify price-sensitive opportunities or adjust pricing tactics in real time. Prior attempts at automation relied on predefined authorization thresholds, which resulted in concentrated pricing activity around administrative checkpoints instead of strategic pricing decisions.
From a technology standpoint, Steelcase faced challenges related to seamlessly integrating and operationalizing models for pricing approvals within a fast-paced commercial environment. Managing huge datasets of up to 60–70 million rows and delivering pricing recommendations rapidly required advanced feature management, real-time inference and collaborative data science capabilities. Previous tools lacked the ability to unify diverse data sources, orchestrate complex workflows and deliver model outputs with the necessary speed and consistency.
“We needed a smarter, faster way to respond to complex pricing scenarios across thousands of deals,” Katie Rey, Consulting Data Scientist and Agile Product Lead at Steelcase, said. “Manual processes just couldn’t keep up with the pace of business or the scale of our data.”
Intelligent pricing starts with a unified data platform
The Steelcase data science team designed and deployed an AI-powered pricing engine built end to end on the Databricks Data Intelligence Platform on Azure. This comprehensive solution unified massive data volumes from critical sources, ingested and orchestrated using Azure Data Factory and Delta Lake on Azure Data Lake Storage Gen2 and structured with the Databricks medallion architecture to ensure quality and freshness.
Key components included the Databricks feature store, which enabled the preprocessing and caching of over 70 predictive features derived from large-scale datasets. By precomputing these features and updating them weekly, the system minimized online calculation overhead, sustaining inference-in-second speeds crucial for the pricing approval workflow. Collaborative development thrived using Databricks Notebooks, allowing multiple data scientists to share work seamlessly, iterate rapidly and leverage MLflow for experiment tracking and model lifecycle management. The solution deployed random forest models via serverless real-time inference to serve instant discount recommendations directly into Steelcase’s internal Pricing and Contracts Management (PCM) application through API endpoints.
Unity Catalog played a vital role in providing centralized governance, secure data access and streamlined data cataloging across the engineering and science teams. This enabled strict data lineage tracking and consistent reproducibility, key for compliance and operational confidence. For reporting and evaluation, dashboards leveraged live tables maintained by Databricks and directly connected to Tableau, allowing near real-time monitoring of model performance and adoption metrics.
“Working in Databricks made it possible to handle massive data with speed and collaboration. The feature store was a game changer, allowing the model to respond in seconds, even with complex, multisource features,” Katie said.
Eric Anderson, Data Science Manager at Steelcase, added, “Databricks provided the full-stack platform we needed, from feature engineering to model deployment to monitoring, allowing us to build an automated pricing solution unmatched in our industry.”
Accelerating deal velocity with data and AI
The Steelcase AI-powered pricing system fundamentally transformed the way the company manages commercial operations at scale. By automating previously manual pricing workflows, including discount recommendations that once took 20 minutes per request, Steelcase boosted automatic approvals from 35% to nearly 50%. This advancement enabled faster deal execution, reduced friction across sales channels and empowered teams to respond more quickly to complex deals involving diverse product bundles and multiple stakeholders.
Ben Krill, Vice President of Pricing and Incentives at Steelcase, emphasized the shift in how teams work. “AI-driven pricing recommendations have dramatically improved the speed and competitiveness of our pricing process. We reclaimed over 1,500 hours annually that salespeople can now spend working with customers rather than pricing, and reduced unproductive bias in our pricing approvals to win more.”
The AI models also supported more consistent pricing decisions at scale, helping Steelcase identify and respond to market dynamics with greater agility. Technologically, the solution leveraged the Databricks Data Intelligence Platform on Azure to provide a robust, scalable architecture capable of handling concurrent pricing projects and growing data volumes. This infrastructure now supports ongoing innovation across Steelcase data science initiatives, including D3 (the Data-Driven Design program), which brings AI-powered insights into dealer designer tools to generate product recommendations and automated floor plans.
Eric reflected on the broader transformation: “This project is a clear example of how embedding AI and robust data management on Databricks continues to unlock new business models and competitive advantages for Steelcase.”