Transforming B2B quoting with platform integration and machine learning
Modernized B2B quoting through CAD integration and machine learning, reducing manual effort, improving accuracy and accelerating revenue growth.
Client background
For nearly 50 years, the client has grown into an international leader in lighting technology, manufacturing innovative lighting solutions across multiple industries.
The business challenge
The client engaged Baker Tilly’s Digital Solutions team to guide the development of a new customer-facing web portal designed to reduce customer support call volume while increasing satisfaction and efficiency.
When the organization sells theatre equipment to dealers, distributors and sales representatives, users at those organizations must collaborate on complex projects — often spanning multiple years — to design, install and support systems across interconnected platforms. The first step in this process is generating a quote, either independently or by submitting a request to the client’s Quotations team.
The client supported two quoting paths based on project complexity. For complex projects — such as multi-phase, multi-space installations over several years — quotes were created using a Computer-Aided Design (CAD) tool to develop system designs and bills of materials. For simpler projects, customers used a self-quoting tool. However, this self-quoting tool was not integrated with the CAD system and required users to manually assemble systems from extensive part lists and adjust quantities one by one. Users also had to select an installation type from more than 40 options to route the quote to the correct sales manager; a step frequently completed incorrectly, causing routing inefficiencies.
Strategy and solution
Baker Tilly collaborated with the client’s engineering teams to embed new self-quoting functionality directly into the CAD tool using a WebView user interface. This in-app experience allows users to visually design systems and generate their own quotes without leaving the CAD environment or manually selecting products from long-part lists.
Within the WebView, users complete a quoting form powered by a machine learning model that predicts the appropriate installation type based on system names and historical project data. This automation significantly improves routing accuracy to the correct sales manager. Users then complete the remaining quote fields, preview the quote and generate it directly within the CAD tool.
Results and impact
Within the first month, more than 300 quotes were submitted through the new platform, generating over $1 million in revenue from orders placed without employee assistance. Six months after release, revenue reached $8 million, outperforming the legacy platform. Based on the client’s internal time-to-quote benchmarks, the solution has saved more than 600 hours of employee quoting effort.
By combining a machine learning predictive model with rigorous user interviews and testing, the enhanced self-quoting tool eliminated manual back-office steps and enabled the client’s Quotations team to focus on higher-priority work and customer inquiries.
