Industrial Data Labs

Starting with business problems

By leveraging data and implementing machine learning, they sought to automate key steps, streamline workflows, and enhance productivity. This case study highlights the approach used and the resulting improvements in response times and efficiency.

Mapping Existing Workflows

Industrial Data Labs began by conducting a detailed assessment of the current workflow for responding to RFQs. This included observing the entire process from receiving a proposal to sending a quote, identifying manual steps, and noting bottlenecks. By mapping the existing workflows, they gained a clear understanding of areas where time and resources were being underutilized.

Identifying Useful Data

After mapping the workflows, the team focused on collecting and analyzing relevant data points that were critical in generating accurate responses to RFQs. This data included customer details, historical pricing patterns, product specifications, and lead times. Identifying and categorizing this data was crucial to train the models and streamline decision-making.

Identifying a Model

Industrial Data Labs identified a machine learning model that could be trained on the categorized historical RFQ data to predict optimal responses. The model was a simple classifier that could take an unstructured part discription and break it into part attributes.

Quantify Potential Time Savings

The team then developed and tested new workflows incorporating the trained model to automatically fill in quote details and suggest responses. Early prototypes indicated a potential reduction in response time by up to 80%, enabling inside sales reps to focus more on high-priority tasks and relationship management.

Reinforcement Learning

To improve the model's accuracy, a feedback mechanism was incorporated to capture user edits and adjustments. Each time a sales rep modified the model's suggestion, the system logged the change and reason, allowing for continuous reinforcement learning and more precise responses over time.

Create a Proof of Concept

A proof of concept was created using a limited set of RFQs to demonstrate the effectiveness of the new system. The prototype integrated seamlessly into the existing sales platform and was able to handle a diverse set of RFQs, showing a significant reduction in time spent on repetitive tasks.

Product Rollout

After successful testing, the refined product was rolled out to the entire inside sales team. The implementation included training sessions and support for sales reps to ensure they were comfortable using the new automated solution. Adoption was quick, as the interface was designed to be user-friendly and aligned closely with the reps' existing workflows.

Identify Key Success Metrics

The primary success metric identified was the average time saved per RFQ response, with a secondary focus on the accuracy and completeness of automated responses. Post-rollout, the average response time dropped from 2 hours to 15 minutes per RFQ, while maintaining a high accuracy rate.

Monitoring and Iteration

Industrial Data Labs implemented a system for continuous monitoring and iteration to ensure the model remained accurate and up-to-date. Regular feedback sessions with the sales team and data analysis helped identify areas for improvement and refine the model further.

Results

By reengineering the RFQ response workflow, leveraging existing data, and implementing a machine learning model, Industrial Data Labs significantly reduced the time spent responding to RFQs. The automated system not only improved response efficiency but also empowered sales reps to focus on strategic tasks. With a continuous feedback loop and regular updates, Industrial Data Labs is well-positioned to maintain and enhance its competitive advantage in the industry.