Retail Training With AI: Using AI To Build Confidence In Retail Associates

The Impact Of Retail Training With AI On 100,000 Associates

As a boutique custom learning company, Cinecraft Productions has always been committed to designing high-quality eLearning solutions that align with our 7 Better Learning Principles: authentic, timely, accessible, relevant, engaging, fun, and efficient. When a global retailer with nearly 6000 stores and 100,000 store associates approached us for help modernizing their retail training, we saw an exciting opportunity to leverage Artificial Intelligence (AI) to meet their needs.

Learning Strategy

The retailer employed the same sales process for 15 years. The sales process was effective, but unfortunately it was underutilized because it had too many steps that caused confusion. So the retailer condensed the sales process into three straightforward steps. Ultimately this change would result in an increase in the average shopping basket value.

To achieve their goal, we recommended a blended approach including a behavioral modeling video, video-based simulations, and refresher scenarios with an AI-generated coach that provides instant and authentic feedback.

Because the new sales process is meant to guide associates rather than provide a script, we recommended using a dynamic, AI-driven approach for the refresher simulations. Associates write in their own responses to the customer rather than select a multiple-choice option. A custom Language Learning Model (LLM) that was trained on the sales process powers the feedback for these simulations. Almost like a real training coach, training the LLM (also called a model) allows it to provide specific feedback based on what associates type in for their answers. This approach helps build associates’ confidence and allows them to get personalized feedback.

Retail Training With AI: How Did We Do It?

There were many factors to consider as we developed the AI coach for the refresher simulations. In addition to our standard process, here are the steps we followed to create an effective, secure solution.

Step 1: Determine The Client’s Needs

This required a thorough analysis of their existing IT infrastructure along with their legal and security requirements.

The client did not have an existing AI platform but wanted to host the new AI solution within their existing infrastructure. This requirement necessitated a robust and adaptable platform that could integrate seamlessly into their current ecosystem while maintaining full client control over the environment.

To meet these requirements, the AI platform and all associated data had to be securely sandboxed to ensure that the client maintained ownership and governance of their data and workflows. In addition, we proposed using an intermediate server to ensure the safety of data processing and minimize risks. This ensures that the learners’ responses and AI feedback remain secure and private.

Step 2: Determine The Best Technology

The next step was to select the right technology for integrating AI into the retailer’s sales process training. Like all effective learning solutions, accuracy and responsiveness are key. The AI model had to provide relevant and immediate feedback to associates to support an engaging and dynamic training environment.

To ensure quality standards, we tested multiple AI models to determine which provided the most accurate results and the quickest response times. This rigorous evaluation process allowed us to select the model that best aligned with the client’s needs for efficiency and precision.

While effective AI integration can be costly, it’s influenced by two primary factors: the amount of data input into the model and the number of queries or users accessing the service. To navigate these variables and find the most efficient solution, we created a detailed cost matrix. This matrix evaluated various configurations and usage scenarios to determine the optimal balance of performance and cost-effectiveness for the client’s specific use case.

The solution we chose ensured affordability without compromising on quality, providing a scalable solution that aligned with the client’s budgetary requirements and operational goals.

Step 3: Determine The Technical Workflow

The hardware retailer wanted to use Articulate Storyline 360 to build their course, which required us to figure out a secure way for learners to interact with the AI through the Storyline interface. After extensive research and discussions, we implemented the following workflow:

• Type answer into storyline – The learner watches a video of a customer entering the store or asking a question and types their response into a Storyline course.
• Intermediate server processing – The learner’s answer is securely sent to a server that is owned and controlled by the client for preprocessing.
• AI platform processing – The intermediate server sends non-sensitive data to the AI platform, which generates feedback that is contextually relevant to the learner’s answer. Sensitive information is stored on the intermediate server and not passed on to the AI platform.
• Intermediate server processing – The AI’s feedback is sent to the intermediate server, where it is refined and formatted for delivery back to Storyline.
• Feedback delivery to Storyline – The learner receives immediate and actionable feedback from the AI coach directly within the Storyline training module.

This behind-the-scenes process occurs every time the learner answers a question, and it takes only seconds to complete!

Step 4: Train The Model

We needed the AI model to act as our hardware retailer’s ideal in-store performance coach for associates. This means we had to teach it everything about the client’s new sales process, as well as other expected behaviors for associates, systems, and resources they might use on the job. This was a meticulous process. Instead of developing a custom model, we used a base model from our AI platform. We provided detailed instructional context to align with the retailer’s specific goals. This included training the model to recognize industry-specific terminology, common customer scenarios, and the retailer’s policies and procedures. This content was outlined in a scenario grid and full storyboard similar to our process for regular training courses.

Step 5: Test The Model

After providing all of this information to the model, we needed to make sure it was trained effectively. If the AI coach provided answers that were consistent with the client’s goals, then we were successful! If not, then we needed to retrain the model by providing different information. The testing process was first conducted by users who were familiar with the training content but were not store associates. After refining the model, the simulations were launched as a pilot for a select number of store associates to try out. They provided their thoughts on the training’s usability, relevance, and feedback accuracy.

Step 6: Refine The Model Based On Feedback

The testing phase revealed areas for improvement. For example, we needed to refine the model’s responses to better match the retailer’s communication style and ensure consistent tone and accuracy. After multiple iterations and adjustments, we achieved the desired performance and learner satisfaction levels.

Conclusion

The integration of AI into retail associate training proved transformative for this global hardware retailer. By leveraging cutting-edge technology alongside sound Instructional Design principles, we created a scalable solution that increased associate confidence, improved customer service, and delivered measurable business results. For Learning and Development professionals exploring AI, this case study highlights the importance of thoughtful implementation and a commitment to quality eLearning principles.