How Retailers are Leveraging AI and LLM to Improve Operations

2023 was a pivotal year for artificial intelligence software, as key innovations were made and commercialized into the market. Although a relatively new technology, many retail companies have already begun leveraging AI and LLM capabilities throughout their workstreams. This increasing interest in implementing AI solutions is forecasted to continue in 2024, with an estimated 6 in 10 retailers planning to adopt AI solutions over the next 12 months. Organizations should review how these tools can fit within their larger data strategy and foundation to get the most value from their investments. To date, these are some ways we’re seeing retailers leveraging AI and LLM to improve their operations.

Chatbots and Virtual Assistants

With AI-powered chatbots and virtual assistants, companies can provide personalized, quality responses to users’ questions around the clock, improving customers’ shopping experiences. Large models enable chatbots to answer frequently asked questions, provide order tracking information or product recommendations, and help with item returns. These bots can also serve as a novel data source, gathering customers’ question data to pull out insights that the company can leverage to avoid future confusion.

Beyond the above, AI chatbots can be implemented for employee use. They can schedule meetings, take and summarize meeting notes, and assist with code development. They can also empower internal knowledge bases, helping enable sales representatives with easier access to product information.

Implementing AI chatbots can reduce wait times and improve customer experiences, particularly with level 1 help desk interactions. However, brands built on familiarity and high-touch experiences should be cautious of too much automation, as it may feel impersonal to longtime customers.

Looking ahead, the integration of chatbots into existing software platforms will become more prevalent, which will help improve adoption and ease training, change management, and the overall transition of working with a new system. However, new challenges may arise for retailers as they strive to maintain and improve these bots throughout their life cycle.

Software Examples: Intercom and Kustomer

Away, a popular high-end luggage company, makes being customer-obsessed and providing a top-tier customer-centric service a core value proposition. However, their CX team took a significant amount of time to locate customer information and pair customers with the correct team members.

Away decided to implement Kustomer, an AI-powered connected service and chatbot platform, that can help to significantly decreases lag time in answering customers’ questions. Kustomer allowed Away to optimize the customer experience team’s response time to inbound messages and used insights from the customer data and search to pair the customer to the correct service agent quicker and with more personalized, informed information.

Click here for more information on this use case.

AI-Assisted Supply Chain Management

AI-assisted supply chain management uses predictive analytics to predict future demand more accurately by analyzing mass amounts of real-time data. These tools enable optimizing inventory levels, streamlining supply chain processes, and reducing the risk of stockouts or overstocking.

Many tools can also find the most efficient and cost-effective way to transport goods, track shipments, and facilitate on-time delivery. AI-assisted supply chain management has been found to reduce labor costs, minimize manual tasks, and improve accuracy and efficiency.

Although AI-assisted supply chain management software can better manage unforeseen risks like COVID and improve future demand planning, they require high-quality data inputs to achieve these results. Retailers should improve their data foundation before investing in AI supply chain management programs. Additionally, implementing supply chain software that has AI built-in can lead to “black-box” predictions, where users will have a more difficult time explaining the models versus if it was developed in-house by the company’s own data teams.

Software Examples: Blue Yonder and O9 Solutions

A retail and wholesale business, which offers an assortment of merchandise and services, was using antiquated and disconnected processes and systems that required extensive manual effort and prevented them from optimally serving customers. Additionally, they lacked tailored assortments based on consumer demographics and regional variances, which created misalignment with space allocation and product offerings. They also did not leverage enterprise analytic insights to optimize space allocation and assortment development processes.

To address these challenges, the company implemented o9’s supply chain management software. With o9, the company finally had a single platform that integrated critical business decisions throughout the assortment planning process. They also could plan granular assortment differences that consider regional nuances, such as store size profile and local preferences. Additionally, the company was also able to leverage advanced analytics and AI to develop volume groupings and determine unconstrained demand impacts, along with other key insights. O9 increased revenue and gross margin, enhanced process efficiency, and stakeholder collaboration.

Click here for more information on this use case.

AI-Enabled Data Visualization and Analytics

With AI-enabled data visualization and analytics, companies can quickly analyze massive amounts of data and create complex visualizations and interactive UI dashboards. These visualizations and dashboards save time, minimizing labor on tedious tasks and allowing for consistent, quick decision-making. It is also found to reduce decision-making bias and increase prediction accuracy.

AI innovation can help accelerate the process and improve outcomes of analytics, democratizing data, and self-service functionality. This would allow the teams that understand their domain best to develop dashboards and analyze performance without requiring in-depth support from the data team. However, retailers should be cautious about seeing this as a magic bullet, as AI-developed dashboards may not follow standard definitions for key metrics and may be challenging to comprehend compared to those developed internally.

Software Examples: Sisense and H20.ai

Macy’s, a large-scale department store, found that it took the company months to make predictive models, constraining Macy’s only to be able to develop long-term or seasonal trends. They wanted to have the ability to shorten the model-building process to days or even hours so that they could find patterns they could use the same week to increase sales and customer satisfaction.

To try and resolve this, Macy’s implemented H20.ai and drastically improved the speed at which they can analyze data and create models. Model creation now only takes hours instead of months. They also found that with H20.ai, they could scale their forecasting operations and have improved accuracy.

Click here for more information on this use case.