Implementing AI in the Supply Chain: Delivering a Profitable Brand and Customer Experience for Retailers

Artificial intelligence (AI), particularly AI-enabled decision-making, streamlines the integration of market and consumer insights into the retail supply chain, significantly minimizing the need for manual labor. From package tracking to demand sensing to product recommendations for out-of-stocks, AI brings real-time supply chain data to the surface in a way that benefits shoppers.

However, most retailers are not yet using AI, or even manually analyzing market and consumer insights when it comes to supply chain, says Ed Robbins, global lead of supply chain management at Publicis Sapient.

Despite the challenges, AI is essential for retailers who want to future-proof their business.

While many organizations are now familiar with or have heard of AI and machine learning (ML), their approach to implementation often falls short, for four reasons.

1.  The implementation remains confined to specific solutions, lacking the     capacity to drive end-to-end value in the supply chain.
2. The complexity of AI-related deployments hinders their ability to keep   pace with today's business needs and challenges.
3. Potential outcomes are often rejected due to mistrust in AI-generated numbers and recommendations, which are limited by the nature of the technology’s siloed and overly complex deployment.
4. A lack of skilled talent slows or halts AI adoption and usage. Oftentimes, the average employee isn’t equipped to use AI in their day-to-day workflow, and there’s a lack of training and guidance around implementing it to its fullest potential.

For example, many organizations have implemented AI to predict and reduce order-to-delivery lead times and have resolution capabilities to prevent hold-ups. However, these benefits don’t transfer into inventory strategies, network capacity requirements or potential improvements in end-to-end planning and performance.

Successful AI integration cannot be purely technology-led, but instead must be joined up into a common approach around the customer, the operations and the employee experience. This is where AI and decision intelligence comes in.

Decision intelligence is the key to operationalizing and scaling AI and ML, and to extracting maximum value. By embedding AI and ML into business decisions to deliver outcomes rather than just enforcing process compliance, a platform can be created that supports that supports a holistic understanding of use cases. Businesses can improve both employee experience and operational performance aimed at a customer, creating a market-first seamless experience across multiple channels.

Where can retailers get started?

Embedded workflows, rules engines and optimization capabilities:

These tools can provide decision intelligence and automation capabilities to act and affect change at levels of detail and accuracy not previously possible. 

For example, when companies manually direct supply chain planning activities, they are limited by their ability to consistently deal with and manage large volumes of granular levels of detail.

This results in supply chain planners making attempts to aggregate and simplify things to deal with the high levels of complexity which in turn results in a sub-optimal answer when compared to a machine.

Artificial intelligence has the ability to manage large volumes of granular detail while also providing recommendations based on the embedded rules, outputs and pattern recognition a machine can deal with.

Conversational artificial intelligence can go a long way in analyzing supply chain data to help employees make quick decisions. Acting as a virtual personal assistant, a proprietary conversational artificial intelligence tool can provide answers to internal questions asked.

Conversational artificial intelligence can go a long way in analyzing supply chain data to help employees make quick decisions. Acting as a virtual personal assistant, a proprietary conversational artificial intelligence tool can provide answers to internal questions asked.

Predictive analytics capabilities:

These tools can provide decision intelligence to anticipate and respond to market and consumer insights, predict future trends and provide recommendations to keep businesses on track.

For example, predictive analytics enables more successful digital media campaigns. An end-to-end decision intelligence platform and analytics capability with optimization rules can better understand where the demand is located, allowing retailers to flex fulfillment methods and, potentially, manufacturing locations based on any spare capacity and supply capabilities.

This prevents out-of-stocks of promotional items, and also reduces the risk of obsolescence or late delivery of the item after the promotional item is no longer in demand.

Predictive analytics can also dynamically assess and predict net revenue management options, and make order quantity recommendations for improved profitable promise-to-delivery commitment and margin protection.

Demand sensing:

This involves leveraging diverse public and private data sources (including social media, events, advertising conversion rates, weather, etc.) to analyze historical sales data and better predict future demand states.

Through comprehensive data feeds, AI and ML can be deployed to analyze the correlations between various factors to derive probabilistic statements about future demand and improve demand accuracy. AI allows analysts to run what-if scenarios to generate probabilistic views of demand states.

Now retailers not only stand to capture a short-term opportunity by putting the right product in front of the customer at the right time, with the right price that the customer desires but also to delight them with something that they themselves didn't even realize they wanted. That is, retailers can move away from traditional forecasting based on purely historical data; thinking about who the customer will be, not who the customer was.

Once this is in place, businesses can synchronize core functions around a single view of the customer to better enhance responsiveness to future demand states, and begin collaboratively solving problems that the customer truly cares about. 

Use case example: Using AI to improve demand forecasting and promotions planning

Retailers have the opportunity to bring together market and customer insight data via online advertisement clicks, visits to social media sites and behaviors.

Typically, promotions are static and tied to historical data rather than market trends. Sales teams are unable to keep up with competition in the fast-changing world due to inaccurate predictive data models, leading to ineffective trade investments and failed customer negotiations. To optimize promotions to negotiate with customers more effectively and react to critical events, sales teams need support to make proactive decisions.

By embedding AI and ML into supply chain business decisions, retailers can:

1.  Anticipate short-term demand by combining real-time customer insight data with other market data to feed into demand sensing capabilities.
2.  Locate sourcing options and channels, and feed into the demand and forecasting process requirements for promotions.
3. Compare promotional pricing intelligence as part of the demand shaping scenario modeling and optimize based on desired outcomes.
4. Integrate this new intelligence with end-to-end planning and production activities to deliver a step change in availability of product for promotions.
5. Reallocate and relocate existing stock and production plans more quickly than before.
6. Dynamically adjust safety stocks to reflect the new situation, constraints and lead times.
7. Evaluate better routing optimization decision opportunities to help guarantee delivery precision.

The benefits of combining data, AI and automation in the intelligent supply chain

Publicis Sapient can help retailers develop a tailored strategy for implementing AI into supply chain decision intelligence through our digital brain solution.