Industry
Retail
Omnichannel CX and Intelligent Innovation – The Future of Retail
HIGHLIGHTS
- The use of AI in supply chains is still in early stages, and the complexity of the processes and the dynamic nature of sub-functions offer a perfect space to consider the induction of hybrid Intelligence.
- The hybrid intelligence maturity model offers a strategic lens to assess and advance the AI capabilities for retailers navigating the evolving supply chain transformation space.
- Embracing hybrid intelligence can help retailers unlock new potential of agility, precision, and resilience across their supply chain operations.
On this page
On this pageinpage
Overview
Supply chains are a complex and significant function especially in retail.
Ensuring the availability of the right products at the right time and at the right place requires complex processes that need to be constantly tuned. Any failure to do so results in in-store shortages as has been observed during calamities, both natural and man-made.
Even though it is generally possible to tune designed processes to near perfection, external and unforeseen events like pandemics (COVID-19), and trade route disruptions (Suez Canal blockage) have derailed the smooth functioning of supply chains. Over the years, organisations have used in-house human-centric expertise to recover from such disruption based on past experiences and learnings. However, with recent advancements in artificial intelligence (AI), retailers can use hybrid intelligence (HI)–human and artificial intelligence working in synergy–to overcome these challenges. Hybrid intelligence can help supply chain teams in retail overcome, or at least minimize, the impact of trade disruptions.
Unified intelligence
While the use of AI in supply chains is still in early stages, the complexity of the processes and the dynamic nature of the sub-functions offer a perfect space to consider the induction of hybrid Intelligence.
Before we dive into the application of hybrid intelligence in supply chain, here’s a quick look at its maturity model. We have identified five stages, each with unique characteristics, that can help improve the supply chain process with AI.
Figure 1
Hybrid intelligence maturity model
Figure 1
- Stage 1 - Assisted: AI automates repetitive tasks, with humans in full control, using AI as a tool.
- Stage 2 - Augmented: AI provides insights and suggestions, playing an advisory role to enhance human decision-making.
- Stage 3 - Collaborative: Humans and AI work together, learning from each other to significantly boost productivity. AI becomes a co-creator.
- Stage 4 - Adaptive: AI adapts to human preferences and context over time. Humans become supervisors, and AI takes on the heavy lifting with context-aware systems.
- Stage 5 - Autonomous: AI operates independently with minimal human intervention. Humans oversee the ethical aspects while AI powers a self-operating system.
The hybrid intelligence maturity model offers a strategic lens to assess and advance the AI capabilities for retailers looking to transform supply chains.
The AI value wheel demonstrates the use-cases in these five stages of hybrid intelligence.
Figure 2
Illustrative AI value wheel with segregated use cases
Figure 2
Action on the five-step framework: Managers can begin with automating routine tasks like workforce allocation and dock assignments using assisted intelligence to boost efficiency. In the augmented stage, AI can help enhance decision-making in areas such as demand forecasting, improving storage utilisation, and eventually move to collaboration and more complex tasks such as co-creating solutions for various challenges ranging from optimizing slotting to managing supply chain transparency. In the advanced stages, we foresee systems that are adaptive, autonomous and context-aware, dynamically adjusting pricing, inventory, and delivery routes in real time all the way to self-operating supply chains.
Key considerations
For optimal results, retailers must evaluate the processes and ensure effective implementation of AI for supply chain.
Here are a few considerations to achieve success:
- Value discovery: To begin with, a structured value discovery process can help uncover operational inefficiencies and align AI interventions that augment personas and drive business goals such as improved sales; reduced selling, general and administrative expenses; and enhanced productivity.
- Identify: Ensure a holistic understanding of the existing supply chain, including the key performance indicators (KPIs) for an impactful implementation of the approach. For example, a US-based specialty retailer asserted their workforce was optimised, but a closer look at the data revealed the workforce was peaking on certain days, leading to overtime. This could not have been verified without taking a closer look at the processes to understand and validate the data and reports associated with key KPIs before offering solutions.
- Validate: Augment the understanding and analysis of the data with site visits. This will ensure first-hand insights and help unlock additional information that the datasets may not provide. For example, the distribution centre of a large US-based retailer had blocks of idle inventory due to disorganised store returns. As cartons weren't organised as per category and duct-tape held items together in pallets, processing returns was a challenge. Following a visit to the centre we could offer to resolve this challenge by using vision AI to help operators quickly process return pallets.
- Map: Stakeholder interviews, data analysis, and on-site visits to distribution centres and stores can help develop a deeper understanding of the existing challenges across domains like transportation, inventory management, and distribution centre operations. AI interventions can be evaluated and augmented with personas to counter challenges and drive business goals once the processes are mapped and pain points have been identified. The focus of this intervention is to effectively augment the persona with additional intelligence so humans can focus on the core and more value-added tasks.
AI value quadrant
The AI value quadrant is a powerful and intuitive tool for prioritising and allocating resources.
The horizontal axis of the quadrant charts feasibility to measure the relative ease of execution, while the vertical axis–business value--qualifies the potential impact of a business value. The quadrant also has two additional dimensions--time to market and time to value, which help project viability evaluation.
Figure 3
AI value quadrant matrix
Figure 3
Here's a three-step method to ensure better actionable insights using the quadrant:
- Visualise each use case with the AI value wheel.
- Assess each use case based on its potential business value, feasibility (including data availability), and time to market, time to value.
- Plot these use cases in an AI value quadrant matrix on the four parameters ie, business value, feasibility, time to market and time to value.
The top right corner represents the low-hanging fruit, the areas that can quickly adopt technological inventions and are ideal for a quick pilot project. This exercise can help retailers create a phased road map.
The hybrid intelligence-first supply chain approach, powered by AI value quadrant matrix and AI value wheel, is a strategic framework designed to transform traditional retail supply chains into intelligent, adaptive ecosystems. At its core, this methodology focuses on identifying, evaluating, and prioritising AI use cases that deliver measurable business value.
Ambiguity to action
While the approach is conceptually straightforward, its application involves significant challenges.
Some of these include:
- Foundational data challenges: Multifaceted data challenges, including data quality, have to be overcome to build reliable AI systems. AI models demand accurate, complete, consistent, and timely data for training. Otherwise, they risk producing biased or flawed output. In addition, data quantity and availability are crucial. Scarcity in niche domains or data silos within organisations can starve AI models of the necessary training data, impacting performance. Recognising and mitigating bias in training data and algorithms is essential for ethical and equitable AI deployment. Finally, data integration and regulatory compliance are vital for secure and seamless AI implementation.
- Understanding and incorporating user needs: Human-centred AI requires a deep understanding of user needs, preferences, and contexts. It should move beyond functional utility to craft experiences that resonate emotionally and address the nuanced realities of human interaction with process and technology. However, eliciting and capturing these needs effectively, especially across diverse user groups, can be challenging. The multi-skilled assessment group should focus on understanding user activities—manual processing of advance shipping notice (ASN), for example—to identify bottlenecks and ascertain if an AI intervention can help the user. .
- Balancing automation with human control: Finding the right degree of automation while preserving meaningful human oversight and control is essential. Designing systems that empower humans to intervene, clarify decisions, and ultimately remain in control can be technically complex and will require innovative design solutions. Using the AI value wheel can help understand the level of automation needed.
- Integration with existing systems: Integrating new AI solutions with an organisation’s existing legacy tech stack can be a major challenge. Large enterprises often still use older systems, giving rise to technical and logistical hurdles. Retailers can start by identifying key applications and evaluating pilot options to find a solution that works with their current integrations.
- Transparency and explainability: Users need to understand how AI systems arrive at their decisions to build trust and accountability. However, the black box nature of some advanced AI algorithms is a hinderance, particularly for end-users. This can be countered by understanding the precise bits of explanations in a use case and choosing the right models or algorithms.
- Lack of knowledge and skills: Many businesses face a knowledge gap regarding AI implementation, including where to start, selecting the right tools, and effectively integrating AI into the existing decision-making processes. Seeking the right partners can help contextualise AI for organisations.
Prioritise and go
The journey toward a human-centred AI supply chain is as much about adopting technology as it is about reimagining the way people and machines collaborate to solve complex challenges. Embracing hybrid intelligence can help retailers unlock agility, precision, and resilience across their supply chain operations. As organisations move through the maturity stages of assisted, augmented, collaborative, adaptive, and autonomous intelligence, they gain the ability to shift from reactive problem-solving to proactive orchestration. Going forward, this can help organisations prioritise what matters most, accelerate what works, and build an intelligent, human-centred, scalable, and future-ready supply chain.