Executive Summary

Pharmaceutical supply chains are among the most complex and regulated in the world. The decisions made within these supply chains often involve navigating intricate trade-offs between cost, speed, quality, and many other factors. Historically, product segmentation has been used by pharmaceutical supply chain leaders to manage complexity in trade-off decision making, but with the increasing use of artificial intelligence (AI), segmentation will cease to be necessary.

The team at HASH are working on next-gen AI optimization for supply chains. This article presents three key insights from our research:

1. →

   Planners can (and should) look beyond the traditional time-cost-quality triangle to optimize for other important factors, such as baking in redundancy (improving robustness) and reducing environmental emissions (contributing towards sustainability initiatives).

2. →

   AI is already enabling planners to move beyond fixed segmentation, facilitating dynamic responses to changing market conditions, and enabling continuously optimized point-in-time decision making.

3. →

   Increased granularity of segmentation, again enabled by AI, will going forward allow for even more precise and optimized decisions, unlocking additional value.

Using AI in these ways can not only support and improve performance of human made decisions, but can also pave the way for increasingly autonomous AI decision making and supply chain coordination. Unleashing optimal decision-making for each product, in every situation, will drive more value, and remove more waste. Consequently, leaders can be confident that the best possible choice was made based on the available information.

Contact us to hear more about our thinking and to explore how we can help with managing your supply chains.

Navigating Trade-Off Decisions in Pharmaceutical Supply Chains

Our research into the AI-enablement of safety-critical supply chains has focused largely on pharmaceutical ones so far, which given their complexity, competitiveness, highly-regulated nature, and direct impact on human health, present the perfect challenge environment for rigorously trialling new "safeguarded" AI.

Pharmaceutical companies must often choose between cost-effective logistics and maintaining high service levels. For example, while air freight ensures rapid delivery, it is significantly more expensive and generates much higher carbon emissions than sea freight, which is slower and riskier for sensitive products. Maintaining high inventory levels can safeguard against stockouts, but leads to increased holding costs and potential wastage, particularly for temperature-sensitive biologics. Using reusable shipping containers or green transport options may compromise delivery speed, but align with corporate ESG goals.

Modern supply chain professionals face the significant challenge of daily decision-making, navigating and balancing these numerous influencing factors. For organizations seeking to optimize their supply chains, understanding and strategically managing these trade-offs is crucial for ensuring both commercial success and positive patient outcomes. In many cases, supply chain segmentation is used to provide a rule based method for fast and supposedly optimized supply chain decisions.

Segmentation for Rule-Based Trade-Offs

Often segmentation provides the framework to guide trade-off decisions in a way that aligns supply chain strategies with business goals, customer needs, and product characteristics, seeking to avoid one-size-fits-all strategies. This involves categorizing products into segments, such as New Products or Tail Products, and then establishing the key optimization priorities for each category. This approach allows for distinct supply chain operational models, such as agile models for clinical and launch phases (focus on high speed and service level) and efficient models for generic products (focus on low cost).

In the day-to-day running of a supply chain, knowing the segment of a product allows a supply chain professional to quickly identify where on the cost, quality, and speed triad optimization should occur. This information is used to inform the trade-off decision they would inevitably need to make when weighing the various possible options that could be taken.

Heuristic-driven supply chain decisions, akin to outdated AI expert systems relying on manually written rules, are no longer the most effective approach to optimization of supply chain decision making. Just as AI has moved beyond these systems, supply chain management needs to evolve past heuristic methods. It is not enough to have a static, high-level optimization target to inform trade-off decisions: for example, always focusing on cost reduction for a certain product, regardless of the supply situation at that present moment. Supply chains are dynamic, and so much more value can be gained through precise optimization based on the business needs and the specific scenario being faced.

The HASH Supply Chain Research team highlights how the use of AI allows supply chains to move past these challenges, and proposes a way to gain the value promised and unfulfilled by existing segmentation strategies.

Beyond The Triad

While traditional supply chain optimization uses product categorization and segmentation to help make speed, quality and cost trade-offs, the use of AI enables additional factors to be considered as well, helping strengthen supply chain robustness and resiliency, while minimizing emissions and environmental footprint — all without compromising the existing focuses of optimization.

While minimizing expenses, accelerating delivery, and ensuring high product standards remain critical goals, overlooking other factors important to the business can lead to long-term risks and missed opportunities. The use of AI can enable the addition of these factors, giving clear guidance on what the optimization priorities for each decision need to be.

Taking sustainability as an example, sustainable supply chains reduce carbon emissions, minimize waste, and promote ethical labor practices, aligning operations with increasing regulatory pressures and consumer expectations. By integrating sustainability into decision-making, such as optimizing transportation routes to lower emissions, or selecting partners based on environmental performance, organizations can build more resilient, future-proof supply chains that deliver value beyond short-term financial gains. Unlike humans, AI can understand the interdependencies across all the differing factors that are important to a business, and recommend the right trade-off to make in a specific scenario.

In using this approach, AI not only provides guidance to human decision makers about what to optimize for, it becomes particularly valuable where the AI itself can subsequently recommend or execute a decision optimized for the specified driving forces.

Getting Granular

Another huge strength that AI brings to supply chain management is the level of detailed information that can be gathered and assimilated. Supply chain professionals keep in their brains a vast library of the nuances that inform their decision making, and AI can formalize and unlock this information to improve decision making. With more information being brought together, comes the ability to make decisions at more granular levels.

Thinking about how this applies in the pharmaceutical manufacturing sector, segmentation is frequently conducted at the Stock Keeping Unit (SKU) level. However, this is insufficiently granular to ensure patient safety and supply continuity, as a single SKU often serves multiple distinct markets, each with unique dynamics.

Each market's availability of therapeutic alternatives varies, and optimization based on information at the stock-keeping unit (SKU) level can amplify the patient impact of supply disruption. In addition, other market differences like reimbursement, distribution channel maturity, and demand fluctuations necessitate different approaches. Granular SKU-market optimization increases visibility into local nuances, optimizes decision-making, and improves the organization's ability to serve patients in diverse global markets.

Achieving Dynamic Optimization Of Supply Chains

As external pressures mount and requirements change faster than ever before, in the face of increasing regulations, trade wars, and geopolitical uncertainty, setting optimization priorities through a yearly segmentation refresh is obviously no longer sufficient. At the core of dynamic optimization lies the fundamental concept that each product-and-market combination possesses a dynamic set of priorities across cost, speed, quality, and other factors determined by the business (for example, sustainability). These priorities must adapt in real-time to any shift in prevailing conditions.

Upon any alteration in business conditions, such as demand fluctuations, supply disruptions, or emissions budget constraints, the optimization priorities adjust accordingly, and subsequent decisions are based on a revised optimization recommendation. Triggers to change the priorities can include competitor stock-outs, batch failures, and real-time breaches of carbon budgets. Whether a decision is being made by a supply chain professional or an AI, leadership can be confident that based on the situation and information available, the best decision for the business was made because there was a clear, up to date, and specific optimization goal driving it.

In essence, optimization recommendation supersedes existing static segmentation models with a dynamic and continuously updated representation of business intent. This new approach can allow a company to optimize on a certain output, for example, by improving performance on certain factors to the best degree possible without compromising on others. By explaining in detail trade-offs, this methodology harmonizes human and machine decision-making, expedites root-cause analysis, and transforms consideration involving cost, speed, quality, and sustainability from implicit subjective assessments into transparent, data-driven decisions.

Understanding that "these are the promises of AI-optimized supply chains" is markedly different, though, to being able to realize such benefits. If you’d like to understand what data, technology and systems (including processes and policies) are required to practically enable the dynamic AI-assisted optimization of supply chains, we'd love to chat.

Over the last six months we've interviewed senior leaders and practitioners from dozens of biopharmaceutical manufacturers, developing frameworks and pilot solutions addressing these needs. If you're curious about how AI can support transformation of supply chain processes in your organization, or if you have your own thoughts or experiences to share regarding any of the perspectives we’ve discussed here, please also reach out to us.

Get in touch with us at HASH to:

• →

  Explore this optimization methodology in more depth

• →

  Co-develop or pilot AI based solutions to problems in your organization

• →

  Co-publish research with us and our academic partners at the University of Oxford (or debate, challenge, and refine the future of smarter supply chains with us)