It’s been 25 years since the dot-com bubble burst, rocking the nascent world wide web and the global tech economy. Despite being at the point where every boardroom conversation starts with “our AI strategy…”, today, it’s the artificial intelligence market that is having a crisis of confidence. The dissonance between hype and delivery is driving a wedge between investors and innovators, with outsized and unrealistic promises increasingly bringing their return on investment into question.
This uncertainty is reflected in a recent Gartner report showing that generative AI applications in supply chain and procurement have already entered the “trough of disillusionment”. With poor integration, unmet expectations, and almost 70% of supply chain AI pilots still failing to scale, it’s no surprise that “AI-washing” and “agent-washing” have already entered the lexicon.
When Hype Outruns Reality
Whether it’s online businesses, NFTs or even tulip bulbs, investor enthusiasm can rapidly transition to panic when FOMO (fear of missing out) morphs to FOLE (fear of losing everything). The AI boom isn’t there yet but, currently, companies are still racing to adopt unproven technologies even though their rate of ingestion outstrips their feasible rate of absorption.
The logic is understandable: no executive wants to be left behind. However, an arms race based in quantity over quality risks pragmatic solutions and too often distracts from the value of strategic patience. Most organizations are still experimenting in isolation, often duplicating effort across business units in hope of finding the secret sauce that will give them an edge over the competition. The cumulative result is that fewer than one in five supply chain organizations have successfully deployed AI at scale, and only 23% even have a formal AI strategy (Gartner).
So, what’s the root cause of this instability?
Unsurprisingly, it has a lot to do with the quality of the technological foundations: too many companies are trying to buy progress instead of building it. They stack new digital tools precariously atop legacy systems and call it “transformation”, when it’s little more than a patchwork of “franken-systems”. These disconnected pilots and duplicated data models lead to teams that spend more time reconciling dashboards than making actual decisions, with every hour spent chasing hype being an hour not spent tending to those foundations. As one supply chain executive recently summed-up, “we’ve been running ten AI pilots at once, but none of them speak to each other.”
This belief – that tools alone can deliver transformation – highlights the real cost of AI: strategic drift. By focusing on the shiny new AI tool, companies are losing sight of what make real intelligence possible: clean data, aligned processes, and a shared definition of value.
Stop Buying AI and Start Fixing the Value Chain
Most supply chains still move as slowly as the goods they manage, that is, step by step, handoff by handoff, with information passed along like a relay baton, often between teams who can’t even see one another. The result is lag, blind spots, and decisions made too late to matter. Automation can never compensate for this sort of fragmentation.
The goal isn’t AI for its own sake and that’s exactly why it’s time to mature the conversation. If, for instance, your planning still lives in a file called “V7_final_FINAL.xlsx”, you don’t have an AI problem, you have a value chain problem.
A truly connected value chain – one that moves at digital speed, where information flows faster and more freely than materials – should be enabled by AI, not hamstrung by it. When data travels ahead of the goods, every node in the chain can see the same reality, anticipate change, and respond together.
But making that journey a reality begins with a few fundamental truths about where AI truly fits in the value chain.
Four Truths Everyone Should Know Before Buying AI
Truth 1: Treat AI as a system, not a silver bullet
AI isn’t one thing: it’s a collection of complementary capabilities, each designed for a specific purpose. Indeed, “artificial intelligence” can be used to describe a wide array of loosely-related technologies and processes – including machine learning, algorithmic optimization, digital twins, and intelligent agents – and every type can add value. In the realm of the supply chain, however, no single AI tool will deliver true impact in isolation.
Leaders like Sanofi and Straumann show what happens when AI operates as part of a broader system. Sanofi connects predictive and prescriptive AI through a single analytics and decision-making layer, giving 23,000 employeesend-to-end visibility across 540 product families. Straumann combines forecasting and optimization to create one global planning model, improving baseline accuracy by 15% and aligning decisions across brands.
The lesson is clear: AI is most powerful when it functions as part of an integrated network, one that connects data, models, and people into a continuous flow of intelligence.
Truth 2: Make data your competitive advantage
Operations generate data. In the analog world, when many organizations and supply chains were last restructured, information was unwieldy and hard to manage. As such it was largely constrained within the ‘need-to-know’ hierarchies of siloed departments.
However, as operations have shifted to a digital-first model, the amount of data has expanded prodigiously, but so too have the technologies that handle it. Data wants to spread, and more integrated enterprise systems have increasingly forced it to spread horizontally, breaking the barriers between those old silos.
In today’s AI economy, data is no longer a byproduct of operations; it is the product. That’s why Gartner’s concept of Data-Centric AI reframes success around improving the quality, accuracy, and context of data instead of endlessly tuning models. To quote the Harvard Business Review, “good AI comes from good data.”
Johnson & Johnson’s Data Acceleration Program proved how powerful this statement can be. By synchronizing master data with real-world measurements, the company improved planning efficiency and reduced lead times.
The takeaway? Stop treating data like an inconvenience and start leveraging its operational value.
Truth 3: Design the architecture before you buy the algorithm
Scaling AI is not just about technology, it’s about structure. The best-performing organizations balance a centralized model core with decentralized interfaces, allowing each team to work autonomously while sharing one version of the truth.
This creates what we call a distributed intelligence network: a single backbone that feeds role-specific tools, dashboards, and AI copilots for planners, schedulers, and finance teams. When people interact with AI in their own language and context, adoption improves, and so does trust.
Froneri, the ice cream producer behind Dreyer’s and Häagen-Dazs, applied this principle by connecting data from factories, warehouses, and logistics into a single optimization engine. The result was an architecture that allowed local teams to make daily deployment decisions in real time while staying aligned on global objectives. The outcome was lower logistics costs, faster execution, and higher service reliability, all driven by connected design rather than another new algorithm.
In short, you should start with a data strategy that maps the key data sources and sinks across your network and allows you to plan a joined-up approach, instead of investing heavily in short-term, local fixes.
Truth 4: You don’t scale technology; you scale trust
As head of Google Brain, Andrew Ng, once said, “if you can’t scale it, it’s just a prototype.” The hardest part of scaling AI is rarely technical; it’s cultural.
Bluecrux research highlights four levers of AI scale: power, expertise, automation, and people. The first three can be bought but that last must be built. Real adoption depends on explainability, feedback loops, and governance to create confidence in the tool.
Sanofi provides a clear example: its AI-driven digital twin succeeded because it was transparent and traceable. Every planner could see how recommendations were made and could adjust them with context. That visibility builtconfidence, which in turn made adoption self-sustaining.
So, while technology can automate a choice, people still have to trust it before they will act on its recommendations. That trust is what turns a pilot into true transformation.
Proof That It Can Work
Real progress happens when AI is embedded within a healthy value chain. Companies such as Sanofi, Straumann, Johnson & Johnson, and Froneri show that success does not come from algorithms alone but from disciplined integration of data, process, and people. Each began by building a common data backbone, aligning decision flows, and ensuring transparency across teams. Only then did AI amplify their results by improving forecast accuracy, speeding up decisions, and creating visibility across complex global networks.
These examples demonstrate that, when organizations fix their foundations first, AI becomes more than just a tool. It becomes a catalyst for collaboration, agility, and lasting value creation.
The Shift That Separates Leaders from Followers
Ultimately, the responsibility rests with leaders to fix the technological foundations before layering in intelligence. This means green-lighting programs to build a living model of how work actually flows, that harmonize the data, and that define what value truly means for their organization.
Only then should AI be used to accelerate progress, by following a defined strategy and by supplementing the work of the people on the shop floor. Because the future of intelligence may be artificial but the future of supply chains isn’t. It’s augmented, powered by human judgment, connected data, and insights that makes people smarter, not obsolete.
So, the message is simple: AI won’t save a broken supply chain – but it can be the key to transforming that supply chain into a sustainable and scalable value chain.
Stop buying AI. Start building value.
Scott Barnard is the Managing Director of North America at Bluecrux, with full P&L responsibility for driving the region’s strategy and growth across supply chain and operations consulting and the Binocs and Axon digital products. He has 20+ years of leadership in SaaS, advisory services, and regulated supply chain and manufacturing operations within the medical device sector.
