September 2026  ·  Keynote

The container moment

GenAI in supply chain planning

KA
Knut Alicke
EyeOn Experience Event, Frankfurt
29 September 2026
Download slides (PDF, 5.6 MB) Browse the slides
Title slide: EyeOn Experience Event, The container moment, GenAI in Supply Chain, with Frankfurt skyline
Event
EyeOn Experience Event DACH
Theme
Digital transformation, the next chapter: AI in supply chain planning
Format
Keynote with live demo, 60 min
Audience
About 65 supply chain and business leaders from 45 companies

We have spent 30 years and a lot of money on advanced planning systems. Yet most companies still plan in Excel, and experienced planners routinely override what the systems recommend. They are often right to do so. The planner was the system.

In Frankfurt I argued that GenAI changes this equation, in the same way the container changed global trade: not by replacing the planner, but by finally connecting the planner's experience with the logic of the systems. And I showed it live.

Six messages from the stage

  1. AI is scaling, and supply chain sees the cost benefits first. The McKinsey State of AI 2026 survey shows 89% of organizations using AI in at least one function and 44% at least scaling. Supply chain management is the function where most respondents report cost reductions from AI.
  2. The planner was the system. Three generations of planning systems (spreadsheets, APS, probabilistic AI/ML) share the same gap. The most important knowledge sits outside: causal depth, episodic memory, organizational behavior and cross-domain reasoning. The 'August in France' example: a planner learns from one burned summer, a statistical model needs more than 100 delayed deliveries.
  3. This is a container moment. For the same 1,000 EUR you bought about 50 million tokens at the end of 2022 and around 50 billion today. Analysing three years of demand for 10,000 SKUs costs about 3 EUR. Traditional AI predicts, generative AI answers, agentic AI acts towards a goal, and needs governance with clear decision rights.
  4. Onboard AI as a digital colleague. A young colleague who is brilliant at data science but lacks business context. Give it onboarding, a mentor, first tasks and feedback. Used as a translation layer, it opens the black box: the planner asks why, and gets an explanation instead of a number.
  5. Vibe coding changes the tech stack. I demonstrated the end-to-end S&OP app I built in 30 hours, with natural-language input, a BOM navigator, scenario planning and a P&L view. It is an MVP, not industrial software. But planning software can provide the core, while experts vibe code the specifics and let GenAI write the specification for IT.
  6. Trust was the consequence. Five AI agents (Sales, Supply Chain, Manufacturing, Procurement, Finance) ran the S&OP cycle with two settings each: locality and trust. A realistic mix delivered an EBITDA of €2.25M against €2.55M in the aligned case, while almost every function's own KPI looked better. Across 121 scenarios: the more local, the worse for the company. Trust is not a parameter you install in an offsite. It is the equilibrium your incentives pay for.

Align your incentives first. Then hold the offsite.

The slides

All 32 slides. Use the arrows or swipe, click a slide to open it in full size, or download the complete deck as PDF.

Title slide: EyeOn Experience Event, The container moment, GenAI in Supply Chain, with Frankfurt skyline
1 / 32 · EyeOn Experience Event: The container moment, GenAI in supply chain.
Grey rhinos and black swans in a container port
2 / 32 · Grey rhinos and black swans: most disruptions are predictable.
Chart of AI adoption 2017 to 2026
3 / 32 · State of AI 2026: organizations are increasingly reaching the scaling phase.
Cost change from AI by function
4 / 32 · Cost benefits from AI show up first in supply chain management.
Revenue change from AI by function
5 / 32 · Revenue gains from AI show up most in marketing and sales and in product development.
AI objectives of high performers versus others
6 / 32 · High performers set growth and innovation objectives, not only efficiency.
Illustration of an experienced planner at a desk with fax machine
7 / 32 · The planner: 30 years tenure, great network, hard to replace during two weeks of PTO.
Chart of planning activity, traditional versus AI-enabled
8 / 32 · Wouldn't it be great to spend most of our time on value-adding tasks?
Infographic: three generations of planning systems, three generations of the same gap
9 / 32 · Three generations of planning systems, three generations of the same gap.
Four layers of semantic intelligence inside a planner's head
10 / 32 · Inside the planner's head: four layers of knowledge no system possesses.
The August in France anomaly chart
11 / 32 · The 'August in France' anomaly: a planner learns once, a model needs 100+ delayed deliveries.
1950s shipping container being loaded in a port
12 / 32 · AI and GenAI will disrupt the way we operate, similar to the container and the internet.
Chart of AI tokens per 1,000 EUR from 2022 to 2026
13 / 32 · Same money, exponentially more AI: tokens per 1,000 EUR.
Three generations of AI: traditional, generative and agentic
14 / 32 · Three generations of AI reshaping how operations actually works.
Onboarding AI as a new digital colleague
15 / 32 · A new digital colleague joining: onboarding, mentor, reverse mentoring, feedback.
GenAI as the natural-language layer between planner and planning systems
16 / 32 · GenAI opens the black box between planner and planning systems.
How I vibe coded an end-to-end planning app in 30 hours
17 / 32 · How I vibe coded an E2E planning app in 30 hours. Followed by the live demo.
Vibe coding workshop with operations teams
18 / 32 · Bring vibe coding to operations: teams solving their own challenges.
Trust was the consequence: title illustration
19 / 32 · Trust was the consequence: what a multi-agent S&OP simulation taught us.
The S&OP meeting we sell versus the one we sit in
20 / 32 · The S&OP meeting we sell, and the one we sit in.
Five functions as agents with locality and trust settings
21 / 32 · Every function as an agent with two settings: locality and trust.
Screenshot of the simulation app
22 / 32 · The simulation running inside the vibe-coded S&OP app.
Comic of a single S&OP cycle driven by local incentives
23 / 32 · Inside a single cycle: no disruption, no bad luck, just local incentives.
Screenshot of agent dialogue in the simulation
24 / 32 · The agents justify their decisions in plain language.
Transcript excerpts of the agents' S&OP dialogue
25 / 32 · Demand, supply and finance review, and the escalation to executive S&OP.
Company scorecard versus each function's KPI
26 / 32 · The controversial result: worse for the company, while local numbers look great.
Heatmap of EBITDA across 121 trust and locality settings
27 / 32 · 121 scenarios: the more local, the worse for the company.
Five findings from the dynamic simulation
28 / 32 · Trust is a consequence, not a parameter.
The cure: seven measures in the order that works
29 / 32 · The cure: seven measures, starting with aligned incentives.
Bringing it home: seven takeaways
30 / 32 · Bringing it home.
Ayrton Senna quote: you cannot overtake fifteen cars when it is sunny, but you can when it is raining
31 / 32 · Ayrton Senna: you cannot overtake fifteen cars in sunny weather, but you can when it is raining.
Thank you slide with further reading and the book From Source to Sold
32 / 32 · Thank you, with further reading.
1 / 32

Bringing it home

  • AI is today's container moment. Treat it as a change in how supply chains operate, not as another tool.
  • Foster curiosity in the organization and in every individual.
  • Impact back, not tech forward. Start from the decision that hurts most.
  • Build strong capabilities on all levels, from the planner to the board.
  • Start with a small pilot and scale fast.
  • Move towards a new tech stack: standard software plus vibe-coded specifics.
  • Treat AI as the new digital colleague.

One practical rule from the Q&A: never put enterprise data into a private AI account. Use enterprise-licensed models, and bring IT in early.

The complete deck
32 slides · PDF · 5.6 MB · English
Download PDF

Further reading

More keynotes

Prewave Day · Oct 2026The container moment: resilient supply chains enabled by Agentic AI
Tangier Logistics Days · Sep 2026AI transformation: the supply chain of the future

Thanks to Pascal van den Boogaard, Ieke le Blanc and the EyeOn team for a full house in Frankfurt and a day of excellent discussions, and to Martin Hendel for sharing his customer story on AI in supply chain planning.

Discussion on LinkedIn. Speaking enquiries: knut@alicke-scm.com.