“A time of turbulence is a dangerous time, but its greatest danger is a temptation to deny reality”
— Peter F. Drucker*

No predictions
“I never predict. I just look out the window and see what’s visible—but not yet seen,” Peter Drucker once explained (Lenzner & Johnson 1997). So let’s skip another prediction about AI and look at what is already visible. Generative AI has arrived, and agentic AI is moving from generating answers to taking actions. Davenport and Tiwari (2024) already asked the critical question: Is your company’s data ready for generative AI? The shift to agentic AI makes better data the top priority on the C-level agenda.

Start with what will not change: Responsibility for results
When almost everything is changing, a powerful strategy is to start with what will not change. Jeff Bezos makes precisely this point: strategy can be built around things that are “stable in time” (Amazon Staff 2019). For management, one thing will not change: someone remains accountable for results and liable for failures. That matters because generative AI is probabilistic. It produces what is most likely, and without sufficient relevant data, it can confidently make things up. When agents act on those outputs, uncertainty becomes operational risk, including compliance breaches and financial loss. The response is familiar. For decades, we have reduced uncertainty with BETTER DATA … more relevant information content and quality. Three everyday examples show where better data makes all the difference:
     •  Success in surgery: The problem is knowing precisely what is inside the patient and where. Moving from flat X-ray images to computed tomography (CT) combines measurements from many angles into detailed cross-sectional images, providing far greater diagnostic visibility before anyone makes the cut (FDA 2023)
     •  Quality in manufacturing: Defects mean scrap, rework, warranty cost, or recalls. Six Sigma uses better measurement data to identify variation, isolate root causes, and control processes, targeting only 3.4 defects per million opportunities, about 99.99966% without a defect (Schroeder et al. 2008)
     •  Safety in autonomous driving: The proverbial “last 1%” is hardest because the rare events that matter most are almost absent from normal driving data. In safety-critical perception, there can be more than 1,000,000 common examples for every rare one. Better data on these edge cases is therefore essential to closing the gap (Liu & Feng 2024)

Analytics versus data: “It’s the data, stupid!”
After 35 years helping Fortune Global 100 companies use data analytics to create business advantage, DATA is becoming more important than ANALYTICS. We have worked through successive waves in which analytics innovation took center stage and data itself often barely warranted mention: multivariate statistics, machine learning, early neural networks, genetic algorithms, complex adaptive systems, swarm- and agent-based simulation, and now agentic AI. Across those waves, one constant has been the steadily rising importance of data—from relational databases and data warehouses, through Big Data, to the data lakes described by Porter and Heppelmann (2015), to data mesh (Dehghani 2022), and increasingly dependent on shared semantics and ontologies.

Data factories for data products
For this next phase, we leaned early on proven management practices for scaling and automation: Procter & Gamble’s product-management logic became data productization (Low & Fullerton 1994; Crosby & Schlueter Langdon 2019); Henry Ford’s industrialization logic became data factories for producing reusable, analytics-ready data at scale (Schlueter Langdon & Sikora 2020). This data harmonization work at scale is rarely glamorous. Kai-Fu Lee, former president of Google China and author of the bestseller AI Superpowers (Lee 2018), recently attributed much of China’s AI catch-up to precisely this kind of “dirty work”: nearly comparable results with only 1–3% of the GPU power of leading U.S. rivals, enabled by far more engineering effort on what is “not glorious, not interesting, but gets the job done” (Husain 2026). Our own large-scale projects point to the same conclusion: Do not start with the latest analytics and AI. Start with the business result and the data required to deliver it. Three multi-partner programs provide proof at industrial scale:
     •  RealLab Hamburg: cross-provider data chains demonstrated up to 30% faster travel speeds -> Read our Auto 5: Mobility super-app disruption – Case study
     •  Catena-X: Governed data sharing across companies and tiers is enabling next-level automotive supply-chain performance, from end-to-end quality to demand & capacity management and cumulative product carbon footprints
-> Check Catena-X on Tractus-X – Compendium
     •  RoX: Trusted, reusable machine and vision data is enabling Physical AI and robotics
-> Use Physical AI in action: RoX

References
Amazon Staff. 2019. Live from Las Vegas, it’s re:MARS 2019. About Amazon (June 2), link
Crosby, L., and C. Schlueter Langdon. 2019. Data as a Product to be Managed. Marketing News, American Marketing Association (October 10th), link
Davenport, T., and P. Tiwari. 2024. Is Your Company’s Data Ready for Generative AI? Harvard Business Review (March): 44-53, link
Dehghani, Z. 2022. Data mesh: Delivering data-driven value at scale. O’Reilly Media: Sebastopol, CA
Husain, M. 2026. Kai-Fu Lee: China Will Win the AI Race for Reach. Bloomberg Weekend Interview (September 4), Bloomberg, link
FDA. 2023. Computed Tomography (CT). Medical X-ray Imaging, U.S. Food and Drug Administration (May 1), link
Lee, K.-F. 2018. AI Superpowers: China, Silicon Valley, and the New World Order. Houghton Mifflin Harcourt: Boston, MA
Lenzner, R., and S. Johnson. 1997. Seeing Things as They Really Are. Forbes 159(5) (March 10): 122-128, link
Liu, H. X., and S. Feng. 2024. Curse of Rarity for Autonomous Vehicles. Nature Communications 15: 4808, link
Low, G., and R. Fullerton. 1994. Brands, Brand Management, and the Brand Manager System: A Critical-Historical Evaluation. Journal of Marketing Research 31(2): 173-190, link
Porter, M., and J. Heppelmann. 2015. How Smart, Connected Products Are Transforming Companies. Harvard Business Review 93(10) (October): 96-114, link
Schlueter Langdon, C. 2024. Auto 6: Sustainability with Catena-X. Case study (2024-12-31), Drucker Customer Lab, Drucker School of Management, Claremont Graduate University, Claremont, CA, link
Schlueter Langdon, C. 2023. Auto 5: Mobility Super-App Disruption. Case study, Drucker Customer Lab, Drucker School of Management, Claremont Graduate University, Claremont, CA, link
Schlueter Langdon, C. 2025. Agents (Are Back). Research note, Drucker Customer Lab, Drucker School of Management, Claremont Graduate University, Claremont, CA, link
Schlueter Langdon, C., and R. Sikora. 2020. Creating a Data Factory for Data Products. In K. R. Lang, J. Xu, B. Zhu, X. Liu, M. J. Shaw, H. Zhang, and M. Fan (eds.), Smart Business: Technology and Data Enabled Innovative Business Models and Practices. Springer Nature, Switzerland: 43-55, link
Schroeder, R., K. Linderman, C. Liedtke, and A. Choo. 2008. Six Sigma: Definition and Underlying Theory. Journal of Operations Management 26(4): 536-554, link
U.S. Food and Drug Administration (FDA). n.d. Computed Tomography (CT). U.S. Food and Drug Administration, link

* Drucker, P. 1980. Managing in turbulent times. Harper & Row: New York, NY