IBF Editorial Team
AI in demand planning is everywhere right now.
But if you walked the halls at IBF’s AI EDGE Supply Chain Planning & Forecasting Conference in Scottsdale recently, you heard something different. Not hype. Not certainty. Instead, a mix of progress, skepticism, and a growing realization:
We are still figuring this out.
Across organizations such as MICROSOFT, CLARIOS, GILDAN, NIAGARA BOTTLING, UNILEVER, and RISE BAKING COMPANY, there were real examples of advancement. But just as noticeable was the tone in the room. There was optimism, yes, but also hesitation, and in some cases, blunt honesty.
One attendee put it this way: “Some companies are using exponential smoothing and calling it AI, just to make their boss happy.”
That comment may sound cynical, but it reflects an important truth. While AI capabilities are advancing quickly, their application in planning is still uneven, often misunderstood, and at times overstated. What emerged from IBF Scottsdale was not a story of transformation already achieved, but one of an industry in transition, moving forward, but still grounded in reality.
AI Is Starting to Participate in Planning, But the Gap Remains
There is no question that the direction is changing.
As demonstrated in MICROSOFT’s session, we are moving toward a model of “digital labor,” where AI agents can execute workflows, manage exceptions, and support decision-making alongside planners. The vision is compelling: planning systems that continuously sense demand, generate insights, and help coordinate decisions across functions.
But for most organizations, that vision is still ahead of their current state.
Many are still working through foundational challenges: data quality, fragmented systems, and inconsistent processes. The gap between what is possible and what is operational remains significant. AI may be advancing quickly, but planning maturity is not always keeping pace.
Forecasting Is No Longer the End Goal
One of the most important shifts discussed in Scottsdale had less to do with AI itself and more to do with how organizations think about forecasting.
As NIAGARA BOTTLING emphasized, “a forecast without context is just a number.”
This resonates because it highlights a long-standing issue. Organizations have spent years trying to improve forecast accuracy, yet still struggle with stockouts, excess inventory, and slow response times. The reason is becoming clearer: accuracy alone does not drive outcomes.
Forecasts only create value when they inform decisions: what to produce, where to position inventory, and how to respond to changes in demand. Without that connection, even a statistically accurate forecast can fail in practice.
Forecasting, in other words, is no longer the endpoint. It is one input into a broader decision-making system.
There Is a Path Forward, But It’s Not a Shortcut
GILDAN and BOARD INTERNATIONAL introduced a framework that helped bring structure to many of these discussions: Analytical AI, Generative AI, and Agentic AI. Each layer builds on the previous one; analytical models identify patterns, generative tools translate those patterns into business context, and agentic systems begin to act on them.
The insight is not just in the framework, but in the sequence.
Organizations attempting to jump directly into generative AI or autonomous agents without a solid analytical and process foundation often end up with disconnected tools and limited impact. In practice, many companies are earlier in this journey than they assume.
As one participant noted informally, “We’re trying to jump to the future without fixing the present.”
AI Is Amplifying the System, Not Fixing It
This reality became especially clear in the practitioner sessions.
At RISE BAKING COMPANY, AI models showed potential improvements in forecast performance. Yet the organization chose not to move forward with the top-performing solution. The issue was not the model; it was the surrounding process. Without alignment in supply planning and execution, the gains in accuracy would not translate into financial value.
Similarly, CLARIOS emphasized that their improvements were driven first by investments in people, processes, and data, before layering in more advanced AI capabilities.
The takeaway is straightforward, but often overlooked:
AI does not fix planning problems. It exposes them and often amplifies them.
Explainability Is Where AI Is Delivering Value Today
While some AI capabilities remain aspirational, one area is already proving its value: explainability.
Across multiple sessions, planners emphasized the need to answer simple but critical questions: why did the forecast change, what is driving it, and what should be done as a result.
This is where generative AI is gaining traction. By translating model outputs into clear, business-ready narratives, it helps bridge the gap between analytics and decision-making. It improves alignment across stakeholders and increases confidence in planning outputs.
Unlike more ambitious AI initiatives, this capability fits naturally into existing workflows and addresses a long-standing challenge in planning.
Agentic AI Is Emerging, But Selectively
There is growing interest in agent-based planning systems, and early use cases are beginning to take shape.
Organizations are applying AI agents in targeted areas such as data ingestion, signal interpretation, exception management, and scenario analysis. In some cases, these agents are starting to interact with one another, forming the early stages of connected planning systems.
But these implementations are still focused and incremental.
The broader vision of fully autonomous planning remains a longer-term aspiration, and not all practitioners are convinced it should be the goal. As emphasized in IBF’s perspective, the future is not fully automated; it is fully human-informed.
The Role of the Planner Is Expanding
Despite concerns about automation, one point was consistent throughout the conference: AI is not replacing planners.
It is changing how they work.
Manual tasks are being reduced, planning cycles are compressing, and expectations are increasing. Planners are spending less time preparing data and more time interpreting it, connecting insights to decisions, and aligning across functions.
As discussed in sessions on LLM-enabled workflows, the planner is evolving into a decision orchestrator and business advisor, someone responsible not just for the forecast, but for what happens because of it.
We Are in the “Age of Potential”
If there was one takeaway that captured the overall sentiment of the IBF Conference, it is this:
We are in the age of potential.
There are early successes.
There are promising use cases.
There is real momentum.
But there is also:
- limited quantified proof at scale
- uneven levels of maturity across organizations
- and ongoing confusion about what truly qualifies as “AI.”
That combination is not a weakness; it is a stage. It reflects an industry that is learning, experimenting, and gradually building toward something more transformative.
Final Thought: Progress Will Come from Discipline, Not Speed
The message from Scottsdale was not that AI has fully arrived, nor that it will immediately transform planning.
It was something more grounded and more useful.
The organizations making real progress are not chasing AI as a concept. They are focusing on execution: strengthening their data, aligning their processes, embedding AI into real workflows, and building trust over time.
AI will play a central role in the future of demand planning. That direction is clear.
But that future will not be defined by the companies that moved first.
It will be defined by the ones who learned, adapted, and built it right.
AI and planning will continue to be a focus at IBF events. If you missed the AI EDGE Conference, you can explore IBF’s upcoming conference schedule here:
https://ibf.org/forecasting-and-planning-conferences-2026