Prescriptive AI in the Factory: Accessible for SMEs or Enterprise-Only?

As part of IIoT World Manufacturing & Supply Chain Day 2025, the panel explores what industrial AI really means in manufacturing, where it is delivering value today, and why adoption still remains difficult for many companies, especially small and midsized manufacturers. The discussion covers practical use cases like predictive maintenance, computer vision, and prescriptive recommendations, while also tackling the bigger issues of data readiness, ownership, workforce change, trust, and how AI will reshape manufacturing work in the years ahead.

Top 5 overall takeaways

1. Industrial AI is a different game entirely
   Manufacturing environments introduce variability, risk, and precision requirements that consumer AI never faces. The challenge is not just building models. It is making them reliable enough to operate in real production conditions.

2. The real value is in focused, high-impact use cases
   The biggest returns are coming from targeted applications like predictive maintenance, computer vision, and prescriptive recommendations. Companies winning with AI are solving specific operational problems, not chasing broad transformation narratives.

3. Most failures are organizational, not technical
   AI projects rarely fail because the algorithms do not work. They fail due to unclear ownership, poor data foundations, lack of trust, weak change management, and inability to connect insights to action.

4. Small manufacturers are more capable than they think
   While large enterprises are investing more, smaller manufacturers often have simpler, more usable data environments. That makes it easier to pilot, prove value, and scale without the burden of legacy complexity.

Trust and decision-making (not data) are the bottlenecks
The constraint is not collecting or analyzing data. It is whether organizations trust AI outputs and act on them. The companies that win are the ones that connect insight to execution quickly and consistently.

Jeff’s top 5 contributions

1. Industrial AI requires a higher standard across the board
   Jeff broke down why industrial AI is harder, highlighting four constraints: environmental complexity, near-zero tolerance for error, stricter security requirements, and the challenge of scaling across heterogeneous systems.

2. AI must be embedded into business strategy, not treated as a side initiative
   He emphasized that AI is not another digital program. It is a capability that should shape how the business competes, operates, and makes decisions across every function.

3. Ownership matters more than org structure
   Jeff clarified that whether or not a company needs a Chief AI Officer depends on how central AI is to the business. But what cannot be optional is clear ownership, governance, and accountability.

4. Smaller companies may have a structural advantage
   He highlighted a key insight: large companies lead in deployment, but smaller companies often have better data usability. Less fragmentation means faster progress if they choose to act.

5. AI will change how work is done
   Jeff’s broader point was that AI is not just another tool. It changes who does the work, what gets automated, and how decisions are made. The companies that treat it as a core capability (and educate their people accordingly) will move ahead fastest.

---