Manufacturing is undergoing massive digital transformation, and industrial AI is at the center. Unlike previous transformations in media or banking, this one is reshaping physical production. But most manufacturers get it wrong: industrial AI isn’t about better systems—it’s about rethinking how sales and production collaborate. Companies winning with industrial AI break down organizational silos and integrate functions. In this episode of Data Driven Voices, Antti Rantanen, Principal Consultant of Industrial AI at EFESO, explains why some manufacturers will thrive while others disappear.
In this blog, we summarize some of the key takeaways from the episode.
Digital transformation is not a new ERP system
Antti challenges the most common misconception in manufacturing: implementing SAP or an MES system is digital transformation.
It is not. It is replacing software.
Real digital transformation happens when 30% of companies go bankrupt because they cannot operate in the new world. This happened in music (iTunes destroyed the industry). This happened in media (Google and Meta took advertising). And it is happening right now in manufacturing, driven by competitors from China producing better quality products at 30% lower costs with perfect execution.
“Digital transformation is really when the workflows and capabilities change.”
The difference between implementing a system and transforming a business is massive. Most manufacturers are still working the same way, just with different software. Real transformation means people work differently. Processes work differently. The entire operating model changes.
The “lights out factory” concept: what’s really coming
Antti introduces the concept of “lights out manufacturing”, fully automated production that runs without humans on the shop floor.
Chinese competitors like BYD operate this way. They produce better products faster at lower costs. Their competitive advantage is not better technology. It is better systems thinking. Their workflows are optimized by AI and algorithms, not by experience and intuition.
The startling reality: Chinese manufacturing companies now must report in their annual reports how many biological workers, AI agents, and robots they employ. Not total headcount. But the mix of different worker entities. Stock markets and regulators want to know this breakdown. It tells you the scale of transformation underway.
The question for Nordic and Western manufacturers: are you preparing for a workforce that includes AI agents and robots alongside humans? Most are not.
Why sales and production still do not talk to each other
Here is what Antti found in manufacturing after six years working on shop floors: sales and production operate in completely separate worlds.
Sales makes demand projections based on CRM data (when they have access to it at all). Production schedules based on guesses and historical data. The two systems do not integrate. CRM and ERP systems sit in different silos.
One manufacturer Antti worked with had production schedulers who did not have access to actual sales data. They received demand projections six months old. For a factory producing specialty paper where humidity, batch sequencing, and machine availability matter, this is guesswork.
“The shop floor produces what the sales organization sells. They don’t produce things just for the hell of it.”
Yet these two functions barely communicate. The business case is staggering. For a 1 billion euro manufacturer, improving demand prediction alone could be a 60 million euro business case just from reduced inventory, logistics, and waste.
The GE Aviation case study: how to actually do it
The clearest example of manufacturing transformation done right: GE Aviation in 2015.
GE faced a problem: complex airplane engines with thousands of variations, but they were struggling on “on time” delivery while costs skyrocketed. They realized the issue was not production capability. It was coordination between sales and production.
Here is what they built:
Sales reps used iPads to take engine specifications. The system automatically projected production losses based on those specs, giving salespeople a real-time pricing and feasibility tool. Then GE added an algorithm that crawled the internet for buying signals (when airlines ordered specific aircraft models). This told them when engine demand was likely coming.
All of this fed into one integrated system: CRM, ERP, MES all connected. When a purchase order was confirmed, the entire production sequence auto-designed itself based on sales specs, projected demand, and real-time production capacity.
The result: a single algorithm that ran from buying indication all the way to delivered product. Sales and production working as one system, not two silos.
The real blocker is not IT. It is mindset
Emma asks the obvious question: is IT the biggest hurdle?
Antti’s answer: no. Not even close.
The data exists. CRM, ERP, and MES data are usually in decent shape. The real problem is organizational silos and mindset.
At one large industrial toolmaker, demand planning people did not speak to production people. Demand planning only spoke to salespeople. They made projections based on historical data (excluding 2020-2023 because “those years don’t count” due to the pandemic) and emailed them to production. No discussion. No collaboration. No feedback loop.
At another company, manufacturing automation teams designed automation without asking shop floor workers what they actually wanted automated.
The CEO of GE said it directly: the biggest hurdle was that salespeople asked “Why the hell do we need to speak to production?” and production asked “What do we care about salespeople?”
“It really is a mindset thing.”
Systems and data are enablers. Collaboration is the transformation.
What leaders must do right now
For Antti, the path forward for manufacturing leaders is clear:
First: Go see what good looks like. Not a corporate experience center. But actual world-class manufacturing. See how they organize marketing and sales. See how they integrate systems. Understand what is possible.
Second: Be courageous. Transformation means being willing to not optimize for EBITDA for a couple of years. Most of the companies that failed in previous transformations could not make this leap.
Third: Hire people who have implemented this before. Do not try to reinvent the wheel.
Fourth: Study history. Kodak, Blockbuster, EMI Records. The mistakes are identical. It always comes down to speed of decision-making and willingness to invest enough in change.
The window is closing. Antti believes manufacturing companies have two to four years to completely reorganize end-to-end processes. Those who do could become the Apple of their industry. Those who do not will be part of the 30% that disappears.
Leadership tools: AI is not optional
Antti’s final practical advice: if you are not using at least three different AI tools daily, you are behind as a leader.
He uses GPT-5, M365, Claude, and Perplexity. He uses them for decision-making, for understanding complex topics, for scenario planning.
If your organization’s AI strategy is “buying co-pilot licenses,” you have probably already failed. That is how fast the landscape is moving.
Boards should be consulting AI tools the way they consult strategy consultants. Using them for scenario planning, for understanding options, for exploring “what if” questions at speed.
Digital twins and scenario-based planning are no longer complex theoretical tools. They are accessible. Available. Ready to use. A leader and an AI tool can model production scenarios, sales scenarios, competitive scenarios in an afternoon.
Key takeaways for manufacturing leaders
- Digital transformation is not a system: It is fundamentally changing how people and processes work. If workflows do not change, you have not transformed.
- Sales and production must integrate: Not just systems, but mindset. They produce what you sell. They must collaborate.
- CRM-ERP integration is a massive business case: Better demand prediction alone can be worth tens of millions for large manufacturers.
- The blocker is mindset, not technology: Data exists. Systems exist. The issue is organizational silos and willingness to collaborate across functions.
- 30% of manufacturers will not survive: This is not speculation. It is pattern recognition from every major industry transformation.
- The window is 2-4 years: Manufacturers have a narrow window to reorganize end-to-end processes before competitors leave them behind.
- Leadership is the critical capability: Speed of decision-making, willingness to invest, courage to transform. This matters more than technology.
Manufacturing leaders face a stark choice: transform now or disappear. The data, tools, and examples exist. What is missing is urgency and courage.
Inspiration for marketing, sales, and data professionals
Data Driven Voices is a podcast where Avaus together with industry experts, thought leaders, and partners discuss how to harness data, technology, and strategy to drive meaningful change and business results in primarily marketing and sales. The podcast shares actionable insights, success stories, and thought-provoking challenges to help professionals with new perspectives.