More than 75 percent of industrial AI pilot projects never reach large-scale deployment, and over 80 percent of manufacturers struggle to extend AI beyond isolated use cases, according to a report by Everest Group and Capgemini Engineering titled “Operationalizing Adaptive AI Across the Manufacturing Shopfloor,” relayed by Challenge magazine. For Morocco, where industry 4.0 ambitions must contend with a large existing industrial base, the findings serve both as a warning and a roadmap for moving from experimentation to real industrial transformation.
The report’s core finding is that the main obstacle isn’t a lack of capable technology or algorithms, but the inability of industrial architectures to move data and decisions between systems. In many factories, IT systems, including ERP and PLM platforms, and OT environments, including MES, SCADA, controllers and production equipment, continue to operate separately, leaving data poorly connected or arriving too late for automated decision-making.
Nicolas Rousseau, head of engineering technologies at Capgemini Engineering, said digital ambition isn’t the constraint in manufacturing today; execution is. The report identifies four main obstacles: rigid processes, legacy technology, fragmented data and a shortage of cross-functional skills, challenges especially relevant to Moroccan plants modernized gradually through added machinery and software without a full architectural rethink.
Rather than starting with algorithms, the report recommends manufacturers first build an architecture capable of connecting equipment and structuring data, through tools such as a Unified Namespace, the MQTT protocol for local connectivity and, in some cases, streaming platforms like Kafka. This approach can be rolled out progressively, starting with one production line before expanding, reversing the sequence often seen in AI projects by building infrastructure first and applications second.
Hybrid edge-cloud architecture forms another central element, processing data close to equipment to reduce latency and dependence on constant cloud connectivity, an approach the report says suits Moroccan manufacturers working with ageing infrastructure and technical debt including outdated data formats and proprietary protocols.
The report warns against rushing to fully automate operations: nearly half of manufacturers still struggle to integrate IT and OT systems, and deploying autonomous agents on ageing infrastructure risks adding complexity rather than reducing it. As operational systems open toward the cloud and AI applications, cybersecurity and compliance become integral to the architecture rather than an afterthought, a concern the report says is heightened for Moroccan manufacturers handling sensitive data for international clients, alongside growing interest in “sovereign AI” offerings from major cloud providers.
The report closes by stressing no single vendor controls the full technology chain, with industrial platforms, cloud providers and AI infrastructure specialists occupying complementary roles in the transformation.
