Manufacturers back MCP-linked AI for productivity gains

Propel has published research on how manufacturers expect MCP-connected AI to improve operations. The survey covered 400 senior manufacturing leaders across several industries.

The findings suggest manufacturers have moved beyond general interest in artificial intelligence and are now linking model context protocol, or MCP, to specific operational goals. The most commonly cited expected benefits were improved employee productivity at 35%, faster decision-making at 34%, and faster product development at 31%.

Talker Research conducted the survey on Propel's behalf among senior professionals in high tech and electronics, industrial equipment, and medical devices. The results also showed clear differences between departments in what they want connected AI systems to do.

Role priorities

Product management teams placed particular value on using AI to create a consolidated view across product and operational data. That view includes requirements, bills of materials, open changes, ownership, and quality status, reflecting the need to connect technical and commercial information in one place.

The broader dataset also showed where respondents think connected AI could have the biggest effect across the product lifecycle. Product design and engineering ranked first, while quality management came second, underlining the importance of linking quality information into wider workflows.

Among product management respondents, 69% said it was extremely or somewhat valuable for an AI agent to pull together information from PLM, CRM, ERP, and market data. The goal is to surface product, customer, and competitive information in a single view for teams making product decisions.

IT leaders, by contrast, focused on the controls needed to expand AI use across an organisation. The survey found 95% rated the ability to securely connect approved AI tools to PLM and QMS systems, with access controls and audit logging, as extremely or somewhat valuable.

Another 86% said a single standardised MCP interface would be valuable as an alternative to multiple point-to-point integrations. A further 84% highlighted the value of governing what AI can access and act on from one place.

For technology teams, the issue is not simply whether AI tools can connect to manufacturing systems, but whether those links can be managed consistently and in a controlled way. In businesses where different departments are requesting AI-based tools at the same time, that governance question becomes central to wider deployment.

Launch materials

Marketing leaders identified a different use case tied to a longstanding operational problem: late changes to product specifications during the development cycle. In those situations, launch teams often need to revise materials quickly using information held across engineering, quality, and product systems.

The survey found 85% of marketing respondents rated AI assistance that brings those details together for launch materials as extremely or somewhat valuable. That suggests commercial teams see practical value in reducing the manual work involved in reconciling technical information before a launch.

The research points to a broader trend in manufacturing technology spending, with interest shifting from generic AI experimentation toward more defined use cases anchored in existing business systems. In this case, the common thread across functions is access to connected product data rather than the AI model alone.

Manufacturers have spent years managing data across separate platforms such as product lifecycle management, quality management, customer relationship management, and enterprise resource planning systems. The survey results suggest many leaders now see the ability to connect those sources as the main condition for getting useful outputs from AI in day-to-day work.

Propel, which develops product value management software built on Salesforce, framed the findings around the role of a connected data foundation spanning product, quality, and content information. The pattern of responses, it said, shows organisations are becoming more specific about the tasks they want AI to support.

"This research shows many manufacturers have already identified what they want AI to do," said Ross Meyercord, Chief Executive Officer of Propel.

He added that the main obstacle is often making data usable across systems. "Organisations that start from a connected product, quality, and content foundation remove the first and most time-consuming obstacle of making their data usable for AI. Propel and Salesforce take care of the rest with a unified data structure and AI being an integral part of our platform," said Meyercord.