How Propel is Raising the Bar: Looking Back at Propulsion 2026

Every year, Propulsion gathers the product, engineering, supply chain, and quality leaders building the next generation of product manufacturing. And every year, the bar gets higher.

This year in Denver, the event delivered pivotal takeaways for the second half of 2026: 

• How MCP will change how manufacturers interact with data across systems
• Cross-industry customer stories with numbers that speak for themselves, including MaryRuth’s, JMA Wireless, Zoetis, and Nextern
• Strategic perspectives from Deloitte, KPMG, and Google Cloud on how regulated industries can make the biggest impact with AI while remaining compliant

Across every session, one message was consistent: the manufacturers who will lead the next decade are building the foundation today, instead of waiting for a "stable" landscape in an imagined future.

MCP Is Here. This Is What It Means for Manufacturers.

Last year at Propulsion, the conversation was about AI agents. This year, it was MCP, and during the Opening Keynote, Propel’s CEO Ross Meyercord and CPO Eric Schrader emphasized why it is the most significant shift in how manufacturers will interact with their systems in years.

Now being adopted rapidly across enterprise software, Model Context Protocol is a standard that allows AI systems to interoperate across applications without custom integrations. 

Instead of logging into multiple systems to resolve a single business problem, you describe what you need in plain language and AI determines where to find it, across your PLM, ERP, and third-party data sources.

What Can MCP Do? See It in Action.

Propel’s Mike Prom led a live demo of MCP that wowed the room, showing the following scenario:

A product manager facing a production-halting supply chain disruption uses Propel's networked AI to identify a back-ordered component, search for alternates inside Propel, expand the search to the web, evaluate options against specification requirements, create a new component record pre-populated with sourced data, and generate a complete change order with justification. 

Jump to 39:39 in the keynote video to watch demo.

One conversational thread, from problem to resolution.

Propel's networked AI via MCP will be available later this year, enabling Propel data to be queried and updated from outside of Propel, all in one natural-language chat.

The ROI in the Room: Customer Wins and Success Stories

This year's customer sessions made one thing clear: transformation is not a future state. It is happening now, across industries and at every stage of implementation.

MaryRuth's started with a single QMS use case in 2022 and built methodically from there — from PLM to PIM to agentic AI. An independent Nucleus Research study validated the outcome: $3.6 million in measurable business value generated over four years.

JMA Wireless went live in four months against a 12-month estimate, and more than doubled their process efficiency. Year one results exceeded every target set: scrap costs down nearly 40%, product releases nearly tripled, and ECR throughput up 30% with cost per change order down more than 60%.

Nextern migrated a 700-person global organization across four countries from Oracle Agile to Propel while processing 6,500 change orders annually. The presentation included a candid, persona-by-persona account of what large-scale PLM change management actually looks like in practice.

Zoetis, the world's largest animal health company, demonstrated that phased implementation done right is not slow. Building from parts and BOMs and now toward field service connectivity, their approach to cross-functional trust and early visible wins is a model for regulated enterprises navigating transformation at scale.

The AI Readiness Gap: Insights from Deloitte, KPMG, and Google Cloud

Propulsion 2026 brought experts from Deloitte, KPMG, and Google Cloud to the stage, and their sessions were designed intentionally to cut through the broader AI noise and provide attendees with real actionable advice for AI adoption.

The data foundation question is not optional. 

Deloitte's Kevin Prendeville, drawing on 30 years and more than 50 PLM transformations, and Google Cloud's Arnab Biswas both made the same case from different angles: AI is only as good as the data underneath it, and most manufacturers are not ready. End-to-end traceability is the prerequisite, not the bonus.

Regulated industries need to be ready for their first AI audit.

KPMG's Zachary Macht grounded this in the FDA's first warning letter to explicitly cite AI misuse, and his message was direct: build the oversight frameworks before the auditors ask for them. 

Stop optimizing AI for small wins. Go after transformation.

KPMG's Rajesh Misra challenged the room to stop optimizing for incremental gains and pursue the AI use cases capable of 20-30% bottom-line impact. The companies building AI into the fabric of their QMS are already pulling ahead.

Embrace AI as our co-engineer.

Propel CTO Kishore Subramanian showed how Propel has pulled ahead of the competition by embracing AI-first development. He demonstrated how AI “co-engineers” operating within human-defined parameters will accelerate execution while humans retain accountability for the judgment calls that matter.

A Community Built on Action

Propulsion would not be Propulsion without what happens between the sessions.

Across the Welcome Reception, Community Dinner, breakout sessions, and all the moments in-between designed to further conversation, attendees turned the week's big ideas into specific plans to enact back at their desks.

They were clear about what the experience was worth:

"Gets better every year. Kudos to the team that organized the event."

"I had a fabulous time. The people were wonderful. I learned so much. I had questions when I came, which I now have answers for."

"It left me very excited about the future."

For those who could not make it, the recordings are available now and packed with the kind of insight that turns aspiration into execution.

Watch all Propulsion 2026 sessions here.