MedTech manufacturers are navigating greater regulatory complexity, shrinking margins, and growing expectations for highly specialized products. They’re also under pressure to generate revenue beyond the initial sale via service contracts, consumables, and digital offerings, while maintaining compliance over the life of the device.
Yet quality management (QMS) and product lifecycle management (PLM) platforms that these manufacturers depend on were largely built for either just quality assurance teams or siloed engineering teams. That’s a problem, because a product touches every part of the business. Regulatory, quality, marketing, sales, and service. All of these functions need product data, and most of them have been left to fend for themselves.
The walled garden problem
Most manufacturers intimately know this scenario: Their PLM system won’t connect to their enterprise resource planning (ERP) without a costly integration. Their QMS is siloed from engineering. Their customer relationship management (CRM) system has no visibility into which revision of a product a customer actually purchased. The data exists. It just doesn’t flow to the right person at the right time.
In a regulated environment, that fragmentation isn’t just a productivity problem, it’s a compliance risk. Take labeling. Getting a medical device label right requires alignment across regulatory, quality, engineering, and commercial teams simultaneously. When those groups are pulling from disconnected systems, errors quietly creep in. And “quietly” is dangerous in an industry where labeling mistakes can trigger FDA warning letters, recalls, or delayed clearance.
Field service faces the same problem from a different angle. When a clinician flags an issue with a device like an infusion pump or pacemaker, the service team needs maintenance protocols, complaint histories, field replaceable units, and component traceability on the spot. Digging through multiple systems to find root cause data isn’t just inefficient. When the device is life-sustaining, it affects patient outcomes.
The infrastructure shift
AI technology continues to advance at a rapid pace, while the industry has been slow to adopt. Anthropic released the Model Context Protocol (MCP), an open standard defining how AI systems connect to external tools, data sources, and business logic. The simplest way to think about it is USB-C for AI. MCP has one universal interface that eliminates custom connectors between every AI model, and every enterprise system.
Before MCP, wiring an AI agent into your enterprise meant building separate integrations for every system with unique authentication flows, data models, error handling, etc. The complexity scaled faster than the value. MCP collapses that problem. Any MCP-capable client connects to any MCP-capable server through a single implementation rather than rebuilding the connection every time.
And, adoption has moved fast with 97 million monthly developer toolkit downloads over the subsequent sixteen months. Major cloud and AI providers including Google and Microsoft have long adopted the technology.
The practical implication for manufacturers? AI agents can now reach across PLM, ERP, QMS, customer relationship management (CRM), and supplier portals and reason across all of that data in real time. And, they can do so without paying for a multi-year integration program.
AI that works inside the product record
MedTech companies can’t treat AI as something layered on top of existing systems. The regulatory environment won’t allow it. Every AI-generated action needs to be logged. Every data source needs to be traceable. Every recommendation needs to be grounded in controlled, approved product data.
When that’s true, the use cases start to compound. A regulatory specialist can ask for a summary of all design changes since the last FDA submission and get one, pulled from engineering change orders, design review records, and verification test results, traceable and audit-ready, in minutes instead of days. A quality team gets proactive flags on corrective and preventive action (CAPA) risk patterns before they surface as observations. A document control team asks AI to build a training quiz from updated product documentation and gets a draft back immediately.
None of this requires navigating a new interface or learning a new workflow. That’s the whole idea behind the no-UI model. The best user experience is no user experience. A plain-language question. A response grounded in the actual product record. The interface gets out of the way.
What getting this right actually requires
Five things matter most when deploying AI against regulated product data.
Data quality comes first. AI is only as reliable as the records it reasons over, so completeness and accuracy are preconditions, not nice-to-haves. Human oversight is non-negotiable for high-stakes decisions: design changes affecting device performance, CAPAs responding to patient complaints, final regulatory submissions. AI should accelerate those decisions, not make them. Behavioral testing and risk-based assessments are also essential. Continuously validate how the AI responds across scenarios to establish assurance and maintain documented confidence for its intended use. Business rules keep AI grounded in existing workflows, defining what happens when a specific prompt is entered and what guardrails keep responses in scope. And, access controls ensure the right data reaches the right roles, and only those roles.
These aren’t obstacles. They’re what makes AI adoption durable in a regulated business.
What comes next
The PLM and QMS market is in the middle of a substantial platform shift. The vendors that come out ahead won’t necessarily be the biggest, or the longest-tenured, they’ll be the ones whose underlying architecture has already been built for the open, composable world MCP is enabling. Platforms where product records, quality records, BOMs, change orders, and compliance documentation live in a single unified data model don’t need a translation layer when AI capabilities arrive. The product thread is already structured, governed, and ready.
For manufacturers still asking when AI becomes practical for product operations, the answer has arrived. The walled garden never fit how these enterprises actually work. The connected, agent-ready enterprise is the model that wins, where compliance and intelligence operate together.
A field engineer simply types a single prompt and gets answers that used to require digging through three enterprise systems. “Which components are failing on similar products?”, followed by “What replacement parts address this issue?”, then “Are any of them still under warranty?” That streaming flow of conversation is more of a reality than most of the industry expects.
Photo: StationaryTraveller, Getty Images
Chuck is the VP of MedTech and Life Sciences Industry Marketing at Propel Software. Formerly, as a Solution Architect and Program Manager at Stryker Corporation, he implemented and supported global PLM, QMS, and digitalization projects. Chuck has deep domain expertise on the development, compliance, and commercialization of medical device products, along with providing high-quality support in launching new products. Over 20 years of experience across senior positions in enterprise software solutions with companies such as Agile Software, Oracle, and PTC.
This post appears through the MedCity Influencers program. Anyone can publish their perspective on business and innovation in healthcare on MedCity News through MedCity Influencers. Click here to find out how.
