Propel Software has outlined five trends it expects to shape product innovation in manufacturing through the second half of 2026. Among manufacturers exploring artificial intelligence, Model Context Protocol drew the strongest interest.
The trends focus on enterprise integration, product development, governance and quality management as manufacturers apply AI across design, supply chain and commercial operations. Propel based its assessment on discussions with manufacturing and technology leaders, including participants from Deloitte, Google and KPMG.
Integration shift
Top of the list is Model Context Protocol, or MCP, which Propel described as a way for AI systems to query business software through prompts rather than custom-built integrations. It argued that this approach could help manufacturers address long-standing problems caused by data being spread across separate product, quality, supply chain and sales systems.
For manufacturers, the challenge is not simply connecting software tools but creating a single view of information used across a product's lifecycle. Many businesses have layered new digital systems onto older infrastructure over several years, leaving engineering, operations and commercial teams working from different records.
Propel said MCP is moving into advanced evaluation discussions across large organisations. That points to a broader shift in how companies are approaching AI deployment, from isolated productivity tools to systems that span multiple functions.
Co-engineer role
The second trend is the emergence of AI as what Propel called a "co-engineer" rather than a basic assistant. In practice, that means using AI within design, quality review and product iteration processes instead of limiting it to drafting documents or answering general questions.
According to Propel, context, memory and access to tools determine whether an AI model can support engineering work in a meaningful way. That could allow teams to test more design options and shorten development cycles while leaving final judgement with human staff.
This reflects a wider debate in manufacturing and product development over where AI fits within regulated and high-risk workflows. Companies are under pressure to raise output and reduce delays, but they also need to maintain traceability over changes, approvals and technical decisions.
Changing product record
Another trend is the expansion of the digital thread as product definitions move beyond physical goods. Manufacturers increasingly sell combinations of hardware, software, services and subscriptions, meaning the underlying product record must cover more than engineering specifications.
That shift affects how companies manage time to market, margins and after-sales performance. A connected product record can link decisions made in development with outcomes in the field and in the commercial organisation, while fragmented systems make that harder to track.
Propel said manufacturers gaining ground are treating product data as a live business asset rather than a static engineering document. The implication is that product information management is becoming as much a commercial issue as a technical one.
Governance pressure
The fourth trend is growing accountability for AI-driven decisions. Propel said companies are moving from treating AI governance as a recommended practice to treating it as an operational requirement, especially where automated systems influence design, quality or compliance decisions.
For manufacturers, the concern is whether they can show who was responsible for a decision that AI helped shape, how unexpected behaviour is detected and whether systems continue to operate as intended over time. In sectors with strict compliance demands, those questions affect risk management as well as internal controls.
Propel argued that businesses need to move beyond a passive "human in the loop" model towards "human in command", where responsibility remains explicit and auditable. That stance aligns with a broader industry emphasis on traceability and oversight as AI systems become embedded in core processes.
Quality and revenue
The fifth trend is a change in how manufacturers view product quality. Propel said AI-enhanced quality management is moving from an early-stage practice to a competitive requirement, with leading companies using connected quality data to catch failures earlier and guide design choices.
In that framing, quality is no longer treated only as a cost of control and remediation. Instead, it is linked to growth, customer retention and product performance by reducing defects, limiting field issues and improving confidence in launches.
Propel said this depends on a digital thread that links engineering data, quality events and field performance in real time. When those elements are connected, manufacturers can identify problems sooner and decide what to change before committing further resources.
Ross Meyercord, Chief Executive Officer of Propel, set out the themes in the company's market assessment. "The question isn't whether companies are using AI, it's whether they can prove who was accountable for every AI-influenced decision. How is unexpected behavior detected? How do they know their agents are still performing as intended months after deployment? Governance frameworks that define which processes are non-negotiable for agents aren't optional. They're what keeps automation from becoming a liability."