AI use is no longer the differentiator it once was. McKinsey’s State of AI research finds that 88% of organizations now use AI in at least one function, while only a small group, 6%, are seeing significant bottom-line impact.
For engineering organizations, the more consequential question is what happens as AI increases the volume and pace of engineering work. Teams can explore more designs, generate more software, and create more tests. But the business advantage comes from understanding how those outputs will interact and influence system behavior before implementation and integration.
Behavior emerges from the interactions across domains.
These interactions cannot be fully understood through written requirements, architecture reviews, or source code alone. When unexpected behavior is discovered late in development, organizations face higher costs, program delays, compliance challenges, recalls, and reputational risk.
The strategic question for leadership teams is simple: How early can we determine whether our products will behave as intended?
Organizations that answer this question sooner make better-informed decisions.
Key Takeaway
Business advantage depends on understanding and validating system behavior before decisions lock in. The earlier teams know, the faster they can move.
Every successful product begins with intent. Business strategy defines objectives. Systems engineering defines requirements. Architects define structures and interfaces. But architecture alone does not prove behavior.
Many organizations discover late in development that products that appear sound on paper do not behave as expected under real-world operating conditions. Teams still need to evaluate interactions, test assumptions, analyze trade-offs, and verify requirements before implementation and integration.
The strategic objective is not simply documenting intent. It’s making intent executable.
Executable models transform requirements and architectures into representations that can be simulated, tested, evaluated, and verified before implementation becomes the primary source of truth. This is exactly what Lucid’s digital twins do, combining multi-physics plant models, controller models, and virtualized ECUs so engineers can test design assumptions in simulation instead of relying solely on documentation and design reviews.
Key Takeaway
Organizations achieve more predictable outcomes when intent can be executed, evaluated, and verified early.
Many organizations have already invested in systems engineering, digital engineering, verification automation, continuous integration, and AI-assisted workflows. What matters now is how well these capabilities work together.
When disconnected, requirements, simulations, tests, code, and documentation become isolated artifacts. When connected, they become a digital thread:
- Intent flows into behavior.
- Behavior flows into simulation and verification.
- Verification produces engineering evidence.
- Evidence supports implementation, certification, compliance, audits, and future decisions.
This is also where Model-Based Design with MATLAB® and Simulink® is strategically relevant. By turning system intent into executable behavior, it creates the digital thread that connects simulation, verification, and engineering evidence across the lifecycle, thus enabling organizations to scale AI-assisted engineering with confidence.
Key Takeaway
Decision confidence is built on evidence, not assumptions.
Generative AI and agentic AI are dramatically increasing engineering output. The question is no longer whether AI can accelerate engineering work. The question is whether organizations can govern, evaluate, and trust that output at scale.
Gartner notes that successful AI adoption requires organizations to “balance automation with human oversight, considering business criticality, risk and workflow complexity.” For engineering organizations, that balance comes from evaluating AI-generated work within established engineering workflows rather than treating AI as a separate activity. Engineers define objectives, constraints, and review points. AI accelerates execution, but accountability remains with engineers.
Without scalable verification, AI can increase complexity faster than organizations can manage risk. Executable systems provide a way to evaluate AI-generated outputs in context, connecting requirements, models, simulations, tests, and implementation through a common engineering thread.
For example, when a requirement changes, engineers can trace its relationships to design artifacts, tests, and results to understand what must be updated and revalidated. As a result, organizations can scale AI-assisted engineering without sacrificing traceability, or governance, or skipping over the established engineering workflow.
Key Takeaway
AI creates value only when organizations can verify and govern AI-assisted outcomes.
Innovation is no longer constrained by the generation of engineering artifacts. As AI and automation accelerate development activities, the new constraint becomes an organization’s ability to determine whether change produces the intended behavior.
Companies can generate more designs, more models, more code, and more tests than ever before. The limiting factor becomes how quickly they can evaluate those outputs with confidence.
| The Constraint Shift | |
|---|---|
| Then | Now |
Engineering Capacity
|
Decision Capacity
|
Verification throughput becomes a strategic capability because it determines how quickly organizations can evaluate change, verify results, and move forward. Simulation, automated testing, continuous integration, traceability, and engineering evidence create a repeatable path for evaluating change and making decisions with confidence.
Key Takeaway
The organizations that scale fastest will be those that can validate change with confidence.
This shift is already underway across industries.
| Organization | What Changed | Business Outcome |
|---|---|---|
| Nokia | AI-assisted verification | 43% fewer tests required for coverage closure |
| Collins Aerospace | Digital thread across development and certification | Earlier validation and improved traceability |
| Oceaneering | Digital twin connected to real-time PLC data | Reduced hardware risk and support for rapid prototyping |
| Medtronic | Earlier evaluation of system behavior | Accelerated innovation and validation |
Wireless Systems
Nokia Improves 5G Algorithm Testing with Machine Learning
Nokia developed a machine learning system that can categorize a large set of design tests and choose the best subset that provides coverage closure.
Challenge
- Testing complexity limited the ability to achieve coverage efficiently, with large volumes of redundant test cases.
- Traditional constrained random testing was limited in providing coverage closure in a reasonable time frame.
Shift
- Nokia used MATLAB to develop a machine learning approach for functional coverage.
- The team used an autoencoder model integrated with the circuit verification flow to help identify tests that targeted new, previously unchecked functionality.
Outcome
- The number of tests required for coverage closure was reduced by 43%.
- Verification quality improved through the reduction of redundant test cases.
- Cumulative simulation time decreased, reducing computational resource usage.
Key Takeaway
Better verification efficiency can reduce development costs while improving coverage and quality.
Aerospace and Defense
Digital Engineering at Scale: Concept to Certified Reality
Explore digital engineering at scale, and learn about emerging industry practices, from digital twin ecosystems providing a unified operational view for multiple stakeholders, to AI-powered knowledge management and engineering copilots.
Challenge
- Developing increasingly autonomous and software-intensive aerospace systems requires managing complexity while maintaining rigorous certification standards.
Shift
- Collins Aerospace established a digital thread that connects requirements, design, validation, and certification through a common source of engineering information.
Outcome
- The approach enabled earlier validation of design decisions across complex aerospace programs.
- Traceability improved from requirements through certification.
- AI detected anomalies, predicted failures, optimized designs, and helped accelerate engineering lifecycles.
Key Takeaway
Organizations that connect requirements, validation, and certification through a digital thread can reduce risk while improving engineering speed and confidence.
Industrial Automation and Machinery
Digital Twin Training with Real-Time PLC Data at Oceaneering
Using Simulink and Simscape, Oceaneering created a digital twin of its plant that communicates with a real PLC, enabling safe, realistic operator training.
Challenge
- Oceaneering wanted a virtual-first approach that replicated real plant behavior without risking actual hardware or interfering with existing automation setups.
Shift
- Engineers developed a virtual plant application using Simulink and high-fidelity Simscape models.
- The digital twin relayed simulated inputs through standard industrial protocols to a real controller, then received simulated feedback to create a closed-loop system.
Outcome
- Oceaneering built a functional simulator integrated with PLC logic.
- The same logic used with the real system was used in interactions with the virtual plant, reducing risks to hardware and improving safety.
- The workflow also supported rapid prototyping, validation, and deployment of analytics and control logic without disrupting existing automation setups.
Key Takeaway
Digital twins reduce operational risk while improving training, validation, and deployment readiness.
Medical Devices
Developing a Simulator and Multivariable Control Algorithms for Artificial Pancreas Systems
Using MATLAB, researchers at IIT develop artificial pancreas control algorithms that interpret physiological signals generated from a wristband device.
Challenge
- Developing an artificial pancreas required validating complex interactions among sensors, control algorithms, insulin delivery systems, and patient physiology.
Shift
- Medtronic adopted an engineering approach that enabled teams to evaluate and validate system behavior before deployment.
Outcome
- The approach accelerated the design, validation, and launch of new artificial pancreas technologies.
- It also improved Medtronic’s ability to evaluate and validate closed-loop system behavior throughout development.
Modernizing product development does not require transforming every legacy workflow simultaneously. Organizations can begin with a focused opportunity, demonstrate value, establish a repeatable approach, and then extend successful practices across programs and business units.
Teams do not need to build these capabilities alone. Training, consulting, and implementation support can help accelerate adoption while building internal expertise.
| A Practical Path to Scale | ||
| Phase | Leadership Focus | Intended Outcome |
|---|---|---|
| Access | Identify where earlier simulation, verification, or continuity would improve decisions | A focused opportunity with clear business value |
| Prove | Apply the approach to one program, capability, or use case | Evidence of value with limited organizational risk |
| Operationalize | Connect requirements, executable systems, verification, and engineering evidence | A repeatable and governable engineering workflow |
| Scale | Evaluate system behavior earlier | Broader gains in speed, predictability, and confidence |
Executive Questions
- What’s the smallest program where we could prove this work?
- Can we evaluate product concepts before physical assets exist?
- Can we scale AI without increasing engineering risk?
- Can we improve launch predictability?
- Can we strengthen certification and audit readiness?
- Can we modernize development without disrupting active programs?
Key Takeaway
AI adoption and digital engineering transformation are most effective when built on a foundation of executable systems and verifiable evidence, proven first on a narrow scope and then scaled.
The companies that lead their industries tomorrow will not be those generating the largest volume of engineering artifacts. They will be those that can move to verified outcomes with the greatest speed, confidence, and predictability.
As products become more software-defined and engineering workflows become increasingly AI-assisted, competitive advantage will depend on how quickly an organization can understand system behavior, evaluate change, and produce the engineering evidence needed to make informed decisions.
The future of engineering lies not in producing more, but in knowing sooner what will work. That knowledge leads to better decisions and better products.
