FDA Issues Comprehensive Draft Guidance for Developers of Artificial Intelligence-Enabled Medical Devices
Guidance Shares Strategies to Address Transparency and Bias, while Providing Key Considerations and Recommendations on Product Design, Development and Documentation
Today, the U.S. Food and Drug Administration issued draft guidance that includes recommendations to support development and marketing of safe and effective AI-enabled devices throughout the device’s Total Product Life Cycle. The guidance, if finalized, would be the first guidance to provide comprehensive recommendations for AI-enabled devices throughout the total product lifecycle, providing developers an accessible set of considerations that tie together design, development, maintenance and documentation recommendations to help ensure safety and effectiveness of AI-enabled devices. This guidance complements the recently issued final guidance on predetermined change control plans for AI-enabled devices, which provides recommendations on how to proactively plan for device updates once the product is on the market.
“The FDA has authorized more than 1,000 AI-enabled devices through established premarket pathways. As we continue to see exciting developments in this field, it’s important to recognize that there are specific considerations unique to AI-enabled devices,” said Troy Tazbaz, director of the Digital Health Center of Excellence within the FDA’s Center for Devices and Radiological Health. “Today’s draft guidance brings together relevant information for developers, shares learnings from authorized AI-enabled devices and provides a first point-of-reference for specific recommendations that apply to these devices, from the earliest stages of development through the device’s entire life cycle.”
The draft guidance includes recommendations for how and when, in marketing submissions, sponsors should describe the postmarket performance monitoring and management of their AI-enabled devices. The proposed recommendations reflect a comprehensive approach to the management of risk throughout the device total product life cycle. The FDA encourages sponsors to engage with the agency early and often, and to use this guidance, once finalized, to guide their activities throughout the life cycle of the device, including during planning, development, testing and ongoing monitoring.
Importantly, this draft guidance also includes the FDA’s current thinking on strategies to address transparency and bias throughout the life cycle of AI-enabled devices. The draft guidance describes specific recommendations intended to help a sponsor demonstrate they have addressed risks associated with bias and provides suggestions for the thoughtful design and evaluation of AI-enabled devices.
Notably, this announcement is specific to AI-enabled devices. Today, the FDA also published draft guidance with recommendations regarding the use of AI to support development of drug and biological products. The publication of these guidances, among other actions, continues to demonstrate the agency’s efforts to provide transparency and to help ensure product safety and effectiveness while supporting innovation in this rapidly growing field.
The FDA is requesting public comment on this draft guidance by April 7, 2025. In addition to general comments, the FDA is specifically requesting public comment on the draft guidance’s alignment with the AI lifecycle; the adequacy of the recommendations to address concerns that may be raised by emerging technology such as generative AI; the approach to performance monitoring (including use of a performance monitoring plan as a means of risk mitigation for AI-enabled devices); the type of information about AI-enabled devices that should be conveyed to users and the most appropriate approach to deliver that information. The FDA will also hold a webinar on February 18, 2025, to discuss the draft guidance.
https://www.jamasoftware.com/media/2025/01/FDA-AI-Guidance.jpg5121024Jama Software/media/jama-logo-primary.svgJama Software2025-01-15 03:00:052025-01-13 12:27:06FDA Issues Comprehensive Draft Guidance for Developers of Artificial Intelligence-Enabled Medical Devices
In this blog, we’ll recap a section of our recent Expert Perspectives video, “Integrating Safety of Intended Functionality (SOTIF) Into the Automotive Requirements Engineering Process” – Click HERE to watch it in it entirety.
Expert Perspectives: Integrating Safety of Intended Functionality (SOTIF) Into the Automotive Requirements Engineering Process
Welcome to our Expert Perspectives Series, where we showcase insights from leading experts in complex product, systems, and software development. Covering industries from medical devices to aerospace and defense, we feature thought leaders who are shaping the future of their fields.
In this episode, we speak with Dr. Hasan Ibne Akram on the topic of Integrating Safety of Intended Functionality (SOTIF) Into the Automotive Requirements Engineering Process.
Watch this video to learn more about:
The differences between SOTIF and functional safety
How to define and manage safety requirements addressing system limitations and edge cases
How to conduct a hazard analysis and risk assessment to cover intended functionality
Below is a preview of our interview. Click HERE to watch it in its entirety.
Kenzie Ingram: Welcome to Our Expert Perspective series where we showcase insights from leading experts in complex product systems and software development, covering industries from medical devices to aerospace and defense, we feature thought leaders who are shaping the future in their fields. I’m Kenzie Ingram, your host.
And today I’m excited to welcome Dr. Hasan Ibne Akram, an entrepreneur, computer scientist, book author, and CEO of engineering service company Matrickz based in Munich, Germany. With more than 17 years of experience in the automotive industry and working for two of the major German automotive OEMs, Dr. Akram brings a wealth of knowledge to this conversation. Today, we’re excited to showcase a discussion between Matt Mickle, Jama Software’s Director of Automotive Solutions, and Dr. Akram, on integrating safety of intended functionality, also known as SOTIF into the automotive requirements engineering process. Without further ado, I’d like to welcome Dr. Akram and Matt Mickle.
Matt Mickle: Thanks everyone for joining us today. My name is Matt Mickle. I’m the Director of Solutions for Automotive and Semiconductor at Jama Software. And I’m joined here today by Dr. Hasan Ibne Akram. Thanks very much for joining us today and answering some questions around integrating SOTIF into the automotive requirements engineering process. Dr. Akram, maybe we could start by just you telling us a little bit about yourself and your history with SOTIF and other industry standards and just a little bit about your background.
Dr. Hasan Ibne Akram: Absolutely. Thank you so much, Matt, for having me here. It’s amazing that we are having this conversation because this is very relevant today.
So my background in automotive started way back in 2005. So I was still a student, but I really wanted to go for a start-up. And back then, I landed a project with Continental. It was a braking system calculation project, and that’s how I got into automotive. And kept doing automotive stuff ever since.
And then, when I started my safety journey, I actually had no clue. So the first encounter to safety was a long time ago when I was actually working at a [inaudible 00:02:30] OEM as an external consultant. I was more responsible for the software. And during the lunch break, the functional safety colleague of that OEM, and in German, we call it FuSi, Funktionale Sicherheit, we used to call it FuSi. So I asked him, “What FuSi, the thing that you’re doing all the time? What is it about?” And quite condescendingly, he said, “We assume that whatever you guys are doing over there, every line of code, everything that you will do will go wrong.”
Akram: That was kind of like a light bulb moment for me. “Wow, that’s interesting. What happens when everything goes wrong? What do we do?” That was really my genesis of the functional safety journey. And SOTIF didn’t exist back then, was doing ISO 26262. And during my PhD, I was specialized in automotive cybersecurity, so cybersecurity and functional safety, I really wanted to bring them together.
And then, we realized, the automotive industry realized that, hey, there is something missing. Because with traditional safety, the definition of traditional safety is all about malfunction, if something goes wrong. Even when we’re doing security, it’s beyond malfunction, it’s all about attack now.Now comes autonomous vehicle, kind of like ADAS’s features, active distance control, automated emergency brake, active cruise control, and different levels of autonomy, Level 1, Level 2. The definitions came much later, but the idea of SOTIF was, hey, there’s something inherently required, there’s something required, something missing, inherently missing in the current standard because there can be hazards beyond malfunction.
It’s all about intention and this is where SOTIF was created, that we will talk about safety of the intended functionality. And my involvement, like you wanted to ask, my involvement with all these standards, I was following these standards before from the very ideas because the community is very, very close community. All the safety people in my podcast, I had Hans-Leo Ross, I had people who are the… Hans-Leo Ross even showed the birth certificate of ISO 26262 because he literally wrote the first lines and everything of ISO 26262. And I was privileged to be around these people who are actually shaping the future of these standards and how the engineering work will be done in the autonomous vehicle sphere and safety will be defined. So yeah.
Mickle: Nice. Well, that must’ve been quite enthralling at the time. So you mentioned that there was this gap sort of missing for functional safety and that SOTIF sort of filled that gap. Could you describe some of the key differences that are there between the standards?
Akram: Absolutely. So the key difference is, like I said, there was a gap. The gap was pretty evident, we’re talking about malfunction. If there is a fault, that fault would lead to a hazard, that’s ISO 26262, that’s traditional functional safety.
Now, what happens if there is nothing wrong in the vehicle, no malfunction, and we still have a hazard? So let me give you a metaphor. Imagine that you have a knife and you bought the knife. Your intention is to chop vegetables. So it’s a very sharp knife. The functionality is great, you’re chopping the vegetable, there is no malfunction, you’re chopping the vegetable. Now, by mistake, unintentionally, you cut your finger with it, it’s a hazard. Now, there is no malfunction still in the knife, the knife is 100 percent functional, it’s your intention that was to chop vegetables, but somehow, unintentionally, you cut your finger. And that’s where the safety of the intended functionality came in.
The famous example of such hazard is this high profile Tesla incident that happened, I don’t know, five, six years ago, where in a junction, because of the lighting condition, Tesla’s ADAS system could not recognize a truck that was passing the junction. And the driver happened to be watching Harry Potter and he didn’t pay attention. And this was fatal, I mean, the driver died. This was such a fatal accident. And there was nothing wrong in Tesla’s ADAS functionality, it’s just that this certain condition, there was no malfunction, this certain condition was not trained, and the ADAS system was not able to detect under certain lighting condition.
And that was the reason, but when we entered, when we started with this, it turned out vastly complex, the whole sphere of SOTIF, when you’re talking about the environment. I’ve just given you one example. So the environment is theoretically infinite. There can be infinite situations and there can be situations that we don’t know about. And the fact of the matter is, we don’t know what we don’t know. When you know something, you can take measure, that’s traditional ISO 26262. Now, we have this unknown unknown. You don’t even know what you don’t know. So that makes it extremely challenging and that’s why the whole process of autonomous vehicle development is going to be a continuous development process, we’ll have to continuously learn and incorporate safety and all those.
https://www.jamasoftware.com/media/2025/01/SOTIF-2.jpg10801920Matt Mickle/media/jama-logo-primary.svgMatt Mickle2025-01-14 03:00:482025-01-07 15:18:15Expert Perspectives: Integrating Safety of Intended Functionality (SOTIF) Into the Automotive Requirements Engineering Process
2025 Expert Predictions for Medical Device & Life Sciences: Innovations in Patient-Centered Care and the Future of Medical Device Design
As patient-centered care takes center stage, the medical industry is witnessing an unprecedented transformation in how devices are designed, developed, and regulated. From the rise of direct-to-patient products and AI-enabled diagnostics to the challenges posed by cybersecurity and evolving regulatory landscapes, 2025 promises to be a pivotal year for medical innovation.
We like to stay on top of trends in other industries as well. Read our predictions for Industrial & Consumer Electronics (ICE) HERE, Automotive HERE, Semiconductor HERE, and Aerospace & Defense HERE – Plus, stay tuned for our future topic, AECO.
With patient-centered care becoming increasingly important, how do you see software and digital solutions transforming the design and development of medical devices in the next few years?
Richard Matt: Software development needs to mature to a place of contributing equally with other specialties instead of excelling independently. Far too many companies are being held back by technical siloing, usually led by software. The tragedy is that software personnel are among the most creative and productive employees. Their collaboration needs to mature to a place of creating opportunities for other specialties to be as efficient and achieve this together.
Vincent Balgos: There is a growing trend of direct-to-patient products (hearing aids, CGM’s, smartwatch apps) that includes complex software and digital solutions. In addition, most traditional med devices are now connected to the internet or other devices. I’d expect this trend to continue in this digital age, but with that growth, there will be some unintended side effects. Specifically, cybersecurity threats will continue to become a significant factor during the design, development, and on-market phases of the product.
Regulatory compliance and data security are paramount in life sciences. What advancements in software do you think will be most effective in managing compliance and protecting sensitive patient data?
Matt: Big data will become more available and utilized systematically to provide answers to questions that have been answered inadequately for decades. Protecting patient data is a simple matter of awareness and giving cybersecurity the sliver of attention it needs to close off the dominant attack vectors.
Balgos: In recent conferences I’ve attended (and read about), AI/ML continue to dominate the discussion around software advancements. Whether providing internal value to organizations, or external facing with AI/ML enabled devices, the impact that AI/ML has (generative, predictive) will play a major factor in the security of data (whether positive or negative). The regulatory guidance of AI/ML is still evolving so it’ll be interesting to see how it unfolds in the future.
As AI and machine learning continue to evolve, what role do you see these technologies playing in medical diagnostics, treatment personalization, and device functionality by 2025?
Matt: AI and ML will continue to evolve, as they have for generations. We have made recent, significant steps forward in natural language recognition, but the integration of that forward movement with diagnostics and treatment personalization will continue to be slow and incremental.
Balgos: The expansion of AI/ML in various traditional areas of med devices is continuing to grow at an exponential rate. Looking at FDA’s dataset, the # of authorized AI/ML enabled devices continue to grow YoY as much as 40% with applications in new areas. Currently, predictive AI is supporting medical professions with their clinical assessment/decisions in a supportive role and seems to be the common use case. But there are current talks now on how generative AI could add potential value in these device areas as well.
What are the biggest hurdles the industry faces in adopting cutting-edge software solutions in device manufacturing and patient care? How can companies proactively address these challenges?
Matt: The biggest hurdle is technical siloing, which software leads very capably. Companies can proactively address this challenge by implementing a systems-approach to problem solving / product development that respects all the technical contributions needed to succeed and ensures software personnel use their exceptional abstraction abilities to work in collaboration with the rest of the company.
Balgos: Some of the biggest hurdles my customers talk to me about are the evolving regulatory landscape, continuing pressures to accelerate development, scalability and ongoing resource and budget constrictions. For changing regulatory, I do recommend folks to work with a qualified regulatory affairs profession for guidance (and is now required in EU). For acceleration, scalability and resource constraints, companies are proactively looking for ways to maximize efficiency and looking to innovative ways to help organizations (e.g., AI applications).
How do you foresee the growing demand for remote patient monitoring and telehealth impacting product development for medical devices? What innovations do you think are necessary to support these shifts?
Matt: COVID was an unplanned stress test on our remote work capabilities. We need to analyze the results of that stress test to identify strengths and weaknesses to build into our next generation products and services.
Balgos: Based on the new FDA CDRH Director’s speech at MedTech Conf 2024, these demands coincide with their Home as a Health Care Hub initiative where these homestead digital solutions are becoming more at the forefront of healthcare for many patients. With these increasing interests and demands, I think connectivity of these devices will continue to rise, AI/ML will play a major factor in delivering value, and also improvements in the human factors/usability aspects of these devices will be something to watch. Transitioning from a professionally trained clinical staff to the general population, there will need to be a shift in developing these at home devices to be extremely ‘easy to use’, especially with those that struggle with technology, complex processes, and tedious tactile tasks.
With these innovations, there will be some side effects such as cybersecurity, post market updates, and the interoperability of all these devices. While they seem daunting, I’m confident that industry will rise to the challenge as they have with other previous challenges
Are there any additional insights you have regarding predictions, events, or trends you anticipate happening in 2025 and beyond?
Balgos: In early 2026, the FDA’s final rule on the Quality Management System Regulation (QMSR) will be effective, thereby incorporating ISO 13485 elements into the current Quality System Regulation (QSR). QMSR will supersede QSR.
https://www.jamasoftware.com/media/2025/01/2025-01-08-predictions-medical-2025.png5121024Jama Software/media/jama-logo-primary.svgJama Software2025-01-09 03:00:202025-01-06 11:44:102025 Expert Predictions for Medical Device & Life Sciences: Innovations in Patient-Centered Care and the Future of Medical Device Design
In this blog, we overview our new datasheet – Click HERE to read it in its entirety.
Jama Connect® Enables DevSecOps Through Robust API and Integrations That Connect All Activity to Requirements
DevSecOps involves integrating security into all phases of the software development lifecycle. Rather than waiting to start analysis of potential vulnerabilities until after the software product, system, or subsystem is completed, this strategy puts security at the center of software development from the start to identify issues when the cost of resolving them is lowest. It also enlists everyone to play a part in identifying, assessing, and mitigating security risks in their individual development-related activities.
Comprehensive Security Risk Management and Seamless Tool Integration
The biggest challenge in achieving DevSecOps success is the need to assess and manage security risks across all software development tools and teams in an efficient and comprehensive manner. DevSecOps leaders choose Jama Connect because it is the only requirements management solution that provides the automation and collaboration required. Its robust REST API provides alignment with an integrated CI/CD pipeline including Jira, Azure DevOps, Git, GitLab, Subversion, Jenkins, Splunk, Kubernetes, Visual Studio, and Coverity. The Jama Connect platform delivers Live Traceability™, connecting all DevSecOps activity to the singular common element that defines value across all steps in the process — the requirement. It provides intuitive, accessible collaboration and review capabilities for internal and external teams.
Enable Cybersecure-by-Design Compliance with DO-326A Standards
Jama Connect for Airborne Systems supports a DevSecOps strategy by applying a cybersecure-by-design approach to meeting DO-326A standards. With Live Traceability, Jama Connect overcomes the disconnectedness of processes in the tool ecosystem that causes certification delays, cost overruns, product failures, audit findings, late identification of defects, and lack of visibility. It makes change management between software and hardware easier and reduces the effort needed to demonstrate requirements and test traceability required for certification.
KEY BENEFITS:
Integrate security across all DevOps and testing software Jama Connect’s robust open REST API and market-proven integrations with best-of-breed DevOps and testing software tools make it possible to connect all DevSecOps activity to the common element that defines value across all steps in the process – requirements.
Start identifying security vulnerabilities early in the development process Jama Connect reduces the risk of releasing code with security vulnerabilities by focusing on security requirements and
testing from the early stages of development.
Empower the entire team to contribute to DevSecOps Jama Connect’s easy-to-use collaboration and review capabilities provide an inclusive, safe, and collaborative environment for internal and
external development, security, and operations teams to build software that is efficient and secure.
https://www.jamasoftware.com/media/2024/12/DevSecOps.jpeg512986Jama Software/media/jama-logo-primary.svgJama Software2025-01-07 03:00:252024-12-31 11:21:00Jama Connect® Enables DevSecOps Through Robust API and Integrations That Connect All Activity to Requirements
Jama Connect Features in Five: Live Trace Explorer
Learn how you can supercharge your systems development process! In this blog series, we’re pulling back the curtains to give you a look at a few of Jama Connect’s powerful features… in under five minutes.
In this Features in Five video, Francis Trudeau, Product Manager at Jama Software, will introduce viewers to Jama Connect’s Live Trace Explorer, which auto-detects risk by bringing comprehensive and detailed insights into your complex development processes.
Please note that Live Trace Explorer is currently in beta and available for all Jama Connect Cloud customers to try.
VIDEO TRANSCRIPT
Francis Trudeau: Hello and welcome to the segment of Features in Five. My name is Francis Trudeau, and I’m a Product Manager at Jama Software. This video is an overview of Jama Connect’s Live Trace Explorer feature. Note that Live Trace Explorer is currently in beta and available for all Cloud customers to try.
The Live Trace Explorer is like a real-time map of the V-model, helping you check coverage completeness and validity across your project. It actively tracks metrics to spot gaps and risks between engineering teams so you can address issues early. This leads to a smoother development process, higher quality products, and faster time to market. This capability is a significant step in our vision to provide metrics for managing the development process through data.
To enable the Live Trace Explorer, go to the Admin tab, navigate to the Details section, find the Live Trace Explorer line, click Configure, check the box, and save. Once enabled, the feature appears in Admin Project settings and is available for Organization and Project Admins.
Trudeau: If permission is granted by their admins, users with a creator license can fully utilize the feature to load and configure existing diagrams. Once enabled, the Live Trace Explorer can be launched by right-clicking a project component or set to create a focused diagram for the selected node or right-clicking the project route to generate a comprehensive diagram showing all components and sets in sequence from top to bottom.
The resulting diagram visually represents the V-model with stakeholder needs, system requirements, designs, and components on the left, and their associated verifications and validations on the right. Each tile represents a component or set connected by trace paths. These paths are gray if there are no relationships between items and adjacent tiles, or they turn green and red to indicate the number of healthy or suspect relationships between them.
On the right side, the Verifications and Validation branch shows the number of Test Cases linked to items within the container on the left, no matter where they appear in the project. At the bottom of each tile, you’ll find a metric representing the ratio of these Test Cases included in a Test Plan. On the requirements side, the top part of each tile displays stats, including the number of items by type and any open conversations.
Trudeau: In the bottom half, you’ll find coverage metrics, essentially the ratio of active relationships to expected ones as defined by the traceability information model. For example, the model indicates that each high-level requirement should have two relationships downstream. Out of my four high-level requirements, three are covered by validations, giving me 75% coverage. Two are related to mid-level requirements, resulting in a score of 50%. In the Actions menu, you can access configuration settings to customize what’s displayed and measured. You can globally turn off item types, exclude specific relationships from consideration, or you can configure each tile separately.
A common use case consists of configuring your diagram for disabling relationships you are not expected to have at an early stage of your project. Then you may want to disable lower-level requirement items and relationships pointing downstream to them. Once applied, the coverage and total score will update automatically. Make sure to save your diagram once you have configured it to your liking. Live Trace Explorer updates in real-time, so any changes to project data instantly affect the metrics. For example, I can address a gap by clicking on the incomplete coverage. This will open Trace View where I can then establish a relationship to a mid-level requirement. Back in Live Trace Explorer, the metrics and total score summarizing all coverage will be updated after a refresh. You can keep a record and share these metrics by exporting a diagram as a PDF from the Actions menu at the top.
If you’d like to learn more about how Jama Connect can optimize your product, software, and systems development processes, please visit our website at jamasoftware.com.
https://www.jamasoftware.com/media/2024/12/Live-Trace-Explorer-FIF-2.jpg10801920Francis Trudeau/media/jama-logo-primary.svgFrancis Trudeau2025-01-03 03:00:562024-12-19 09:59:37Jama Connect® Features in Five: Live Trace Explorer™
In this blog, we recap our webinar, “The New ARP4754B: Tips for Engineers & Quality Teams” – Click HERE to watch it in it’s entirety.
Navigating the updates to ARP4754B can be challenging.
Understanding new safety analysis methods, validation and verification flexibility, and strategies to mitigate unintended behaviors is crucial for advancing aerospace development and ensuring compliance.
Join us as Cary Bryczek, Director of Aerospace and Defense Solutions at Jama Software, shares practical tips for engineers and quality teams to navigate the most impactful changes in ARP4754B.
Gain Insights On:
Changes from ARP4754A to ARP4754B
Model-Based Safety Analysis (MBSA) and Cascading effects Analysis (CEA)
Identifying and mitigating unintended system behaviors
Tying your safety analyses to requirements in Jama Connect
The updates to verification and validation methods
Below is an abbreviated transcript and a recording of our webinar.
The video above is a preview of this webinar – Click HERE to watch it in its entirety!
VIDEO TRANSCRIPT
The New ARP4754B: Tips for Engineers & Quality Teams
Cary Bryczek: We’re going to have fun talking about the changes from ARP4754B revision A to revision B. We’ll spend some time a little bit more deeply on its emphasis on model-based design and safety. I’ll talk about enhanced integration of safety and requirements management and some of the changes to validation and verification. At the end, we’ll have some time for Q&A.
A quick refresher on what ARP4754B is. Its title is Guidelines for Development of Civil Aircraft. It’s an industry guideline developed by SAE International that provides recommended practices for the development of complex civil aircraft and systems. It outlines a structured systems engineering process for the integrating of hardware, software, and human factors to ensure safety, reliability, and performance across the system lifecycle. The document emphasizes traceability, verification, and validation from initial concept through to certification with a strong focus on meeting regulatory safety and design assurance standards.
ARP4754B also aligns and is used in conjunction with other key aerospace standards like DO-178C and DO-254 offering detailed guidance on how to meet safety and certification requirements in the context of modern integrated aircraft systems. ARP4754 revision B is meant to expedite consistency with ARP4761 revision A, the safety assessment process, which was it was released on the same day in December of 2023.
The guideline describes generic aircraft system development process, which establishes a framework for discussing the process. ARP4754B doesn’t imply a preferred method or process, nor does it imply a specific organizational structure. At its simplest, it emphasizes the flow down of intended aircraft function through the system requirements management process and allocation of function to systems, subsystems, and hardware and software items.
Integral processes in the context of 4754B refer to key processes that are interwoven throughout the entire development lifecycle of aerospace systems from concept to design, integration, verification, and certification. Now, these processes ensure that various engineering disciplines, your systems engineering teams, your hardware and software engineering safety are fully integrated, aligned, and contribute to the overall success of the project.
Bryczek: This diagram from 4754B outlines the key stages of the aircraft system development process and provides a framework for understanding how safety is integrated into each stage. The safety are the ones that are in the lightest white or gray. The standard approach ensures that the safety risks are identified, analyzed, and mitigated early in the design process, and are continuously assessed throughout the system lifecycle.
I want to point out that lifecycle phases really are iterative and independent. 4754B emphasizes that the phases of system development aren’t strictly linear. For example, design and development may loop back to earlier phases such as the requirement’s definition. If issues are found during those later stages, sort of this iterative approach ensures that safety concerns can be identified and corrected throughout the lifecycle.
You’ll also notice that safety and hazard analysis is integrated throughout the development phases. Safety assessments are continuous activities throughout the development process. Safety considerations such as your functional hazard assessments, your fault tree analysis to your cascading effects analysis are embedded within multiple phases, particularly the design, development, and verification phases.
Let’s get to the meat of what has changed. So ARP4754B builds on the foundation laid by 4754A but offers a much more structured, detailed, and modern approach to developing complex aerospace systems. This is in response to the increasing complexity of our modern aircraft, tighter safety requirements, and evolving certification processes, particularly the need for rigorous system integration, traceability, and safety assessment practices. It provides greater clarity around the development assurance levels and how they relate to the overall system and safety requirements.
Bryczek: While A provided a basic framework, B refines the application of DALs throughout the system lifecycle. B expands the understanding of development assurance levels in the context of aircraft and system development, and it places a greater emphasis on safety, traceability, and integration across the lifecycle stages. The updated standard provides a more comprehensive guidance on managing the DALs and aligning the safety assessments with the system requirements, and it ensures that development processes are rigorous enough to meet the increasing complexity of the modern aircraft systems.
With the increased use of model-based techniques, 4754B highlights the benefits of using models to perform safety assessments. It recognizes that simulation-based safety analysis can help engineers assess the safety of complex integrated systems much more efficiently by modeling different failure scenarios and responses, so the standard supports using simulation tools to model those failure scenarios and validate the robustness of safety-critical systems. And this all just improves the accuracy of safety analysis, and it helps identify the potential issues earlier in the design process.
https://www.jamasoftware.com/media/2024/12/The-New-ARP4754B-Tips-for-Engineers-Quality-Teams.png9001600Cary Bryczek/media/jama-logo-primary.svgCary Bryczek2024-12-30 03:00:462024-12-19 11:27:49[Webinar Recap] The New ARP4754B: Tips for Engineers & Quality Teams
2025 Expert Predictions for the Semiconductor Industry: Innovations, Sustainability, and Globalization
The semiconductor industry is navigating a transformative era, marked by groundbreaking innovations and pressing challenges. As AI and machine learning demand faster, more efficient chips, semiconductor design and manufacturing are evolving at an unprecedented pace.
In part three of our annual predictions series, Michael Luciano, Senior Account Executive at Jama Software, explores the key trends shaping the industry. From advancements in silicon photonics and memory technologies to innovations in cooling systems and power delivery, these developments are poised to revolutionize chip performance while addressing critical energy efficiency needs.
Michael also addresses growing concerns about the environmental impact of chip production. With the immense power demands of AI-driven data centers and the continued use of harmful chemicals in manufacturing, the industry is turning to nuclear energy, novel materials, and refined processes as potential solutions. Emerging markets like India and China also play a pivotal role in future growth, highlighting the importance of global collaboration and infrastructure investment.
We like to stay on top of trends in other industries as well. Read our predictions for Industrial & Consumer Electronics (ICE) HERE, and Automotive HERE – Plus, stay tuned for future topics, including Aerospace & Defense, Medical Device & Life Sciences, and AECO.
With AI and machine learning driving demand for faster, more efficient chips, what key innovations in semiconductor design do you predict will transform these technologies, and how can companies balance performance with energy efficiency?
Michael Luciano: This is a great question. Key innovations in semiconductor design coming from increased demand with AI and machine learning (ML) will likely be on-chip optical communication using silicon photonics, continued memory innovation (i.e. HBM and GDDR7), backside or alternative power delivery, liquid cooling systems for Graphics Processing Unit (GPU) server clusters and superclusters.
Do you have any concerns or anticipate any negative impacts as it pertains to AI & ML?
Luciano: It’s understandable that people have concerns. Like every other tool that man has created, it’s important to create safeguards to prevent misuse and abuse. Agreeing on the exact safeguards and corresponding regulations is a highly contested and complex topic with wildly ranging global opinions. It’s undeniable that as AI systems and tools continue to evolve, these systems will replace some people’s jobs. This is already starting to happen. I am cautiously optimistic. As AI technologies become more advanced, with every negative impact I believe there will be an equal or greater level of positive impact for society and mankind elsewhere. Artificial superintelligence (ASI) is a hypothetical AI system with an intellectual scope beyond human intelligence. Mankind needs to see eye-to-eye before ASI comes to fruition or we are all in trouble. But don’t worry, we still have some time.
As chip production faces increased scrutiny for environmental impact, what role do you see for sustainable materials and manufacturing practices in the semiconductor industry, and how can software contribute to optimizing these efforts?
Luciano: In the context of the AI boom – the power required to operate gigawatt+ data centers is immense. Nuclear power is likely the most environmentally friendly way to go about it. Amazon and Google are currently investing heavily and recently formalized several key partnerships in this space. In the context of individual chip/device manufacturing – modern fabs also require a lot of energy/power. Nuclear powered systems will be the long-term answer. There are also a lot of nasty chemicals and gases that are used in chip production. I don’t see a clear way to fix this now, but as academia continues to study alternatives and companies continue to invest heavily in Research and Development (R&D) there is a possibility individual process steps can be adjusted/refined to incorporate novel materials or find other ways to help mitigate detrimental environmental impacts.
As the semiconductor industry becomes increasingly globalized, what emerging markets or regions do you see as pivotal to future growth, and how can companies foster effective cross-border partnerships and innovation?
Luciano: I identify Asia-Pacific (APAC) as the largest emerging market – specifically India and China, due to their populations. Companies can foster effective cross-border partnerships and innovation through significant investment in key infrastructure in those markets.
Are there any additional insights you have regarding predictions, events, or trends you anticipate happening in 2025 and beyond?
Luciano: AI Agents will mature and become widely used. This will significantly change how companies operate and go-to-market (GTM.)
https://www.jamasoftware.com/media/2024/12/2024-12-19-predictions-Semiconductor-2025.png5121024Jama Software/media/jama-logo-primary.svgJama Software2024-12-19 03:00:232024-12-20 12:30:022025 Expert Predictions for the Semiconductor Industry: Innovations, Sustainability, and Globalization
In this blog, we recap the “Write Better Requirements with Jama Connect Advisor™” webinar. Click HERE to watch it in its entirety!
Achieve Project Success with Clear, Effective Requirements
In this webinar, the speakers provide insights on how to leverage Jama Connect Advisor™, an easy-to-use, cutting-edge requirements authoring, editing, and analysis tool. Jama Connect Advisor uses Natural Language Processing (NLP) and evaluates and scores requirements against INCOSE EARS guidelines, enabling teams to create industry-compliant requirements, reduce risk, and improve efficiency throughout development.
You will learn how to:
Boost requirements clarity and writing speed as well as develop team skills with guided authoring
Track progress and improve requirements quality over time with downloadable reports
Improve the quality and usability of large volumes of requirement statements effortlessly with Batch Analysis
Save time on authoring, reviewing, and updating requirements
Confidently assess project readiness through requirements maturity analysis
Minimize rework risk due to ambiguity and contradictions
Below is an abbreviated transcript and a recording of our webinar.
The video above is a preview of this webinar – Click HERE to watch it in its entirety!
VIDEO TRANSCRIPT
Write Better Requirements with Jama Connect Advisor
Jeremy Johnson: Thank you so much to everybody that’s joining us today. This is a pretty special time for us to be able to take a new capability to market. From a product management and product development standpoint, it’s an extremely exciting time for us. So again, I appreciate everybody’s time in joining us here today.
Before we transition into the main portion of the session here, I want to provide a short introduction and an overview of our agenda. We’ll talk a little bit, for those who aren’t familiar with us, a little bit about Jama Software. We’ll talk a little bit about the trends in product development, and some of the challenges that we see in requirements authoring. We’ll also of course introduce you to Jama Connect Advisor, who it’s for, and how it works. We’ll get into a demonstration. We’ll also talk a little bit about our customer success program, specifically our customer success authoring workshop, and how we are now including and embedding the technology and the capabilities around Jama Connect Advisor into that consulting offering.
And then, as Juliette mentioned, our special guest, Sheila King will go into the requirements quality focus that she’s helping implement at Rockwell Automation, and we’re super excited and happy to have her. And then, we should have some time at the end of the session for some questions as well.
But again, starting with and moving into Jama Software’s role in the product development ecosystem, our vision and our purpose as an organization is to ensure that innovators succeed. And as you’ll see from today’s discussion and demonstration, that’s really at the core of what drove our introduction of Jama Connect Advisor.
From a broader solution standpoint, Jama Connect is the number one requirements management provider in the marketplace. We help teams with requirement management and product development through live traceability that also spans not only requirements, but the verification and validation components on the test side, risk management, and other key data that drives those processes forward.
The value that we hope these innovative organizations, our customers, derive is really focused around things like cycle time reduction, helping speed time to market, enabling through live traceability the ability to gain visibility and control over the organization’s product development processes, and really drive streamlining, really drive a tremendous amount of value, and ultimately ensure compliance and managing risk.
As far as organizations that we work with, we span medical device, automotive, industrial, machinery,and software, and this is just a sampling of the customers that we have the pleasure of partnering with. We have over 800 customers globally. These organizations span from smaller startup organizations to large global enterprises.
So with that very short intro to Jama Software, I now would like to review some of the complexity and challenges that we see today in product development, and of course to introduce you to Jama Connect Advisor.
Katie Huckett: Thanks, Jeremy. I’m really excited to talk about Jama Connect Advisor today and some of the things that are happening in the environment that led us to develop this solution. Today’s systems have become much more complex, and the emergence of the system of systems architecture has become the dominant approach for devices in all sectors, whether it’s aerospace, automotive, medical, and even consumer products. The system of systems is actually a collection of independent subsystems that are integrated into larger systems and deliver the unique capabilities required by users. The challenge is that it is difficult to predict accurate, predictable models of all emergent behaviors. So global systems of systems performance is difficult to design. That leads to testing and verification. Verifying upgrades to existing systems of systems is difficult and expensive as well, which is hard to scale. These are some of the factors that have led us to think about how we can help.
Another question we asked ourselves is why is requirements authoring so hard? If we look at the industry approaches for requirements authoring, we looked at the International Council on Systems Engineering’s (INCOSE) Guide for Writing Requirements. There’s a need to exercise a core subset of 40 rules in the INCOSE Rules for Writing Requirements, and in addition to that, an assessment of 49 requirement attributes. So just following INCOSE alone requires a substantial amount of training and understanding and then applying it, which can take a lot of time.
We’ve also found that EARS, the Easy Approach to Requirements Syntax, is being adopted by many organizations developing complex systems of systems. That includes Airbus, Bosch, Dyson, Honeywell, Intel, NASA, Siemens, and others. What EARS does is gently constrain the textual requirements. The EARS patterns provide guidance for writing a requirement sentence and provides syntax structure with an underlying rule set. Even these industry preferred approaches are challenging to apply, so we’re looking at how we might address that.
So today, just as a brief example, product requirements quality drives fidelity and efficiency in the product development cycle. If you look at this automotive example, there are many systems. It’s a complex system of systems that are dependent on each other. Any of these systems can lead to confusing the operator or systems operating optimally. If you look at the traditional V model of approaching systems engineering, the requirements are fundamental at the very early phase. So immediately after your needs analysis, you need to have really clear, concise, accurate requirements definitions.
The negative outcomes of poorly written requirements has been well-documented. It often leads to delayed time to market, late stage errors in the product, inaccurate translation of stakeholder needs into product attributes, and the lack of development team synergy. As teams are very organic today, the requirements need to be documented clearly and in an understandable way so that the team can execute with high performance. And then, ultimately failure and verification and validation can happen without high quality requirements.
Huckett: A secondary challenge is the training and reinforcement of requirements authoring skills. The lack of proper requirements can lead to product issues, and it’s a significant challenge in today’s environment. 30% of engineering degree holders are nearing retirement globally, and in the US 79% of American workers agree that to retain or increase their future employability, they need to continue with their learning and development. Computer scientists, 47.5% participate in work-related training to maintain and extend their skills, and engineers almost 60% do the same. So onboarding, retaining, and training system engineers remains a significant challenge.
With those items as a background, I’d like to introduce Jama Connect Advisor. Jama Connect Advisor is an add-on for Jama Connect Cloud. It’s an intelligent natural language advisor that improves the quality of requirements. It allows you to author intricate product requirements quickly, easily, and with precision. It is powered by engineering-based natural language processing, so not a general-purpose aid. It is engineering language-based. The advice provided is based on the industry-recommended best practices for the INCOSE rules and EARS notations.
Jama Connect Advisor has a very significant side benefit, while you use it, it augments skills and reinforces organizational preferences while authoring. So not only is Jama Connect Advisor doing the pragmatic work of improving requirements quality, but your systems engineers are learning how to do that more quickly and efficiently over time with its use.
When we look at Jama Connect Advisor’s capabilities, its features include analysis and advice from industry-leading practices, INCOSE rules, and EARS notation. The application is designed to put these two together to increase the quality, accuracy, and efficiency of requirement statements. So that’s its unique value. The guidance is provided seamlessly while you are editing in Jama Connect, which we’ll demonstrate in a moment. So really, the advantages are that experts can work faster confirming the application of INCOSE and EARS as they go, sharing their expert knowledge across the organization.
In this blog, we recap a section of our Datasheet, “Jama Connect for Defense Systems: Integrate DoD MIL-STE-882E Risk Management with Systems Engineering” – Click HERE to read it in its entirety.
Integrate DoD MIL-STD-882E Risk Management with Systems Engineering Using Jama Connect® for Defense Systems
Align hardware and software systems safety using Jama Connect as your single, secure platform for requirements engineering, risk analysis, and test management.
Military departments and defense agencies must follow the MIL-STD-882E Standard Practice for System Safety to ensure safety throughout the entire lifecycle of military systems, including development, testing, production, use, and disposal.
A key challenge to compliance is the need to integrate risk management and collaboration into the systems engineering process systematically across system and fire protection safety and occupational and environmental health disciplines.
Relying on a manual, document-approach using Word and Excel to manage the MIL-STD-882E risk register is inefficient and error-prone. A more reliable and intelligent solution is Jama Connect for Defense Systems which provides a single, secure platform for requirements, risk, and test management throughout the development lifecycle. It enables alignment of software and hardware development teams to achieve speed and quality, auto-detection of safety and environmental hazards and risks for early identification and mitigation, and robust collaboration and reviews involving internal teams, supply chain partners and government agencies.
Align the hardware and software systems’ team safety activities. Manage risk management for hardware and software system safety in Jama Connect which provides a single source of truth that integrates with best-of-breed software tools chosen by various teams.
Identify hazards and risks for early mitigation. Teams benefit from a single system and integrated data model for architecture, hazard assessment, analysis, safety requirements, and tests.
Accelerate development by streamlining collaboration and reviews. Avoid development delays by making it easy for internal and external teams to participate in MIL-STD-882E activities with Jama Connect’s Review Center and collaboration intuitive capabilities.
Get the most out of your requirements management and traceability solution. Use the same Jama Connect solution for managing and documenting your product requirements AND your MIL-STD-882E activities to maximize your return on investment from Jama Connect.
By leveraging Jama Connect, DoD systems development teams can significantly improve their efficiency, reduce risk, enhance safety, and expedite development while maintaining the highest standards of regulatory compliance with MIL-STD- 882E, contract requirements, defense data standards, interface standards, design criteria standards, manufacturing process standards, standard practices, and test method standards.
https://www.jamasoftware.com/media/2024/11/2024-10_dod-mil-std-882e-risk-management-w-systems-engineering-datasheet-2-1.jpg512986Jama Software/media/jama-logo-primary.svgJama Software2024-11-25 03:00:272024-11-20 11:51:07Integrate DoD MIL-STD-882E Risk Management with Systems Engineering Using Jama Connect® for Defense Systems
In this blog, we recap our webinar, “Key Systems Engineering Skills: Critical Thinking and Problem Framing” – Click HERE to watch it in its entirety.
Key Systems Engineering Skills: Critical Thinking and Problem Framing
Elevate your team’s success by exploring the role of critical thinking in a system engineering competency model.
In this insightful session, Chris Unger, Retired GE Healthcare Chief Systems Engineering Officer and Principal at PracticalSE LLC, and Vincent Balgos, Director of Medical Device Solutions at Jama Software®, discuss how critical thinking and decision-making skills are integral to systems engineering.
In this insightful session, you will learn:
Explore the vital role of critical thinking and decision-making in systems engineering.
Learn practical techniques for decision framing and closure.
Gain insight on how systems engineers should manage design decisions on a project.
See a simple model of how and when to engage with stakeholders in design decisions.
Below is an abbreviated transcript of our webinar.
Chris Unger: We’re going to talk today about a follow-up to the last webinar, where I’m going to talk about some of the most important systems engineering skills, critical thinking, and problem framing. So, how do skills in general, and soft skills, fit into improving systems engineering? So, in prior talks, I’ve suggested you keep your processes very simple but make them effective, and that’s easy to say but hard to do. That means you have to understand the system of the SE processes, how they connect, and where the diminishing value of the processes, the source process heading off, happens. As an example, a topic could be a technical risk, or it could be a trade-off between different possible solutions. So, we want to understand how those to the risk management and the decision process interact.
In order to do that, the best systems engineers have to have really good judgment. In addition, we have to influence people. Being simplistic, hardware and software engineers design things, things do what they’re told. I know it’s oversimplified, but our deliverables are instructions on how the software and hardware engineers do things. So, the best systems engineers here have an area of depth that they’re experts in, so they bring some technical credibility. They have systems of breadth, they understand all the systems processes and how they interact, and they have great interpersonal skills. Today I’m going to focus on how you achieve a balanced and optimized design, how you focus on your cost versus risk, and doing that through basically decision making.
So, first I want to talk about the Helix Model. So, the Helix Project was a project funded by the government and, the US government, and their concern was for big aerospace and NASA projects you tend to produce a major, billion-dollar development every 10 years, and then you do 10 years of support. So, people often move on. They were worried about how you create the truly brilliant leader systems engineers from a team that may be a little bit sparse. They developed this model up here in the front and simplistically, you start with things you learn in school, how to do good mechanical engineering, electrical engineering, and software engineering techniques. You then go into an organization, and so you spend the first five years learning about your company. Things like, well, if you’re going to be doing a say glucose monitor, what does blood chemistry look like? What does a sensor look like? What’s a workflow? So, you become a good organization-specific mechanical engineer.
Then you learn about lifecycle. How do you go from womb to tomb, from customer needs to disposal and disposition with all the regulations across the world in terms of chemical safety? So, after five, maybe 10 years, you understand your domain, you understand the lifecycle and you understand your technology. What differentiates after that? What they found was the skills on the bottom half of this page, the Systems Mindset, so big picture thinking, and paradoxical mindset. You’ve all heard that joke about fast, good and cheap, pick two of the three. Well, that’s the world in which systems engineers live. We make trade-offs between things that are inherently conflicting. The other thing is, we’ve got to make decisions quickly, so you’ve got to have a flexible comfort zone. You’ve got to be willing to wait till you have the critical information but make a decision without all the information you want.
Unger: In terms of the middle column, Interpersonal Skills, just the obvious stuff as I mentioned. You’ve got to influence the other engineers to make a good decision. Then finally here in Technical Leadership, balanced decision-making, and risk-taking. So, I had a general manager one time say, “We’re in the business of managing risks, not avoiding risks.” The least-risk program is also a boring one, but you also don’t want to take moonshots and everything. So, you really want to balance. It’s another case of a paradoxical mindset. Balance risk-taking with hitting a schedule predictably. So, these are the kinds of skills that really differentiate as systems engineering leaders, 10 to 15 years into your career. I’m going to talk more about these, decision-making, stakeholder management, and barrier-breaking.
So, I put together a very simple Systems Engineering Competency Model. I started with the NASA handbook and the NASA lifecycle. I simplified it, into that they had scope and requirements management separated, and I actually agree with those being different. But in reality, on the size of programs that we typically implemented, the people who did one typically did the other. Same thing, the architecture and the design, those were typically the same people. So, you have the upfront design, you have implementation. So, managing the subsystems actually do the implementation of what the design asks them to do, and you integrate it, such that you find your defects early. Then you manage all the lifecycle, the serviceability, manufacturability, disposability, and all the “ilities.”
Then leadership, obviously, there the interpersonal skills. This was developed for GE Healthcare, so I just picked it from our existing leadership skillset and I simplified it. What you’ll notice here is I put down at the bottom, critical thinking, as a technical skill. For many executives, and for other functional engineers, critical thinking is important, but as I mentioned, since we deliver instructions and designs to other engineers, framing decisions, taking vague things from product management and marketing, and turning them into clearer problems or functions to solve, I consider that a core technical excellence of systems engineering. But that’s vague. How do I actually measure that? So, I came up with this fairly simple set of observable behaviors. So, first of all, framing problems takes an ambiguous problem identifies the critical stakeholders, and turns them into a clear problem a more junior engineer can solve.
So, first, let’s talk about framing the problem. Even an entry-level person has to be able to understand a problem that’s been framed for them. But as you get to more senior people, the 10 to 15-year level, you have to be able to frame a complex problem, see around corners, use foresight to sort out essentials from the detail, and identify risks and emergent behavior that need to be incorporated in the decision, that other engineers might not see. Even at the strategist level, you can take a complex and ambiguous problem clarify the ambiguity, and turn it into simply just a complex and interconnected problem.
So, if we’re talking about maybe the 10 to 15-year-old person, not the most senior executives, you’ll be able to take a complex problem, identify ahead of time problems other people don’t see, and capture that. Balance cost, schedule, technical risk, and team capabilities, and make a trade-off based on sound evidence and data. Balance your intuition, when you don’t have all the data with waiting and gathering data where you need it. Then finally, making the decision is maybe the easy part. You have to make sure the team follows your leadership. Take accountability for making the right decisions, delegate where you can, and then ensure that the entire team buys into the decisions that the team or you have made. So, that’s the theory.
Unger: Let’s talk about how we manage design decisions. First of all, why? Why is this a critical skill? By identifying the critical design decisions, it allows the team to focus on the most important thing, and separate out the core from the distractions. It helps teams identify work items. So, for example, one time when I was working with the ultrasound team in Japan, we had a bunch of really experienced engineers and they were working on a new ultrasound probe. It had moved an active component into the probe and there was a thermal issue. They were talking in Japanese for about five, 10 minutes when I was asked to frame the problem and I said, “Yeah, you’re talking too fast and too much. This is not that easy. Come back to me and tell me what you’re actually doing.”
They were figuring out how to measure the thermal properties in the lab. I said, “Well, imagine you had a probe that was safe, with maybe 39°C, but that was uncomfortable to handle. Have you worked with the application people on how much value? If you spent $50 more and took the temperature down by 1°C, would that be worth a trade-off? The team, “Oh, that’s interesting.” They were actually focused on the technical feasibility, not the real market and customer acceptance problem. So, by doing this upfront, you can make sure that you have a complete work process for the team. Then once you’ve made the decision, it minimizes rework by making sure the decisions stay closed.
Now, this decision list and prioritization should start early. It would be comfortable to wait until you know everything, but that’s too late. So, it’s a living document. Don’t wait to get started until you have enough information to make a good plan. Start with what you know, and then build out as you continue. So, one of the first things I talk about is, what is a decision? As an example, I’ve had teams come to me saying, “The operating system selection is a decision.” It’s like, “No. It’s actually not typical. It’s typically a collection of decisions.” So, I draw this little arrow here. It’s basically a decision is a point in which you select between different paths going forward and you pick one way versus another. So, deciding whether to include a stretch item in scope or not is a decision. Deciding between very specific designs and implementing a feature is a decision. Setting a critical to-quality parameter or balancing between different parameters, so cost versus reliability or cost versus performance, is a decision.
https://www.jamasoftware.com/media/2024/03/Key-Systems-Engineering-Skills-Critical-Thinking-and-Problem-Framing.png5121024Vincent Balgos/media/jama-logo-primary.svgVincent Balgos2024-11-21 03:00:452024-11-18 10:07:27[Webinar Recap] Key Systems Engineering Skills: Critical Thinking and Problem Framing
By leveraging Jama Connect, DoD systems development teams can significantly improve their efficiency, reduce risk, enhance safety, and expedite development while maintaining the highest standards of regulatory compliance with MIL-STD- 882E, contract requirements, defense data standards, interface standards, design criteria standards, manufacturing process standards, standard practices, and test method standards.