How AI is fundamentally reshaping software engineering practices, team structures, and developer productivity
Our CEO, Chris Cera, recently attended the AI Engineer Code Summit, where industry leaders gathered to discuss the rapidly evolving landscape of AI-powered software development. As an AI-first healthcare product studio, we’re deeply invested in understanding how AI tools are transforming how we build digital products. The insights from this summit reveal some critical inflection points for the entire industry.
Here are the four key takeaways that are shaping how we think about AI-assisted development at Arcweb. Links to the full presentations can be found at the end of the article.
1. The 90% vs 100% Problem: Why Complete AI Adoption Matters
Dan Shipper from Every Inc. presented perhaps the most thought-provoking concept of the summit: he demonstrated how a single engineer can now build and maintain four separate, complex production products using AI tools.
The real insight extended beyond individual productivity to organizational transformation. Shipper highlighted a critical distinction: there’s a fundamental difference between a company where 90% of engineers use AI and one where 100% do.
Why this matters: When only some of your team use AI tools, you create a two-tiered system with mismatched expectations, workflows, and output quality. The engineers who do not use AI become bottlenecks. Communication patterns break down, code review processes don’t align, and the team can’t fully optimize around AI-assisted workflows because they’re still accommodating traditional approaches.
The Arcweb perspective: This is the issue of our time in software development. As we build AI-focused solutions for our clients, we’re seeing AI adoption trends play out in real time within our own teams and when working closely with clients and partners.
Organizations that commit fully to AI-powered software development demonstrate measurably faster delivery. GitHub’s research on their Copilot tool found that developers achieved up to 55% productivity increases and reduced cycle time by an average of 3.5 hours These gains require organizational commitment beyond just providing access to tools, however. As McKinsey’s 2024 State of AI report notes, while 65% of organizations now use generative AI regularly (double the previous year), most still struggle with integration challenges and scaling across teams.
In healthcare innovation, where time to market can literally save lives, the real value lies in delivering impactful solutions efficiently and at scale. This is why our positioning as an “AI-First” studio is a fundamental commitment to building our processes and culture around these tools from the ground up.
2. Moving Away from Agile: Rethinking Team Structure in the AI Era
Martin Harrysson and Natasha M. from McKinsey & Company presented fascinating research on how AI is forcing organizations to reconsider established software development methodologies, including the sacred cow of Agile.
Key observations:
- Large teams are increasingly splitting into smaller, more autonomous units
- Traditional Agile ceremonies and structures are being questioned
- The problem of large pull requests is becoming more acute as AI generates more code faster
Why this matters: When AI can generate substantial code blocks in seconds, many Agile practices need to be reconsidered. Rituals like stand-ups, sprint planning, and story pointing were built around development at human speed.
The Arcweb perspective: We’re already adapted our processes to accommodate AI-powered software development. Smaller, more focused teams can be even more efficient when empowered with AI tools. We haven’t abandoned all Agile principles, but we have evolved them.
Compliance, security, and quality aren’t negotiable in healthcare development. We structure projects around small, autonomous teams, paired with rigorous code review protocols explicitly designed to maintain clinical-grade standards while capturing AI’s productivity gains. Speed matters, but not at the expense of the ethical, safety, and regulatory requirements healthcare demands.
3. The Mixed Reality of AI Productivity Gains
Perhaps the most sobering presentation came from Yegor Denisov-Blanch from Stanford’s Software Engineering Productivity Research Group, who asked the critical question: “Does AI Actually Boost Developer Productivity?”
The answer? It’s complicated.
Key research findings:
- Results are mixed when you measure real-world productivity gains
- “Reviewer burden” is increasing. Humans need more time to review AI-generated code
- “Rework” is becoming a significant issue, which is code that needs to be fixed or rewritten after initial AI generation
- Recent changes in code quality are raising important questions
Why this matters: The industry narrative around AI development tools has been overwhelmingly positive, but rigorous academic research is revealing a more nuanced reality. Yes, developers can write code faster with AI assistance. But if that code requires extensive review and frequent rework, have we actually gained productivity?
The Arcweb perspective: This research validates what we’re seeing in practice. AI tools are incredibly powerful, but they’re not magic. The real productivity gains come from:
- Strategic AI use: Knowing when to leverage AI and when human expertise is more efficient
- Strong review processes: Investing in thorough code review becomes even more critical, not less
- Training and expertise: Developers need to learn how to prompt, review, and refine AI-generated code effectively
For healthcare applications where errors can have serious consequences, we’re particularly focused on quality and review. Our approach isn’t to blindly accept AI-generated code, but to use AI as a powerful tool within a rigorous quality-assurance framework, where human developers always have the final say.
4. Anthropic’s Ethical Leadership in AI Development
Throughout the summit, one company stood out not just for its technical capabilities but for how it’s approaching AI-powered software development: Anthropic.
Key observations:
- Anthropic is gaining significant traction among engineers, particularly for code generation, where Claude holds 42% market share which is more than double OpenAI’s 21%. Developers cite the company’s focus on security, privacy, and ethical AI development as key differentiators for enterprise applications.
Why this matters: As AI tools become infrastructure-critical to software development, the companies behind these tools matter more than ever. Developers are increasingly evaluating AI platforms on factors beyond raw performance. They’re considering the values, safety practices, and long-term vision of the companies building these tools.
The Arcweb perspective: As a B Corporation working in healthcare, we appreciate Anthropic’s early leadership in bringing ethical considerations to AI development. When we’re building solutions that handle sensitive patient data or support critical healthcare decisions, we need AI partners who prioritize safety, privacy, and ethics.
This perspective guides our platform-agnostic approach as we explore the breadth of AI tools such as Gemini, ChatGPT, Claude (especially Claude Code and the Agent SDK), and other specialized platforms for our own AI initiatives. We recognize that different providers bring different strengths to the table, and the combination of technical excellence and ethical foundation across the ecosystem enables responsible healthcare innovation and AI-first solutions that scale.
Check Out the Talks
Didn’t get a chance to make it out to the summit? Interested in diving deeper? Check out the full talks on YouTube:
What This Means for Healthcare Innovation
These insights from the AI Engineer Code Summit have direct implications for how we approach healthcare innovation at Arcweb:
- Clear alignment with AI is our competitive advantage
We are building our entire development culture around AI tools, enabling faster and more efficient delivery for our healthcare clients. This lets us move faster and more efficiently for our healthcare clients. - Quality processes are more important than ever
AI speed without quality controls is dangerous, especially in healthcare. We’re investing heavily in review processes and quality assurance frameworks designed for AI-assisted development. - Ethical AI partnerships matter
Working with AI providers who share our commitment to safety, privacy, and ethical development is foundational to responsible healthcare innovation.
The Road Ahead
The conversations at the AI Engineer Code Summit confirm what we’re experiencing firsthand: we’re in the middle of a fundamental transformation in how software gets built. The developers and organizations that will thrive in this new era will be those who thoughtfully integrate AI tools while maintaining high standards for quality, security, and ethics.
For healthcare innovation specifically, this transformation represents an unprecedented opportunity. AI-assisted development can help us build better clinical tools faster, deploy wellness protocols more efficiently, and ultimately improve patient outcomes at scale. But only if we do it right.
At Arcweb, we’re committed to leading this transformation responsibly: building AI-first solutions that honor the trust healthcare organizations and patients place in us.
Want to discuss how AI-first development can accelerate your healthcare innovation project? Get in touch with our team to explore how we can help you build better healthcare solutions faster.
Check out Arcwell, our open-source platform for delivering wellness protocols and clinical research—built from the ground up for accelerating and empowering clinical trials management and driving better outcomes for patients.
