FYAI: Why developers will lead AI transformation across the enterprise

Developers are at the forefront of adopting AI technology, and they’re transforming industries along the way. By utilising AI tools like copilots and agents, they’re speeding up software delivery, cutting down on manual tasks, and increasing their confidence in building new applications. Much like they did with automation, developers are reshaping customer experiences and making operations smoother, unlocking the true potential of AI.

In this edition of FYAI, a series highlighting AI trends with Microsoft leaders, we chat with Amanda Silver, the Corporate Vice President and Head of Product, Apps, and Agents. Under Amanda’s leadership, Microsoft is embracing open-source collaboration, working towards a future where AI revolutionises how developers create, deploy, and continuously refine large-scale applications.

In this Q&A, Amanda explains the significance of developer-led AI adoption, the impact of agent-driven DevOps on workflows, and what actions leaders can take now to enhance their effectiveness.

How is AI reshaping app delivery for developer teams?  

AI is streamlining handoffs across the entire software lifecycle. While DevOps united building, testing, deploying, and operating, earlier stages—like discovery and needs assessment—were often overlooked. Now, copilots can transform straightforward ideas into specifications and frameworks, while agents manage tests, updates, and runtime operations. This change leads to a more efficient cycle from concept to outcome: reduced iteration costs, swifter transitions, and more flexibility to refine the product until it meets business needs. It’s like the transition to public cloud: before that, teams waited weeks to set up hardware and commit funds; now, cloud environments can be established in seconds, and you only pay for what you use. AI introduces that same responsiveness to product definition and delivery, removing initial friction and allowing teams to iterate based on genuine feedback. Simply put: while the cloud eased infrastructure challenges, AI eases the journey from intent to execution.

How is AI helping developers reimagine their work?  

AI is converting software delivery into a genuine idea-to-operation system. Developers now spend less time on tedious cleanup tasks and more time on creative, high-impact projects. Copilots and agents manage the monotonous, often unnoticed tasks like debugging, updating dependencies, and applying security patches. No longer reliant on quarterly ‘tech debt sprints,’ teams can progressively tackle debts in the background through agent-driven DevOps.

One notable example is the way AI is speeding up migration and updating processes. Previously, updating frameworks or transitioning to new platforms could take months of planning and manual intervention. Today, agents can automate .NET and Java upgrades, handle breaking changes, and orchestrate extensive migrations, trimming down timelines from months to mere hours. This isn’t just about speed; it’s also about maintaining healthy and modern codebases so developers can focus on innovating and enhancing user experiences.

Overall, developers are spending less time putting out fires and more time innovating. Technical debt transitions from a looming threat to a manageable aspect of the process. With AI agents taking care of routine tasks, teams can operate in a more balanced state, ensuring healthier code and quicker delivery.

What are the implications for apps? Are they improving, and how does this affect developers?

Apps are set to improve because they evolve into learning systems. With AI involved, teams move from a ship and hope strategy to a continuous process of observe → hypothesise → alter → validate, allowing them to refine products for better market fit continuously. AI can analyse data (like user behaviour and feedback), identify pain points, suggest improvements (like content, layout, and features), and even manage feature flags or experiments to test change effectiveness. This drastically cuts down the time needed to learn and accelerates alignment with user preferences.

User Interactions: PreAI vs. PostAI  

• PreAI: Users navigate complex menus and intricate architectures, searching for a specific control. Each step carries the risk of encountering dead ends, and context can easily be lost while switching tools.
• PostAI: Users can express their intent using natural language—whether through text, voice, or other modes. The app grasps that intent, maintains context, and guides users to the relevant data or workflows, often dynamically generating the user interface (like filling out forms and filtering results). This transforms the conversation from “Where do I click?” to “Here’s what I need—help me with it.”

What changes for developers?  

• From page builders to experience composers. Developers are now creating intent routers and orchestrations that connect different models, agents, data, and services, allowing the app to adapt intelligently to varied user goals without rigid click pathways.
• From manual analysis to AI-assisted product loops. Instead of manually tracking down insights and creating dashboards, AI identifies opportunity areas, proposes test plans, and suggests code and configuration changes through pull requests. Developers simply review, adjust, and implement—all within set guidelines.
• From “debt sprints” to ongoing modernization. AI agents keep applications up to date by managing framework upgrades (like .NET and Java), fixing dependencies, patching vulnerabilities, and standardising processes—ensuring feature development continues smoothly. This transforms technical debt into a manageable, ongoing task rather than an occasional crisis.

In summary: AI strengthens the connection between what users desire and how the app evolves. Developers can focus less on wiring menus and manual investigations, and more on defining intent, crafting intelligent flows, establishing success metrics, and overseeing safe, measurable changes. Apps are improving rapidly—not just because they are smarter, but because teams can experiment, learn, and adapt as usage increases.

How is Microsoft distinguishing itself in the AI landscape?  

Microsoft’s key advantage lies in our ability to connect AI agents with the systems, data, and workflows that drive genuine business results. We cater to organisations with extensive, complex codebases and intricate operational needs, providing tools specifically designed to meet these challenges. Through GitHub, Visual Studio, and Azure AI Foundry, millions of developers can access advanced models and agent capabilities right within their usual workflows, all backed by enterprise-level security and responsible AI standards.

What truly sets Microsoft apart is the depth of our integrations. AI agents created on our platform can leverage a vast ecosystem of business applications, data repositories, and operational systems—be it enterprise resource planning (ERP), customer relationship management (CRM), human resources (HR), finance, or tailored business solutions. Thanks to open standards like Model Connector Protocol (MCP) and Agent-to-Agent (A2A) connections, agents can securely orchestrate and automate across diverse environments, enabling significant outcomes: workflow automation, modernising legacy systems, and fostering continuous improvement.

Yina Arenas’s Agent Factory series illustrates how Microsoft is establishing the framework for building safe, reliable AI agents—from rapid prototyping to full-scale deployment, observability, and real-world applications. Our platform is more than just about creating agents; it’s about empowering them to work seamlessly with existing systems and data that businesses already depend on, enabling teams to progress from initial experiments to impactful enterprise-wide implementations.

Ultimately, Microsoft’s edge isn’t merely based on scale; it’s our ability to make AI agents genuinely beneficial by linking them to the core of businesses, using tools and standards that ensure safety and security.

When should developers determine which tasks to assign to agents versus handling themselves for optimal impact?  

As my colleague David Fowler puts it: “Humans are the UI thread; agents are the background thread. Don’t block the UI!” Developers should devote their energy to creative and critical thinking tasks—defining intent, making architectural choices, and enhancing product experiences. Agents excel at managing repetitive, time-intensive, or cross-cutting responsibilities that can run effortlessly in the background, like maintaining code quality, upgrading dependencies, analysing telemetry, or creating initial frameworks to overcome creative blocks.

The goal is to assign any task that hinders your workflow or distracts from impactful projects to an agent. If a task is routine, tolerant of delays, or easily undoable, let an agent handle it. However, if it demands extensive context, product insight, or has the potential to shift your app’s trajectory, keep it in the hands of a human. This ensures developers remain quick and focused while agents continuously enhance the codebase and operations simultaneously.

By finding the right balance, developers can spend fewer hours on routine tasks and concentrate on what truly drives products and teams forward.

Software development generates the kind of rich, structured signals that AI thrives on. Code changes, pull request reviews, test results, and performance metrics are all timestamped, labelled, and traceable. This richness makes it the ideal environment for applying machine learning: models can learn from actual work, assessed against objective metrics (like tests and policies), and improve within existing feedback cycles (like Continuous Integration and Continuous Delivery (CI/CD)). We’ve got the data, instrumentation, and validation built-in.

Additionally, there’s a cultural element at play: developers automate to eliminate frustration. From compilers and build systems to version control, CI/CD, and infrastructure as code, generative AI is the next progression. It shifts more tasks from manual creation to intent specification and outcome supervision: copilots assist with exploration and quickening processes; agents ensure ongoing code health, updates, and safe, reversible actions. Investment flows here since an improved developer experience correlates with throughput, quality, and time to value.

Indeed, the future begins with developers. As development teams uncover where AI meaningfully supports their workflow, these patterns can spread throughout the organisation, enhancing how all functions experiment, learn, and deliver.

Empowering developers with AI for lasting impact 

We’re entering a groundbreaking phase of software delivery that is agent-driven, adaptable, and centred on human needs. With copilots and agents involved, developers can create systems that constantly adjust to business demands. At Microsoft, we’re empowering developers to swiftly move from idea to significant impact by prioritising creativity, product vision, and the responsible use of AI.

In fact, leading companies are already showcasing what’s achievable. They view software as a dynamic entity—refined through telemetry, experiments, and insights powered by AI. Across various organisations, promising AI applications are emerging—from customer service to software development—setting new benchmarks for what is achievable with the latest AI tools.

Interested in learning more? Explore resources and tools to enhance your AI journey: