Meet 4 Developers Leading the Way with AI Agents
AI agents are transforming how businesses automate processes and deliver value. These agents work alongside or on behalf of individuals, teams, or organizations to execute complex workflows. According to Microsoft’s 2025 Work Trend Index, nearly half of company leaders report using AI agents to automate workflows, and over 80% expect an agentic workforce within the next 12 to 18 months.
At Microsoft Build 2025, new AI agent tools were introduced to help developers build efficient and secure AI solutions. Here’s how four developers are leveraging AI agents to speed up coding and solve real-world business challenges.
Timothy Keyes: Streamlining Cancer Patient Data for Clinicians
Cancer treatment often involves tumor boards—specialized teams collaborating to decide on complex cases. Preparing cases for these boards is time-consuming, requiring data from electronic health records, imaging, and medical literature. Timothy Keyes, a data scientist at Stanford Health Care and MD/PhD candidate, highlights the administrative burden clinicians face during preparation.
Stanford is testing Microsoft’s healthcare agent orchestrator, now available in Azure AI Foundry, which enables AI agents to gather and synthesize patient data. These agents build timelines, summarize literature, reference treatment guidelines, identify clinical trials, and generate reports—all while integrating with Microsoft 365 Copilot for seamless use within familiar apps like Teams and Word.
The orchestrator directs specialized agents that handle diverse data types—clinical notes, lab results, imaging scans, and genomic data—providing clinicians with accurate, cited summaries quickly. This approach could cut preparation time by up to 90%, freeing clinicians to focus on patient care.
Keyes emphasizes that AI agents assist clinicians without replacing human oversight. The goal is to reduce time spent on documentation and enable doctors to spend more time with patients.
Xavier Portilla Edo: Accelerating Proofs of Concept with AI Agents
Xavier Portilla Edo, head of cloud infrastructure at Voiceflow, uses AI agents to speed up development of conversational AI solutions without coding. Voiceflow’s platform serves diverse clients, from large brands to local businesses, automating tasks through voice or text interactions.
Portilla leverages GitHub Copilot’s agent mode and Copilot Edits to streamline coding workflows. These tools provide context-aware suggestions and make inline code changes across multiple files using natural language, allowing developers to stay focused and iterate faster.
This agent-driven approach significantly cuts the time to validate proofs of concept, often reducing days of work to just hours. It also helps developers work outside their expertise—back-end engineers can generate front-end interfaces, and front-end developers can quickly create back-end components with agent assistance.
While there is a learning curve in communicating effectively with different AI agents, Portilla notes that GitHub’s agent tools provide a smooth experience overall, enabling faster development cycles.
Amit Sethi: Enhancing Software Development Efficiency at JM Family
JM Family Enterprises, a diversified company in automotive and related sectors, uses AI agents to standardize and accelerate its software development lifecycle. Amit Sethi, principal AI and ML research scientist, explains how their multi-agent system supports business analysts and quality assurance teams.
The AI-powered "BAQA Genie" helps write user stories, generate technical requirements, and design test plans. This has cut business analyst work by 40% and quality assurance test design by 60%, while improving standardization across teams.
JM Family began experimenting with Microsoft’s open source AutoGen tool in early 2024 and now builds agents using Azure AI Foundry services. Specialized agents handle requirements, coding, documentation, and more, coordinated by an orchestrator agent.
The company maintains a human-in-the-loop approach to ensure quality and is planning to commercialize the BAQA Genie to offer these benefits to other enterprises facing similar challenges.
Rob Bos: Living in Agent Mode to Boost Development
Rob Bos, a DevOps consultant and GitHub trainer at Xebia, uses GitHub Copilot’s agent mode extensively in his daily work. He describes the progression from simple chat interactions to edit mode and now agent mode, which focuses on speed and automation.
Agent mode not only applies code changes but also validates them by running scripts and tests, iterating until tasks are complete. This capability extends across software systems, allowing Bos to automate reports, bug analysis, and script generation based on repository data.
Bos highlights how AI agents enable non-developers, such as operations teams, to contribute more effectively to software projects by providing actionable insights and enriching the development lifecycle.
His advice for working with AI agents is to start small and refine instructions through conversation, allowing for gradual course correction. With frequent new features emerging, Bos encourages staying curious and exploring the expanding possibilities AI agents offer.
For those interested in deepening AI skills relevant to development and automation, resources like Complete AI Training’s curated courses provide practical guidance on leveraging AI tools effectively.
Your membership also unlocks:
