I recently sat down with Daniel Newman, CEO and principal analyst at Futurum to talk about where enterprise applications are headed next. We covered agents, automation, trust, and governance, but the real takeaway was simpler: the way software gets built and the way it gets used are changing at the same time.
AI agents and business teams together are redefining how modern systems are built
As I told Daniel, we’re seeing a dramatic expansion in what’s possible to build, driven by AI and agents. At the same time, we’re seeing an equally dramatic expansion in who needs to be involved. Companies that want AI to meaningfully change how they operate need the people who are doing the work every day, those who understand finance, human resources (HR), supply chain, and customer service, to be able to shape the systems they rely on. This is an imperative for success.
If more people need to shape systems, the software itself needs to change. Systems designed to require a human to click through every step bottleneck participation. Agents help shift that dynamic. They can monitor, reason, and act—with humans providing intent, oversight, and judgment. Daniel and I talked through practical examples, like refunds and fraud detection in retail where traditionally slow, manual processes can now be rethought. When agents work alongside apps and automation, they can radically accelerate how fast teams build and help them reduce costs. They also give business teams a more direct hand in improving customer experiences.
But my conversation with Daniel wasn’t about handing off everything to AI. Adopting an agent‑first means teams can spend less time wiring together steps and more time defining what should happen, where human input is required, and how outcomes can be governed.
One thing I emphasized to Daniel is that trust and governance are foundational. Systems must be observable, auditable, and adjustable. Agents need boundaries and humans need visibility into what’s happening. Black boxes that fire and forget aren’t the answer. That’s why we’re focused on higher‑level abstraction, not throwing a prompt over the fence and hoping for magic. Great systems start with planning, architecture, data modeling, and policy. We’re building experiences that help people with deep process expertise think like software architects, with AI supporting their work, not replacing human judgment.
Applications stay and they take on a different role, becoming shared spaces for human-and-agent collaboration and oversight. As I said to Daniel, the future interface is about delivering the right information to the right place at the right time.
The broader point Daniel and I kept coming back to is this: the future of enterprise software is not a single killer agent or a clever prompt. It’s a managed environment where intent turns into action—safely, repeatedly, and at scale.
That’s the shift we’re building for. And it’s the shift enterprises need to start planning for now.
Watch my full conversation with Daniel Newman from Futurum to hear us unpack what agent‑first development really means for enterprises.
What’s next for agent-first development
- Turn great ideas into impactful solutions with Microsoft Power Platform
- Learn to transform your business with modern, AI-first applications in Microsoft Power Apps
- Optimize business processes with an end-to-end automation solution built for enterprise with Microsoft Power Automate
