In a recent fireside chat at the startup community and founder hub South Park Commons in San Francisco, Microsoft Chief Technology Officer Kevin Scott discussed five ways startup founders can leverage the AI revolution and maximize its benefits.
Although his advice targeted the startup community, it is equally applicable to more established organizations that are beginning their journey with AI and seeking to navigate this complex and rapidly changing landscape.
1. Anchor Your Work in Reality and Continuously Gather Feedback
Scott explains that while founders — and this applies to any business leaders embarking on AI adoption — can have the drive and optimism to think big and bold, it’s critical to ground their strategy in the realities of the operating landscape.
It’s important to recognize the technological, economic, and behavioral conditions you are working within and align your efforts to these factors rather than ignoring them. Beyond this, Scott emphasizes the need for constant pressure testing and prioritizing what appears to work over what is merely available, as new products and services are emerging rapidly in the AI Era.
This concept of pressure testing and listening to what is actually working — rather than simply jumping on the latest trends — is incredibly relevant. Even when AI, specifically agentic AI, is installed and continuously monitoring progress in the “always on” operating landscape, business leaders must regularly step back to analyze feedback.
2. Capitalize on the ‘Capacity Overhang,’ But Be Prepared for Sustained Effort
Scott points out that AI systems already possess far more capabilities than most business applications utilize, referring to this phenomenon as “capability overhang.” He believes this situation presents significant opportunities for founders who are willing to invest the time to unlock those capabilities.
“All you need to do to harness the potential of these systems is engage in some challenging, less glamorous work, or perform fundamental product building,” Scott said. He highlights that startups often embody a culture of “grind,” a mindset that can also apply to more established businesses.
While the rapid growth of AI and the significant efficiency and productivity gains these tools offer make it tempting to rely on a plug-and-play approach, dedicating teams to do the discovery work can give a company a competitive edge.
AI Agent & Copilot Summit is an AI-first event to define opportunities, impact, and outcomes with Microsoft Copilot and agents. Building on its 2025 success, the 2026 event takes place March 17-19 in San Diego. Get more details.
3. Open and Closed Approaches Aren’t Rivals
Scott encourages companies to be open to both open-source and proprietary AI models instead of limiting themselves to one approach. He believes that the most effective solutions will combine different tools for various purposes. Essentially, it’s about selecting the right tool for the job.
I would build on this by adding that companies shouldn’t be afraid to experiment with AI tools from different vendors. If there is a specific purpose in mind, it’s important to choose the best tool for the job. As we see more open standards, collaborations, and strategic partnerships emerging, interoperability issues are quickly dissolving.
4. Lean into Experimentation, Taking Advantage of Low Costs
One area that is often overlooked when it comes to AI is the fact that the technology has made it both cost-effective and incredibly fast to experiment. Scott commented on this, noting that previously, many concepts requiring extensive manual input from specialized teams can now be executed quickly and affordably with AI tools and agents.
“Do the experiments. Try things,” said Scott. “I would really encourage folks to not be precious about the possibility of failure.” Microsoft has a range of sandbox testing environments where organizations can truly experiment with optimizing AI agents, applications, and processes. These environments are secure and increasingly low-code to avoid isolating less technical users.
5. Stay Focused on Empowering People and Building Real Value
Scott echoed Microsoft’s strong stance on human-first AI, emphasizing that AI developments must always focus on how they empower human users. He stated that even as AI capabilities continue to accelerate, users should always ask whether their work is making someone’s life or work meaningfully better.
Scott emphasized that the focus should always be on choosing a problem worth solving. I would echo that sentiment and expand on it. From a purely business perspective, this is also incredibly relevant. You know your customers, your products, and your areas of expertise. Just because AI technologies enable businesses to think outside the box, even beyond the stratosphere, it doesn’t mean they have to. Instead, tailoring AI agents to improve and expand upon existing successful use cases is a much more effective approach.
Ask Cloud Wars AI Agent about this analysis
