How AI Changes the Math for Startups According to a Microsoft VP

Microsoft’s Amanda Silver explains how AI agents will reduce startup costs and enable smaller teams to launch higher-valuation ventures, comparing it to the cloud computing revolution.

How AI Changes the Math for Startups According to a Microsoft VP

Redmond: Microsoft’s Amanda Silver has worked with developers for 24 years, focusing on helping startups build AI tools. She leads Microsoft’s CoreAI division, working on the Foundry system in Azure. This system helps businesses use AI agents and see how they work in real companies.

Silver says AI will change how startups start and grow. It’s like when cloud computing first came, making it cheaper for new companies to begin. Now AI agents can do many jobs faster and cheaper, like helping with support or legal work. This means more startups can start, and they might be worth more money with fewer people running them.

AI agents are already helping with real work at Microsoft. They can update old computer code to newer versions much faster, saving 70% to 80% of the time. Another agent helps watch websites and fix problems without waking people up at night. Before, people had to stay ready to fix things anytime, but now the computer can handle many issues alone.

Some people thought AI agents would come faster, but Silver says there are reasons why it’s taking time. Many companies don’t know exactly what job they want the AI to do. They need to be clear about what success means and what information to give the AI. People who try these systems see that they give back more money than they cost.

Silver thinks humans will still check important AI work sometimes. For example, when returning a package, computers can now look at pictures and decide if it’s damaged most of the time. But for very important things like legal contracts or putting new code into live systems, humans should still check. The goal is to let AI do what it can do well while keeping humans involved where they’re needed most.

Leave a Comment