Scaling AI and agents at Microsoft beyond this year

- I'm super excited because I really think this year for Microsoft as well as for our customers, is one of going from putting your toe in the water to really broad adoption and enablement across our employees and transformation of the processes that we drive for our customers. (upbeat music) - Today, working without a computer feels unimaginable. Tomorrow we'll feel the same way about AI, with every employee using it as instinctively as they use laptops, smartphones, or the internet each day. The promise of agents is extremely exciting and we believe that agents are key to realizing ROI on your AI investments. - We're expanding usage, integrating AI deeper into workflows, and increasing agent sophistication. When we think about simplifying complex sales processes, agents will be critical to drive faster deal cycle times for our sellers and quicker case resolution times for our support engineers. - We see three levels of evolution happening. The first is human with assistance, where every person is augmented with an agent that understands their job and helps them get work done on a daily basis. The second is human-led agents, where every team includes AI agents managed by humans that can complete tasks autonomously on behalf of marketers. And level three is human-led, agent-operated work where humans set the direction and AI agents run an entire function. - We're already implementing AI and agents to drive connected experiences, simplify interactions, and maximize the value of HR tech investments. We are bringing innovative tech and HR together to further empower the people who enable Microsoft's mission to empower individuals and organizations on the planet to achieve more. - Tomorrow's legal professional will have intelligence at their fingertips and they'll be working with agents in different modalities and different processes. As an example, we have processes today where a human is driving the process and will leverage agents to solve a particular problem. Research is a great example of that where, in my daily work, I'll ask an agent to research something for me and come back to me. And in patent prosecution, we'll have humans directing a set of agents which are taking responses from the patent office on any rejections, and these agents will be able to drive the responses all the way from recommendations to formalizing the response and submitting the response and involving humans where necessary. - What we've seen so far is that most of the AI agents really require a human deeply embedded in each interaction, and that's important and that helps us make massive progress and we will always have humans in the loop. However, what we're most excited about is turning on more of the autonomous agents in the months and years to come. - Our vision is an autonomous contact center powered by Copilot and AI agents that enable faster resolution and a high percentage of customer issues being addressed entirely through self-service. - Scaling AI and agents in the supply chain involves several strategic initiatives. First, we plan to expand AI capabilities across plan, source, make, and deliver domains, connecting different agents into a unified supply chain intelligence platform. This will improve our ability to manage complex processes efficiently. Next, we focus on enhancing user experience by providing the right tools for different user groups, such as engineers, product managers, and business partners. - The balance between human and virtual agents is evolving towards a collaborative approach, and autonomous agents will change the game across the end-to-end support experience. - By making AI ubiquitous in every seller's workflow, tracking business impact, and iterating quickly, we're transforming our organization at scale and setting a blueprint for our customers. - And just as we've seen the role of AI in software development evolve over the past three years from just coding assistance to chat to now agents, the same pattern will apply to knowledge work.