Microsoft executives at the opening of AI Agent & Copilot Summit defined how the future is evolving to capitalize on the power of agentic technology, specifically detailing transformation opportunities and steps toward fully autonomous operations, as well as functionality for building agents and leveraging AI in core Microsoft systems.
At the same time, other top-level executive speakers at the inaugural agent and copilot event laid out guidelines for setting priorities across organizations — from employees to boards of directors — to fully capitalize on AI.
Tying together the themes of agentic transformation and how companies need to manage people in an agentic AI world — as well as the previous night’s keynote on careers in a post-AI world —Microsoft Vice President, AI Agents Ray Smith detailed a “future of work” in which employees will be measured and evaluated at least in part on their agent acumen.
“The future of work may not necessarily be what are Ray’s skills, but what agents can Ray build, manage, deploy, and that’s going to be a representation of my value that I can bring to a company, or what I could bring to the market,” Smith said.
Transformation Opportunities
In his opening keynote Tuesday, Smith detailed three types of transformation that agentic AI unlocks, spanning general employee productivity, process improvement measured in growth and cost reduction, and new products and experiences, while noting that transformation starts with a focus on outcomes, such as improving a process, growing revenue, or saving/eliminating costs.
In any/all of those scenarios, he emphasized how the speed at which AI is advancing necessitates moving quickly to build, test, and iterate, rather than executing long-running transformation programs such as those of the past.
Smith also laid out four levels of agentic transformation that are beginning to play out, progressing from human-focused where every person is augmented by AI that understands their job, all the way through to departments or agents operating autonomously without regular oversight. (See graphic below).
At level 3 in that progression (agents first), there’s a shift whereby humans “manage a pool of agents to get work done and handle exceptions,” Smith said, noting that human contributions will come to be defined by people’s skills at managing agents. “Everyone in this room will be an agent builder,” he said.
Once things advance to level 4, with autonomous agents only, agents perform basic policy checks and reasoning, resulting in “a big ROI unlock,” Smith said. Such agents will have “autonomy to trigger a response, learn, improve,” he added. “That’s where we get to autonomous agent capabilities.”
Companies moving down the path toward autonomous agents will typically start with common processes including IT and HR help desks, but will then look to move up the food chain to business development, supply chain cost reduction, and more.
Smith urged customers to view agents as the new apps, and as reasoning tools that are core to unlocking business transformation while following human guardrails and asking for help while orchestrating other agents.
Tech Stacks Delivering on AI Functionality
Smith and Microsoft colleagues detailed the technology stack and platforms that Microsoft is building out to enable customers to progress to the level of autonomous agents.
When it comes to developing agents, the company offers three core technology platforms:
- M365 Copilot, a no-code tool for “productivity AI”
- Copilot Studio, which falls between no-code and pro code, for “process AI”
- Azure AI Foundry and Github, procode AI platforms for developers
Smith referred to these three platforms as a “transformation stack with tooling aligned to skills” and said customers can adopt each layer individually if they choose to do so, or work with them jointly from a “better together” perspective.
There was an additional element of the Microsoft technology stack and its AI positioning presented by Cecilia Flombaum, Office of the CTO – Business & Industry Copilots.
Flombaum detailed the new Microsoft application and platform stack bringing together people, copilots, and agents, with UIs evolving to copilots, workflows evolving into agents, and data evolving to knowledge. Her discussion highlighted Dynamics 365 managing business data and processes (see chart) while Microsoft 365 manages productivity and collaboration and Power Platform provides low-code solutions.
Setting AI Priorities and Organizational Factors
Beyond the agent technology opportunities and impact detailed by Microsoft, the event also featured extensive discussion about organizational considerations that are so key to AI success. Kirstie Tiernan, BDO board of directors member (pictured above), shared recommendations and insights on how to set organizational priorities for AI and how to do so while engaging everyone from board members to staff.
She discussed the ways that AI raises the bar — in terms of competitive and performance stakes for employees and employers.
Tiernan cited examples and data points indicating employees are using AI tools such as ChatGPT whether they’re allowed by policy to do so or not. For instance, a group of individuals at a meeting she attended were carrying two laptops: one corporate-issue system as well as a personal laptop that they were bringing into the office to skirt a policy against use of ChatGPT on corporate systems. The takeaway: Employees are using AI tools even if policy forbids it.
Taking such scenarios further, she noted that employees will be far less likely — if not outright unwilling — to even interview with employers that implicitly or explicitly restrict use of AI tools.
Further, Tiernan shared data indicating that employees’ use of AI tools — regardless of what policies say — spans all age groups, not just the youngest Gen Z workers as we might be inclined to assume.
Tiernan went on to detail important ways that employees can prepare themselves for the AI world. She noted the importance of developing and refining prompting skills, for example. “If you’re going to be around for the next five years or so, you’d better be developing prompting skills,” she said.
She said another key factor in preparing people for the changes wrought by AI is leaning on power users to successfully executing organizational change management. They are the savviest users and companies should rely on their leadership in this area.
Finally, Tiernan gave visibility into the factors that boards consider when evaluating AI projects and how to lay out plans effectively to obtain board-level buy-in. These include setting a foundation built on education and awareness as well as a cross-functional AI team. A second core factor is strategic considerations including opportunity and risk assessment as well as AI policies and governance. When it comes to execution, boards will scrutinize implementation oversight, ethical AI advocacy, and conduct resource assessment of how projects are being carried out.
Ask Cloud Wars AI Agent about this analysis
