Dive Brief:
- The biggest factor contributing to the success of AI pilots within a company is organizational readiness, according to 67% of respondents in Microsoft’s annual Work Trend Index. In practice, that includes a supportive workplace culture, clear rules, encouraging talent practices and manager support, said the report, which was commissioned along with Edelman Data & Intelligence to survey 20,000 knowledge workers.
- Organizational readiness, alone, however, isn’t the solution. Employees also need to be willing to use the technology. The report found that this combination was rare. Only about one in five workers established a high level of individual AI capabilities and organizational readiness, the report found. Others face challenges with AI adoption because their organizational conditions aren’t set up for success or their individual AI practice is low.
- Enterprises that think of AI as a new operating model and address how jobs will change will succeed with the technology, Karim Lakhani, professor of business administration at Harvard Business School, said in an interview featured in the report. “[Tech leaders] have to invest a ton of time and effort to map the dependencies that exist between people and processes, then say, where might AI make a difference?” he said.
Dive Insight:
Companies finding the most success with their AI adoption are redesigning how work gets done and encouraging and modeling AI experimentation with solid infrastructure, the report found.
Half of knowledge workers reported that they are “emergent” in their readiness, meaning their personal AI skills and the health of their organization’s AI readiness are taking shape. The advantage comes when organizations build the conditions for employees to apply what they’ve learned, Laura Hamill, director of research of Microsoft’s AI@Work, thought leadership, said of the report’s findings.
“Leaders have to get aligned with the leadership team,” Hamill said in an interview featured in the report. “It can’t just be something that IT drives or someone on the leadership team drives, it actually needs to be a leadership team imperative.”
Building infrastructure as AI agents become more commonplace is imperative, the report stated. It requires employees to rearchitect their work with review processes and for leaders to redesign processes around outcomes and agent autonomy. IT departments will need to build infrastructure to check agent operations as they scale, and a security team should build layers of trust into AI systems.
The report recommended tech leaders treat agents as managed entities with identities, permissions, enforcements and lifecycle management.
“IT becomes the control plane for agent operations, extending the same rigor already applied to people and applications so that scale does not come at the cost of visibility,” the report said.
Lakhani said that for AI insights inside of an enterprise to scale, the creators of agentic systems need to be the owner of the processes. An automated learning loop, where every interaction with an agent, both positive and negative, is captured and analyzed, will also help inform the training of an agentic system.
The third thing leaders should keep in mind is the view that the systems should continually change, because the AI will keep changing, he said.
“That’s where they’ll see failures and successes and how to adapt them,” he said.
