Over the past few years, artificial intelligence has dominated economic and business attention. However, the hype cycle is slowing as companies face the challenges of enterprise AI adoption and the need to drive tangible business value.
In the MIT Sloan Management Review article and related video about AI predictions for 2026, Thomas Davenport and Randy Bean say they expect a stable year when it comes to AI.
Davenport, a Babson Professor and fellow in the MIT Digital Economy Initiative, and Bean, an advisor to Fortune 1000 companies, advise business and technology leaders to focus on the organizational structures, tools, and strategies needed to implement technology at enterprise scale over the next year.
Here are five AI insights and action items for decision makers for 2026.
1. Agentic AI is not ready for prime time yet
Davenport and Bean are dialing back expectations for agent AI, despite predicting its exponential growth last year. Agent AI is a type of AI system that can recognize, reason, and complete tasks independently or with minimal human supervision.
Continued hallucinations and mistakes, along with the ease with which hackers can hijack agent AI systems using prompt injection and other methods, are a wake-up call for slow adoption. Davenport said “companies will continue to involve humans” to build guardrails for agent-based AI, but that would undermine the promised productivity benefits.
Some industry watchers predict it will take more than a decade to resolve these issues, but Davenport and Bean are more bullish, predicting that within five years most transactions in many large business processes will be handled by AI agents.
Action items: Companies should start imagining how AI agents can power new ways of working, starting with use cases that can be reused across the organization. It’s also important to start cultivating internal capabilities to create and test agents.
2. The AI bubble will deflate and there will be economic consequences.
AI has dominated board discussions and driven up stock markets. Mr. Davenport and Mr. Bean expect that liquidation to occur this year, perhaps sooner rather than later. The emphasis on user growth over profits is reminiscent of the dot-com bubble, Davenport and Bean write. “Technology is often overvalued in the short term, but its transformative impact is vastly underestimated in the long term,” Bean said.
Action items: Companies need to make the most of the AI technology they already have, while also considering how their investments will impact their future business strategies.
AI Executive Academy
Face-to-face at MIT Sloan
Register now
3. Generative AI needs to become a business, not a personal resource
Organizations are primarily taking an individual-level approach to generative AI, with employees leveraging the technology to improve their own productivity. It is less common for companies to apply generative AI to enterprise workflows and processes. Until that happens, it will be difficult to aggregate the results and quantify business value.
Action items: To drive value, companies should look beyond personal productivity and consider enterprise generative AI use cases, such as accelerating new product development and enriching customer experiences.
4. The optimal reporting structure for AI has yet to be determined.
Support for the leadership role of data and AI in large enterprises is at an all-time high, but responsibilities for AI and what constitutes the best reporting structure remain unclear. In the 2026 AI and Data Leadership Executive Benchmark Survey, 38% of responding companies said they had appointed a chief AI officer or equivalent role, but there was little consensus on who that role would report to. (Currently divided into business, technology, and transformational leadership.) “Diversity in reporting relationships may be contributing to the broader problem that AI, especially generative AI, is not delivering sufficient business value,” Davenport and Bean write.
Action items: What companies should consider Designate an individual to integrate data, analytics, and AI and report to business leaders. “As AI becomes more well-known, it is being elevated to executive roles,” Bean said. He cited JP Morgan as a prime example. There, a new AI-focused executive is part of a 14-person steering committee that reports to Chairman and CEO Jamie Dimon.
5. “AI Factory” helps organizations accelerate value
Davenport and Bean define an “AI factory” as “a combination of technology platforms, methods, data, and previously developed algorithms that makes building AI systems quick and easy.”
“It’s not just building a big data center and filling it with GPU chips; it’s the capabilities within the organization,” Davenport said.
Instead of relying on data scientists and businesspeople to duplicate work or figure out what data is available, these AI factories establish a foundation of tools and business processes that allow companies to efficiently and cost-effectively build AI at scale.
Action items: Forward-thinking companies should use this year to launch their AI factories, expand the number of internal use cases, and unlock more economic value from their AI investments.
Read the article: “5 AI and Data Science Trends for 2026”
watch video
Thomas H. Davenport He is a professor of information technology and management, chair of the Department of Entrepreneurship at Babson College’s Metropoulos Institute of Technology, and a fellow of the MIT Digital Economy Initiative. His latest book is “The New Science of Customer Relationships: Delivering the 1:1 Promise with AI.”
randy bean He has been an advisor to Fortune 1000 companies on data and AI leadership for over 40 years. He is the author of Fail Fast, Learn Faster: Lessons in Data-Driven Leadership in an Age of Disruption, Big Data, and AI.
