Constellation Research analysts were split on whether we're in an AI bubble or not, recapped 2025 in AI agents and gave a hint of what's to come in 2026 around decisions, automation and exponential efficiency.
Here's a look at what Constellation Research analysts said about 2025 and 2026 at the opening panel of Connected Enterprise 2025.
Michael Ni
2025: The great divide between enterprises is not about digital. "It's actually a decisional divide and it will decide winners and losers," said Ni. "Winners and losers will really be about who can actually get to decision velocity while keeping compliance control."
2026: "You'll see a massive shift toward a decision centric architecture. Unified data foundations will need to extend to context and grounding. You'll see process and workflow automation incorporate decisions and governance."
Chirag Mehta
2025: "The biggest adoption barrier for AI is security, control and trust. It's not a lack of use cases. It's not a lack of money," said Mehta. "Enterprises want control. They want a specific outcome system and that's driving a lot of innovation and spend in cybersecurity, which was traditionally considered a cost center. Now people are looking at ROI for security and it's amazing."
2026: "Right now people are building AI agents, but they have no idea how to manage a life cycle of an agent, access and privileges. The innovation will be in AI agent identities."
Me
2025: "Agentic AI is the obvious trend and everyone and their mom wants to be a platform. And this game of musical chairs isn't going to go well."
I also noted that we need to start segmenting the AI market. AI infrastructure is bubblelike. Enterprise AI hasn't really started yet. It's time for nuance.
2026: "We're going to realize later in 2026 that all we did was scale mediocrity with the same LLMs and data. We're going to need to come up with ways to be creative and innovative."
Liz Miller
2025: "The challenge for marketing tends to be that you're always being sold a silver bullet," said Miller. "We've seen so many bubbles. The real trend is that marketing is actually using AI. We're really on the front line, but there's a gap between the expectation and the delivery."
2026: Marketing departments will start hiring people because they'll realize the limitations of agents.
Esteban Kolsky
2025: "The biggest trend right now is the commoditization of subsidized AI."
2026: "In the second half of 2026, we're going to start investing in what really matters. Private platforms with a distributed computing architecture.
Martin Schneider
2025: "Revenue ops will optimize using AI but there will be smaller models that are precise and can get better understanding of workflows," said Schneider.
2026: Revenue ops will continue to be revamped via AI agents.
Holger Mueller
2025: "There is a broad revival as everything is going agent with HCM." the disappearance of the divisional / Departmental HR people.
2026: "There will be frontline worker empowerment to complement the agents that are being built," said Mueller. He also said that AI agents will begin to replace divisional and departmental HR people.
Ray Wang
2025: "There's a romantic notion that there will be an agent per persona. We're going to tell you that it's false. You're seeing automation push across the back office."
"We're seeing a manufacturing renaissance. It's supply chain, precision manufacturing, data centers, distribution and energy."
2026: "One of the biggest things we'll be talking about is the notion of exponential efficiency. You're seeing revenue per employee going from $100,000 to $1 million to $10 million," said Wang. "We're going to see the difference between winners and losers based on who adopted AI."
