A New Stage in the Hype Cycle on AI 

August 4, 2026 by Donald Davidoff

All tech, even (maybe especially) something as bold and society-changing as AI follows the Gartner Hype Cycle. And the most recent BMOC, Claude, is no exception to being over-hyped. 

To be clear, it can feel like it’s hard to under-hype the impact of Claude, particularly Claude Code. I blogged back in March about how Feb 5, 2026, will likely go down in history as an apocryphal date that delineates a major change in world history, positing that henceforth we will refer to dates before Feb 5 will be referred to as “BC” and those following it as “AC”—“before Claude” and “after Claude.” [And no, the presence of an M-dash is not evidence that AI wrote this. This article has been written and edited with no input from AI—I just like m-dashes! 😊] 

As CEO of a SaaS software company, I’m particularly interested in the opportunities that AI presents and also the threats it may pose to our business. The proliferation of the notion DIY AI and BI could be the latter (more on how it could be the former later on). 

At REBA, we’ve studied this as deeply as one can study a rather nascent technology; and we came to the conclusion that DIY AI, simply allowing operators and asset managers to turn Claude loose on systems of record will not likely drive the desired results. In fact, all of the Gartner, McKinsey and HBR articles we’ve read about what separates successful AI projects from unsuccessful ones includes the “not sexy” but true observation that the successful ones are all bit on top of the grungy, but critical, work of cleansing and governing the data first. 

So, it was with great interest that I recently heard the first two comments from senior tech executives in the industry share the first observations indicating we may be starting to move past the peak of inflated expectations and towards a more balanced understanding of the capabilities and limitations of this technology. 

The first comment comes courtesy of one the most experienced and thoughtful tech leaders in our industry, Kristy Simonette of Camden Properties. On a  Steering Committee call planning content for RETTC’s upcoming OpTech conference, the subject naturally turned to AI and the growing ubiquity of its applications in our businesses. Someone talked about how it could be used in virtually every facet of the business, when Kristy blurted out, “Yeah, until we found out the cost.” She went on to explain the range of use-cases Camden had experimented with only to find that, in some cases, the “old school” way worked better and was less expensive than the AI solution. Camden has graduated to the point where they are being more thoughtful about where, and when, AI is appropriate—and more importantly, where not! 

In that same call, another executive shared how they had gone “all in” on Claude. He gave the example of letting their asset managers use Clause against their PMS and other systems of record only to find that there were now multiple reports and dashboards with differing numbers for the same metric…and few matching the governed metrics in their BI system. After spending much time and treasure to get away from “Excel Hell” into a cleansed and governed BI platform, they essentially found themselves in “AI Hell.” 

AI is a force multiplier, but it doesn’t change the fundamentals. As those premier publications reported in their study, if you don’t have cleansed and governed data, you won’t have good AI results. 

That’s the basis of REBA’s tagline which doubles as advice: Buy the Foundation. Build Your Advantage.^SM 

What new technologies like Claude Code and Microsoft Fabric have done is changed the “or” to and “and.” Instead of choosing between buying a solution OR building it, we can now buy AND build.  

We can stop forcing our limited IT and data science resources to build and maintain the plumbing thus allowing them to focus on their best and highest use of leveraging a data platform to do the predictive analytics and build adjacent to, and on top of, a robust foundation that they purchased. There’s no value in building our occupancy, rent growth, lead, lease, survey results and other metrics when someone else has already done that. Leverage their work to more quickly build custom analytic and combine that foundation with other proprietary data to “build your advantage.” 

The result will be better capabilities in much quicker time!