I like a good user conference. When done well, the format allows a level of thought and discussion that industry events cannot match. The need to cover a broad enough range of topics to attract an audience and to get sponsors, members, etc. on the stage limits how deep a content program can go.
The same restrictions do not apply to user conferences, where a narrower set of interests and decision-makers allows the program to indulge the subject matter. The combination of the right supplier and the right content focus can substantially deepen our understanding of important aspects of our industry. And that is exactly what we got this week at the REBA Insights conference, which wraps today in Vail, Colorado.
A Rapidly-Transforming Capability
At this point, we can take the radical transformation of our technology environment as a given. But as change continues quickly, planning is hard. The challenge of reimagining what work needs to be done is to understand how to improve what is already there.
With that in mind, I ran a workshop designed to provoke discussion about how participants and their organizations are changing their approaches to analytics as AI’s role increases. The gist of the conversation was:
- Individuals and teams are developing bespoke skills using LLMs and a growing number of analytical AI tools. They are increasing personal productivity and creating a need for new guardrails. Agents are also beginning to look like employees (one IT leader noted their function is starting to resemble “HR for AI agents”).
- AI is changing enterprise analytics; the key is to understand how. AI can answer questions that were previously unanswerable due to the depth of analysis required. We still need dashboards and reports that establish reality in our organizations. But AI offers more.
- These capabilities depend on a foundation of data that companies can establish and maintain. Pointing AI at PMS data, with its property-specific irregularities, will not yield reliable analytics. Pointing it at normalized, well-curated data from core systems will unlock increasing advantage as AI grows in power.
In Praise of First Principles Thinking
A recurring theme was the opportunity AI offers to return to first-principles thinking about processes and technology. To AI-enable existing processes may miss the point. We should instead ask what needs to exist in an AI-enabled analytical environment.One of the best examples came from a presentation on the future of Revenue Management (RM) technology by Brad Schell, REBA’s RM product leader. The clarity of thought on how AI is changing RM, and—critically—how it is not, given recent legal challenges, stood out.
The presentation began with a principle that may not yet be obvious to multifamily operators. The legal issues of recent years show that the industry must not allow pricing algorithms to incorporate any nonpublic data from other companies. That creates a problem for AI, which requires training data. Low transaction volumes in multifamily yield too little data to train an AI for something as nuanced and multifactorial as pricing.
How AI Should Affect Pricing
This has implications for how AI is deployed in RM. It is helpful to separate the pricing algorithm from the activities revenue managers use to manage it. AI does not belong in the former, but it is radically changing the latter.RM systems have many settings that allow operators to fine-tune strategies. These require maintenance and constant review as strategies and market conditions change, creating risk that settings drift from strategy. I have worked on many consulting projects to correct this issue for RM users. Such projects will no longer be necessary as AI that understands how settings relate to strategy can critique and propose changes automatically.
This separation was captured in Brad’s presentation as “responsible automation.” The principle is that AI automates the machinery that governs the pricing process, but is never involved in the computation of prices.
This approach has another implication: while automation saves time, it increases the need for subject matter expertise. A second principle, “confident execution,” ensures AI does not overstep its expertise. It can automate decisions that would be too burdensome for a human, but the limits on training data mean it cannot make fail-safe decisions. Expert human review remains essential and should be central to how firms organize RM functions.
A Growing Reflection of the Data Environment
There was much more to this event; too much to cover here. One highlight was an excellent presentation by Michael Tcheau, VP of research and strategy at Waterton, on how the firm uses data to understand markets and identify deals.
The presentation did two things well. First, it showed how a range of economic indicators, from employment numbers to comp rents and (in the case of San Francisco) changes in homelessness trends, helps Waterton understand and anticipate market direction and, hence, opportunity.
Second, it showed how the market data brought into REBA through its acquisition of Markerr fits into enterprise analytics. As this blog has argued, comp data belongs outside pricing algorithms. But understanding market opportunities and projecting the future, for example, in budgeting, are central to the analytical environment in which RM operates.
It is this breadth of analytical vision that sets this event apart from other analytics-focused events in our industry. A deep grounding in strategy, clarity on legal constraints, and a commitment to AI-driven transformation make this the strongest analytics forum in the industry today. Attendees will be heading back down I-70 with a clearer sense of direction on our rapidly-changing analytical environment.
