Proptech
CRE property management AI works best in narrow lanes, not broad sweeps
By Real Estate Wire Staff, . Real Estate Wire.
The gap in commercial real estate property management AI is not the software itself, according to Propmodo. It is the absence of the workflow documentation that automation requires before it can function.
Tony Schollum, Chief Technology Officer at Re-Leased, told Propmodo that the core problem is dependency on people to keep systems current. "Property managers all have a system of record, the problem is getting the information into the system of record depends on people and that creates lots of opportunity for mistakes," he said. A system updated by the same staff doing the underlying work will always lag behind what is actually happening at the property.
The fix, Schollum explained, starts with mapping every stakeholder connected to an asset before any automation is configured. "You have to first create the hierarchy for each property, where every person that is affected by the property mapped out," he said. "Anything that occurs in the real world, you have to understand how it needs to push information around that schema." That means owners, asset managers, property managers, tenants, contractors, and accountants, all of them, before a single workflow is handed to software.
The example Propmodo uses to illustrate the point is a leaking pipe. Software has no instinct to fall back on.
Once automations are running, Schollum said the audit approach matters as much as the audit itself. "We can trust agents to do everything but the key is to audit it, particularly at first," he said. "The best practice is to audit the outcome, not each data point." Checking whether a work order was created, assigned correctly, priced within tolerance, and closed with documentation tells an operator whether the process ran correctly without requiring a field-by-field review. Schollum said early auditing should be intensive and then taper as confidence builds.
On agent architecture, Schollum pushed back against the instinct to build one capable agent covering an entire workflow end to end. The structural reason is how large language models handle memory: they have none between turns, so the full conversation history is resent with every new interaction. As a workflow grows longer, so does the context window, and cost and latency grow with it. Vertical software can store state externally to reduce this, but the problem compounds when a single agent is asked to manage too many steps.
"It is important to keep the scope of work for an agent small," Schollum said. "You can ask one agent to do all of the steps in a workflow automation but that gives the agent too much agency. You might not get the same outcome every time."
The alternative Schollum described is decomposition: breaking a workflow into discrete objectives and assigning one agent to each. "Once an agent has a single goal it uses fewer tokens, provides more consistent outcomes, and reduces hallucinations," he said. "If you completely automate the entire property management process the token will be astronomical," Schollum said. "There is just too much context for every step of the way."
Agents built this way can operate across platforms rather than within one.
The practical implication for property management teams is that the preparatory work, mapping stakeholder relationships and decomposing familiar processes into their actual components, is the harder part of AI adoption, and the part most operators have not yet done. The technology is available. The workflow documentation, in most cases, is not. Teams that invest in that mapping first are likely to get more consistent results from automation than those who deploy tools and adjust afterward.
Source: https://propmodo.com/your-cre-software-knows-whats-happening-why-isnt-it-doing-more/
Real Estate Wire provides news and information, not real estate, legal, or financial advice. Consult a qualified professional about your specific needs.