AI and automation in commercial real estate have moved beyond theoretical interest. Firms are using productivity assistants, exploring AI capabilities within Yardi and MRI Software, and implementing focused solutions for invoice processing, bank reconciliation, leasing, and document analysis.
The challenge now is turning those individual experiments into dependable business results, without introducing unnecessary risk, cost, or complexity.
During the webinar From Survey to Strategy: AI & Automation Insights for Real Estate, PredictAP founder and CEO David Stifter joined Neal Cousino, REdirect Consulting's Manager of AI & Automation Services, to examine recent survey findings and discuss what they are seeing across the industry. One principle shaped the conversation: start with the business problem, not the technology.
Most CRE firms remain in the exploration or limited-implementation stage. Adoption generally progresses through several levels:
Portfolios, operating models, and ERP configurations vary widely. AI can adapt to some of that complexity, but it cannot replace sound processes, reliable data, effective controls, or human judgment.
A mature approach is not defined by the number of AI tools a company owns. It is defined by whether the organization can solve valuable problems safely and sustain the results.
Establish how AI will support the broader business strategy, then define approved tools, permitted uses, data-security expectations, validation procedures, ownership, and performance accountability.
Guardrails should enable responsible experimentation. Policies that only prohibit activity may push motivated employees toward harder-to-detect shadow IT.
Personal enablement helps employees understand what AI does well, where it fails, and how carefully its output must be reviewed. Departmental champions can then test relevant use cases and demonstrate tangible results.
Evaluate ideas based on:
A cross-functional AI and automation group can review proposals, share lessons, monitor results, and maintain accountability.
A compelling demonstration is not an enterprise-ready solution. A pilot creates value only when it fits the real process, connects to the right systems, respects access controls, manages exceptions, and produces reliable outputs. Otherwise, as Stifter noted during the webinar, it remains a science project.
The strongest opportunity depends on each organization's bottlenecks, but several areas consistently stand out.
AP is well suited to focused automation because it combines high transaction volume with repetitive work and clearly defined controls. AI-enabled solutions can help capture invoice data, identify the correct property and general ledger coding, detect exceptions, and route invoices for approval.
The objective is not simply faster data entry. A well-designed AP workflow reduces manual touchpoints while improving consistency, visibility, and control. This is also where a specialized solution can be more practical than building and maintaining a custom tool for a common industry process.
AI can support prospect outreach, lead follow-up, and communication triage while escalating conversations that require personal attention.
Large language models can locate clauses and data points within leases, contracts, mortgage statements, and other unstructured documents. Any answer affecting a financial, legal, or operational decision should still be verified against the source.
AI can provide another analytical perspective on investments, portfolio performance, and operating trends. It is particularly useful for spotting patterns, generating questions, and interpreting validated datasets.
Beyond AP, high-volume finance processes offer significant potential, including bank reconciliation, report preparation and delivery, variance analysis, financial-document extraction, and close-related workflow routing.
For example, REdirect's bank reconciliation solution can automate approximately 85% to more than 90% of transaction matching in suitable implementations while retaining human review before finalization. The manual workload falls without removing a key financial control.
Not every process belongs in a large language model. Technology leaders must distinguish between deterministic and nondeterministic work.
Deterministic systems follow defined logic and return the same result from the same inputs. SQL reports, business rules, APIs, and conventional automation are better suited to exact calculations, structured reporting, matching, and repeatable system updates.
LLMs are nondeterministic. They excel at interpreting language, analyzing unstructured information, summarizing, generating options, and surfacing patterns, but their outputs can vary.
For example, a financial report with a governed definition of occupancy should use consistent calculation logic. AI can analyze the validated report, flag patterns, suggest questions, or draft a narrative; it should not replace the calculation itself.
A dependable CRE workflow often combines technologies:
The result is more reliable than forcing AI into every step.
AI can reduce manual effort, but it should not erase controls designed to prevent error and fraud. Segregation of duties remains essential for payments, bank reconciliation, journal activity, and financial approvals.
Agents can make mistakes, and bad actors may exploit rigid rules. Someone could submit repeated invoices just below an automated approval threshold, for example, or target an inactive vendor that technically meets an age requirement.
Fully autonomous payment approval therefore carries a different risk from AI-assisted invoice coding. AI can prepare, classify, flag, and recommend while a qualified person remains responsible for sensitive approvals. Human-in-the-loop design assigns volume and repetition to technology while preserving judgment and accountability.
AI tools can create useful prototypes quickly. Scaling them requires attention to:
Organizations should also distinguish between building with an LLM and building a process that continually runs on an LLM. AI can help developers create a deterministic report or automation without becoming part of every production run. Continuous model use introduces variable outputs, recurring costs, version dependencies, and additional testing requirements.
First, review capabilities already available in the organization's core technology. Many firms underuse functionality within their ERP and existing applications. Before adding another platform, determine whether an existing module or configuration can solve the problem.
If the need remains, consider:
A specialized solution often makes sense for a common process such as invoice processing because its provider tests and improves the technology across many customers. Custom development is more compelling when a workflow is unique, strategically differentiating, or unsupported by the market. A useful principle from the webinar is to buy the core and build the edge.
Outside expertise can help when a company has many ideas but no shared method for evaluating them. A business process review can map workflows, expose friction, and prioritize use cases.
It also adds value when a solution must connect property management systems, specialized applications, reporting environments, files, and approvals. ERP decisions affect AP, budgeting, investment accounting, consolidation, and close, so a cross-functional architectural view matters. Internal teams should remain active participants while specialists provide technical depth and execution capacity.
Three connected areas are likely to attract greater investment.
AI agents will increasingly support routing, triage, follow-ups, and task management. Firms should validate narrow AI assistance before expanding into workflows that span multiple applications or decisions.
As organizations deploy more agents and cross-system automations, they will need to manage connections, permissions, usage, triggers, process status, and audit records. Governance will become an operating capability—not simply a policy document.
Generative AI enables more convincing and scalable fraud attempts. Vendor verification, payment controls, employee training, identity management, and exception monitoring must evolve alongside internal AI adoption.
CRE organizations do not need to automate the enterprise at once. They need a disciplined path from business problem to measurable result.
The firms that gain the most from AI will not be those with the most tools. They will be those that apply the right technology to the right problem, preserve critical controls, and connect innovation to the way the business operates.
To hear the full conversation between PredictAP's David Stifter and REdirect Consulting's Neal Cousino, watch From Survey to Strategy: AI & Automation Insights for Real Estate.
REdirect Consulting helps commercial real estate organizations assess, design, and implement AI, automation, and technology strategies across Yardi, MRI Software, accounting, reporting, and property management workflows.