AI is resetting the baseline for commercial real estate operators, and the laggards are already behind

Regular AI use is now the expected baseline at a growing number of commercial real estate firms. That is the assessment of Adventures in CRE, whose summer 2026 tool guide tracks the models and platforms CRE professionals are actually deploying, updated quarterly as the market shifts. The shift from competitive advantage to table stakes has happened faster than most operators anticipated, and the gap between early adopters and the rest is widening.

Jack Mullen’s June Forbes Finance Council opinion discusses AI companies’ demand for property, rather than faster deal analysis.

From nice-to-have to non-negotiable

Adventures in CRE co-founder Spencer Burton, who also co-founded CRE Agents, a company building AI agents specifically for the real estate industry, describes the shift in direct terms: using AI reflexively, integrating it into daily tasks rather than treating it as a separate step, is what separates top performers today. The firm also runs AI.Edge, which it describes as the industry's leading AI training community for CRE professionals.

The implication for acquisition teams, asset managers, and development groups is concrete. AI tools can now process the volume of financial data that previously required significant analyst time, automate repetitive underwriting and reporting tasks, and surface insights that inform hold/sell decisions, budget variances, and lease comparables. The workflow gains are not theoretical; they are already showing up in how leading firms structure their teams.

He argues that AI concentrates property demand in selected markets, rather than driving a broad rebound.

The model landscape CRE teams are actually working with

Adventures in CRE's summer 2026 leaderboard, which pulls performance data from the Artificial Analysis API and pricing from OpenRouter, gives CRE operators a clear read on which large language models are worth evaluating. The scoring methodology uses the Artificial Analysis Intelligence Index, a composite drawn from benchmarks including MMLU-Pro, GPQA Diamond, MATH-500, and several long-context reasoning tests.

As of July 2026, Anthropic's Claude Opus 5 in Adaptive Reasoning Max Effort mode leads the field with an Intelligence Index score of 60.7 and a token output price of $25 per million, according to Adventures in CRE. The same model's High Effort variant scores 58.9 at the same price point, giving teams a cost-equivalent option with a modest capability trade-off. OpenAI's GPT-5.6 Sol in max mode sits at 58.9 as well, priced at $30 out per million tokens.

For speed-sensitive workflows, such as rapid comparable analysis or tenant communication automation, Google's Gemini 3.5 Flash rates as the fastest option on the leaderboard at 288 tokens per second, with output priced at $9 per million tokens, per Adventures in CRE's tracking data. That pricing gap matters for teams running high-volume, lower-stakes inference tasks versus deep underwriting work where accuracy is paramount.

Where CRE operators should focus adoption efforts

Adventures in CRE identifies five core CRE functions where AI is generating measurable workflow impact: acquisitions, development, asset management, investor relations, and brokerage. The common thread is data volume and repetition. Each function generates large amounts of structured and unstructured data, and each has historically relied on analyst time to process, summarize, and act on it.

Purpose-built CRE AI agents, distinct from general-purpose LLMs, are beginning to address the domain-specific requirements of the industry. Burton's CRE Agents is one example of that emerging category, building tools designed around CRE-specific data types, workflows, and decision contexts rather than adapting a generic model to a specialized use case.

That property-demand argument does not establish which underwriting applications deliver the fastest returns.

What to watch in the second half of 2026

Adventures in CRE commits to updating its tool directory at least quarterly, a cadence that reflects how quickly the underlying model landscape is shifting. Between January and July 2026 alone, both Anthropic and OpenAI released major new model versions with materially improved reasoning performance. CRE leaders who locked in their AI stack decisions a year ago may already be working with tools that have been outperformed.

The practical implication for procurement and technology teams: AI evaluation in CRE is not a one-time vendor selection exercise. It is closer to a continuous monitoring function, with model capability and pricing both moving fast enough to warrant reassessment on the same quarterly cycle that Adventures in CRE uses for its own tracking. Firms that build that review cadence into their operations will maintain the workflow edge. Those that treat AI as a set-and-forget deployment will not.