AI can sketch the outline, but the data that actually matters is locked behind paywalls.

The structure is identical each time: same metrics, same asset classes, same three cities. This is a templated research task that could be run quarterly with minimal reformulation, which is favorable for automation.

The requested outputs (asking rents by class, vacancy trends, sector demand, growth rates) are reasonably well-defined, but 'size the opportunity' introduces a judgment layer — the agent must decide what counts as sufficient evidence and how to weight conflicting data points across sources.

CoStar requires a paid subscription with authenticated access; CBRE and JLL publish selective public summaries but gate their full market reports. Municipal economic-development data is public but inconsistently structured. An agent without pre-provisioned credentials to these platforms will hit walls on the most critical inputs.

This research directly informs a capital-allocation decision about opening satellite offices. Stale, hallucinated, or cherry-picked rent figures could lead to a bad market entry or missed opportunity — both costly and not easily reversible once leases are signed.

Interpreting whether a market's vacancy trend signals opportunity or oversupply requires contextual judgment a broker brings from deal experience. However, the data-gathering and synthesis layer is largely mechanical if the right sources are accessible.