# Real estate

Organize lead emails, property information and visual briefs with checks for accuracy and missing details.

## Separate lead handling from property presentation

A property business can use AI for different tasks: understanding incoming enquiries, organizing property information or preparing visual concepts. Keep those workstreams separate because their evidence and review requirements differ.

## Process incoming leads

The [lead classification workflow](https://seedrouter.ai/solutions/lead-processing) keeps source messages, extracted fields and uncertain classifications connected. Define what qualifies as a lead and review rejected messages as well as accepted ones. Short or forwarded enquiries should not disappear simply because they resemble unwanted mail.

Use your existing email or CRM permissions to read and update records. SeedRouter does not provide access to MLS systems or other property databases.

## Prepare property information

Preserve the source for measurements, amenities, availability and other factual fields. Ask a model to organize or draft from supplied facts, while leaving missing values unresolved. Review the final wording against the authoritative record.

For virtual staging or presentation concepts, keep original photographs available and distinguish generated changes from the actual property. Review structural elements, windows, views and fixed fixtures before using an image. The [image editing guide](https://seedrouter.ai/blog/gpt-image-2-prompting) provides a starting point for describing controlled changes.

## Connect and evaluate

An [n8n integration](https://seedrouter.ai/integrations/n8n) can coordinate message processing and review. [Image models](https://seedrouter.ai/image-generation) can support visual concepts according to their documented inputs. Your application owns records, approvals and publication.

Measure lead-classification mistakes and image acceptance separately. Compare [current API costs](https://seedrouter.ai/pricing) with actual review work, rather than treating every property-related task as one interchangeable automation.

## Related resources

- https://seedrouter.ai/blog/gpt-image-2-prompting