Prompting is how you talk to Giraffe AI: the words and instructions you type to get it to do what you want. Because it's built on a large language model, the clearer and more specific your prompt, the better it can understand your intent and pick the right action, rather than guessing.
Once your API keys and settings are configured, you can chat with Giraffe AI just like you would with any LLM. Ask about your market, find comps, search for parcels, and more! Once you identify a parcel for study, use Giraffe AI to analyze yield, identify zoning constraints and lot restrictions, and even run a simple proforma.
Start from scratch
Click "New Chat" in the top bar to wipe the existing chat and start fresh.
Type in the text box, then click "send."
You can also upload attachments, like text, pdfs, or images to give more context to your request. Click the Paper Clip icon to add an attachment.
Remember: LLMs can make mistakes. Double check the results.
Prompting tips
Be specific about what you want changed. "Make the office 10 levels" is better than "make it taller".
Use system language. For example, Say "usage type" explicitly when you want to change project-wide defaults, not individual buildings.
Say "assumptions" or "usage type" for project-wide changes. "Update the office assumptions" changes the definition for every Office building — not just one.
Say "this" or "the selected" to target whatever you've clicked on the map. Without a selection, the AI will ask you to pick something first.
Name a place and it searches directly — no need to ask it to navigate there first. Naming the place is enough.
Say "add to the map" or "create a layer" if you want to see results, not just read a summary.
Distinguish "chart" from "analytics." A chart or quick number triggers a fast, one-off aggregation. "Analytics," "measures," or "reporting" triggers the persistent Analytics panel.
Be explicit when you say "delete a usage." It could mean deleting the usage type definition or deleting all buildings with that usage — spell out which.
Use real addresses or region names for parcel searches, not vague descriptions like "that lot from earlier."
Ask for "top
N" when you want a shortlist instead of a full dataset.Expect a clarifying question on ambiguous requests — the AI will ask rather than guess when a phrase could mean two different things.
Example Prompts & Use Cases
1. Parcel Search (US & Australia)
Giraffe AI goes straight to the parcel API when you name a place — no need to fly there first.
"Find multifamily parcels over 10 acres in Dallas County."
"Show me the top 5 largest industrial lots near this address, add them to the map."
"Search for R3-zoned parcels within 500m of downtown Sydney."
"Which of these parcels are in a high flood risk zone?"
US queries route through Regrid; Australian queries (Sydney, NSW, Melbourne, VIC, etc.) route through the national parcels database. Both can drop results straight onto the map as a styled layer.
2. Building & Usage Changes
Giraffe AI distinguishes between editing a specific building and changing a usage type's underlying assumptions — the wording you use determines which one happens.
"Make the office building 15 levels." → edits that one building
"Change the residential assumptions so hard cost is $350/sqm." → edits the usage type definition, affecting every Residential building
"Change this building to Retail." → edits the selected feature
"Set all buildings over 5 levels to 10 levels." → batch edit by filter
Say "assumptions" or "usage type" explicitly when you want a project-wide default changed rather than one building.
3. Site & Layer Geometry
Turn your site boundary or layer data into editable buildings.
"Turn my site boundary into a 20-level office building."
"Take the largest parcel from my search results and build a 10-story tower on it."
4. Quick Aggregations & Charts
For fast, ad-hoc numbers without setting up a formal analytics tab.
"What's my total GFA by usage?"
"Show me a bar chart of hard cost by usage type."
"How many buildings are over 5 levels?"
"Give me everything about the residential buildings."
5. Analytics (Persistent Reporting)
For structured, ongoing reporting that lives in the Analytics panel rather than a one-off answer.
"What analytics do I have set up?"
"Add a cost line item for hard costs on office buildings."
"Create a new tab for the sale option scenario."
"Delete the contingency measure."
Say "analytics," "measures," or "reporting" to route here instead of the quick-aggregation skill above.
6. Layer Management
Manage what's visible and how it's styled on the map.
"Hide everything except zoning and parcels."
"Highlight all parcels over 5 acres in red."
"Style the industrial layer by zoning type."
7. Content Library & Flows
Reuse existing templates instead of building from scratch.
"Find an apartment flow in the content library."
"Show me what's inside the podium layout pack before I add it."
"Attach the Basic Envelope flow to my selected site."
8. Custom Data as Map Layers
Bring your own spatial data in.
"Here's a GeoJSON of flood zones — add it as a red, semi-transparent layer."
"Turn these search results into a 3D extruded layer by height."
9. Market Sentiment & Trends
Use Giraffe AI to get a quick read on how a sector or market is trending before committing analyst time to deeper research.
"What's investor sentiment on data centers in Houston right now?"
"How is the multifamily market performing in Austin?"
"Is there appetite for industrial development in the Inland Empire?"
"What's driving office-to-residential conversions in downtown Chicago?"
These lean on general knowledge and web search rather than Giraffe's structured tools, so treat the answer as a starting point for research, not a data source to cite directly.
10. Development & Site Context
Ask about what's happening in a specific neighborhood or corridor before narrowing in on individual parcels.
"Tell me about developments around Center City Philly."
"What major projects are underway in Nashville's Gulch neighborhood?"
"What's being built near the new transit line in Charlotte?"
"What do I need to know before developing near the waterfront in Tampa?"
Good as a first step before a parcel search — once you've got a sense of the area, follow up with something like "find parcels zoned mixed-use in that area" to move from context into structured data.