Plain-language guide

Talking to Waldo

There's no syntax to learn and no command list to memorize. You describe what you want the way you'd describe it to a colleague at the next desk — and if you've never worked with an AI assistant before, the examples below show what that actually sounds like.

The short version

Talk to Waldo in plain English. Name the data you're working with and what you want out of it — you don't need to know which geoprocessing tool does the job, what it's called, or where it lives. That's Waldo's half of the conversation. If your request is ambiguous, Waldo asks rather than guessing.

The one thing worth knowing up front is the Ask / Do toggle at the bottom of the pane. In Ask mode Waldo is read-only, enforced: it answers questions about your project, recommends approaches, and hands you tools ready to open — but changes nothing. In Do mode it does the work: runs the analysis, applies the symbology, builds the layout. Start in Ask if you want a second opinion; flip to Do when you want hands.

That's the whole manual. The rest is examples.

Ask mode — questions and advice

Read-only, free forever. Use it the way you'd use a knowledgeable coworker: quick how-do-I questions, or a method discussion before you commit to an approach.

Ask · everyday question
YouI have a spreadsheet of member addresses. What's the best way to get them on the map as points?
WaldoI took a look at members.xlsx — 1,748 rows with Address, City, and ZIP, but no coordinate columns, so this is a geocoding job (if it had latitude/longitude, plotting would be free and instant). Geocoding spends Esri credits — roughly 70 for this sheet — so I'd confirm the estimate with you before running anything. Here's the tool ready to go when you are: Open Geocode Addresses in Pro.

Waldo read the spreadsheet before answering — the advice is about your data, not a generic how-to.

Ask · analysis question
YouDo neighborhoods with more tree canopy actually stay cooler? I have a canopy layer and a surface-temperature raster — how would I test that relationship with this data?
WaldoGood question for this data, with one wrinkle: your canopy layer is 1 m resolution and the temperature raster is 30 m Landsat, in different projections — so compare neighborhoods, not pixels. I'd summarize both to your neighborhood polygons with Zonal Statistics, then chart canopy against mean temperature and check the correlation. If you want to go further, a regression can control for things like impervious surface. Want me to lay out the exact steps — or flip to Do and I'll run it?

In Ask mode this is all talk, enforced — nothing in your project changes.

Do mode — hand over the work

The same plain language, but now Waldo executes — in your live session, with every action logged and every geoprocessing run leaving a receipt you can inspect or re-run yourself.

Do · everyday task
YouClip the wetlands layer to the county boundary, then symbolize the result by wetland type.
run_geoprocessing_tool · Clip — Wetlands × County_Boundary → wetlands_county
set_symbology · unique values on WETLAND_TYPE — 6 classes
WaldoDone — 4,912 wetlands fall inside the county, now on your map symbolized by type: blues for the open-water and riverine classes, greens for palustrine, so the families read apart at a glance. The Clip receipt above lists the exact parameters, and its link reopens the tool pre-filled if you want to adjust and re-run. Want labels, or a layout for export?

Two steps, one sentence. Every action lands in the log — the receipt is yours whether or not you read it.

Do · a whole project
YouI'm doing a needs assessment for childcare access. I have licensed provider locations with capacity, census block groups, and the county boundary. My goal is a set of maps showing where families with young children are underserved — something I can drop straight into a report. Propose an approach before you start.
WaldoHere's how I'd approach it. Demand: children under 5 per block group from the census data. Supply: licensed capacity at each provider point. Access: a floating-catchment analysis that scores each block group by the capacity reachable within a travel threshold — the standard method for exactly this question. Deliverables: an access-score map, an underserved-areas map, and a summary table, each as a letter-size PDF layout. Two things to settle first: what travel distance is reasonable for childcare in your county — 10 minutes? 15? — and should I include an inset for the city, where block groups get small?

Notice what the prompt didn't include: the method. Data plus goal is enough — Waldo proposes the how, you approve or adjust, and it carries the whole thing through to exported layouts. This is a real pattern: a full needs-assessment of this shape is an afternoon, not a week.

Three habits that help

Lead with the goal

"I need a map the county board can read at a glance" gets a better result than a list of button-presses. Waldo picks better methods when it knows what the output is for — and if you do want a specific tool or method, just name it and Waldo will use it.

Say your standards once

Shop conventions — your projection, your color ramps, "legends never show the basemap" — can be stated once in conversation and Waldo carries them through the session. You don't have to repeat yourself request by request.

Ask why

Waldo explains its choices — why natural breaks instead of quantiles, why that search radius. If something looks off, say so in plain language ("the labels are crowding the coast") and it adjusts. It's a conversation, not a command line.

The best examples are your own projects

Seven days, every capability, no card and no sign-up. Open a project you know well and ask Waldo something about it — that first answer is usually the moment it clicks.

Download the free trial