Complete AI Training
Sign inGet my AI kit

Your job's AI kit

Get your AI kit

Tell us who you are and what you do. We show you your kit right away and email you the link: skills, prompts, AI agents, MCP servers and courses for your job.

500+ jobs ready, and we make a kit for any other job. No payment needed to look.

Share

AI agent for environmental engineers

Air Quality Alert Forecast Agent

Timely draft air quality alerts based on agreement between models, monitors and weather

Air Quality Alert Forecast Agent: what goes in, what the agent does and what you get

What it does

Smoke and ozone alerts are often issued late because the forecaster must read several sources at once. This agent combines smoke model output, monitor readings and the weather forecast to estimate the air quality category by area. It compares each estimate to the alert thresholds and updates as new data arrives. It checks whether monitors agree with the model, and falls back to monitor trends if they disagree. When a threshold is expected to be crossed, it drafts an alert for review. The forecaster approves issuing. Edge case: the model forecasts heavy smoke but the nearest monitor is clean, so the agent checks wind and asks for confirmation instead of assuming the model is right.

How it works

Follow the arrows from top to bottom. The orange dashed arrow is the loop: when a check fails, the agent goes back and tries again.

Start and resultWhat it doesA check on its own workWaits for your OKGoes back and retries
Yes, continueYes, continueApprovedNoNo 1 STARTS WHEN Scheduled data update 2 USES A TOOL Read model output and monitor readings 3 USES A TOOL Read the weather forecast 4 DOES Estimate the air quality category by area 5 CHECKS THE RESULT Do model and monitors agree within one category? If not: check wind and plume path, and weight recentmonitors more. Back to step 3. 6 DOES Compare estimates to alert thresholds 7 CHECKS THE RESULT Is a threshold expected to be crossed within thealert window? If not: continue monitoring and update next cycle. Backto step 2. 8 DOES Draft the alert with areas, category and timing 9 YOU APPROVE Forecaster approves issuing the alert 10 DOES Log the forecast and verify later 11 RESULT Alert issued or hold decision recorded
Read the steps as a list
  1. Scheduled data update
  2. Read model output and monitor readings
  3. Read the weather forecast
  4. Estimate the air quality category by area
  5. Do model and monitors agree within one category?If not: check wind and plume path, and weight recent monitors more. Back to step 3.
  6. Compare estimates to alert thresholds
  7. Is a threshold expected to be crossed within the alert window?If not: continue monitoring and update next cycle. Back to step 2.
  8. Draft the alert with areas, category and timing
  9. Forecaster approves issuing the alertThe agent waits here for your OK.
  10. Log the forecast and verify later
  11. Alert issued or hold decision recorded

How it decides

A category is forecast when the model and the latest monitors agree or the weather supports the model. Disagreement lowers confidence.

  • Use the more recent monitor trend when model and monitors disagree
  • Draft an alert when a category crossing is expected within 12 hours
  • Include confidence level in the draft
  • Update at every data cycle during an event

Make it yours

Every agent is a starting point. You choose these settings for your own situation.

  • Thresholds by category
  • Alert lead time (default 12 hours)
  • Areas covered
  • Update frequency

What keeps you in control

It always asks you first

  • Issuing any alert

Hard limits

  • Never issues alerts
  • Never hides low confidence

It stops when

  • Done: alert approved or no alert needed
  • Stop: monitor and model data unavailable

Set it up

We guide you through the set-up, step by step

Members get the full set-up guide for this agent. No technical skills needed: you copy, paste and upload.

10 minto set it up in your AI
5 AIsChatGPT, Claude, Copilot, Gemini, Grok
  • One set of instructions to paste into your AI, with the clicks for ChatGPT, Claude, Microsoft 365 Copilot, Gemini and Grok
  • The agent then walks you through connecting your own data, one source at a time
  • A downloadable copy with the flow chart, the rules and the full guide
Get access to this agent

An example run

What happensAt 6 am the smoke model put the city at the unhealthy category by noon, but the nearest monitor still read moderate. The agreement check failed. The agent checked the wind from the northwest and the plume path, found smoke arriving by 10 am, and drafted an alert with medium confidence. By 9 am monitors rose to unhealthy for sensitive groups. The forecaster approved issuing.

More agents for environmental engineers