Lesson objective
Triage a landscaping task on consequence if wrong, data class, verifiability and who is qualified to decide, then route it to standard review, enhanced review, redesign or no AI before any tool is opened.
Plain-language explanation
Why this matters now
Level 1 explained how AI tools build answers, where they fail with plants, pests, sites and numbers, and which decisions belong to named, qualified people. Level 2 starts on an ordinary Monday in April. The office has a stack of site-walk notes from the weekend, a property manager asking why the front beds still look bare, a crew leader's voice memo about yellow patches on a customer's lawn, and an approved AI assistant already open on the office laptop.
This programme will not open that week with an industry statistic, because there is no reliable one to give. No verified source records how many landscaping, lawn-care or tree-care companies use AI, and no source measures an accuracy or productivity gain from AI on green-industry work.
The nearest controlled evidence comes from another profession, and it is a warning as much as an encouragement.
Triage is how one person decides, task by task, which side of that boundary a job sits on, before a wrong answer becomes a wrong application, a wrong promise or a wrong cut.
A question to sit with
At 7:30 a.m. an office manager has two items. One is turning Saturday's handwritten site-walk notes into the company's typed visit summary. The other is the crew leader's question: "What should we put down on the Delaney lawn for those yellow patches?" Both could be typed into the same assistant in under a minute. What makes them completely different tasks?
What goes wrong
Choosing the tool before the task. The assistant is open, so the question goes in. Nobody first asks whether the answer should come from AI, from a document, or from a named, qualified person.
Triaging on time pressure. "It saves ten minutes" is a benefit, not a triage answer. Ten minutes saved on a wrong product choice can mean a damaged lawn, an unlawful application and a lost account.
Deciding the data question after the paste. Once a customer's gate code, an HOA's contract terms or a worker's personal details are inside a tool, the decision about whether they belonged there has already been made by default.
Assuming everything can be checked. A typed visit summary can be checked line by line against the handwritten notes. "What should we put down?" cannot be checked against any document the office holds, because it is not a question about a document. It is a pesticide decision.
Treating qualified work as office work. A question about which product, what rate, whether a tree limb is safe, or how to reset an irrigation zone does not become administrative because it arrived in the office inbox or the crew group chat.
Losing track of who owns it. The draft moves from office manager to account manager to crew leader, and each assumes one of the others checked it.
The practical answer
Ask four questions, in order, and write the answers down before any tool opens.
| # | Question | What you are looking for | Typical landscaping answers |
|---|---|---|---|
| 1 | Consequence: if this is wrong and nobody notices, where does it land? | The furthest place the output travels | An internal note; a customer or property manager; a proposal or contract; a crew's work order; a pesticide application; a tree crew's cut |
| 2 | Data class: what would the tool see? | The most sensitive item in the material | Public plant or product information; company records; customer names, addresses and gate codes; HOA or municipal contract terms; worker personal details |
| 3 | Verifiability: can every part be checked against something you hold? | The document, measurement or photograph you would check against | Fully, against supplied notes or a label; partly; only by redoing the work; not at all |
| 4 | Who is qualified to decide? | A named person, not a job title | The account manager; the estimator; the person who signs the proposal; the certified pesticide applicator; the licensed irrigator or arborist where the state licenses that work |
Question 1: follow the output to the end. A "quick note for the crew" is read by a crew leader who acts on it. Rate the task by where an uncaught error finishes, not where it starts. Anything a crew could apply, cut, dig or program from is a field consequence.
Question 2: settle the data class first. Level 1 worked through what a landscaping company holds and where data goes once it enters a tool. At triage you need only the headline: what is the most sensitive thing this task would expose, and has the company approved a system for it?
Question 3: name the thing you would check against. Handwritten notes, the property's measurements, the signed scope, the product label, the photographs. If the honest answer is "nothing I have," the output cannot be verified and the task needs redesign or no AI at all.
Question 4: name the owner. Landscaping has no single licensed professional who owns everything. The owner changes with the output.
| Output | Who owns it |
|---|---|
| A visit summary, status update or complaint reply | The account manager |
| An estimate, takeoff or proposal | The estimator, with the person who signs the proposal |
| Whether to apply a pesticide, and the product, rate, timing, site and conditions | The certified pesticide applicator, or the certified applicator supervising a non-certified one |
| Irrigation design, installation or programming, where the state licenses it | The licensed irrigator |
| Whether a tree or limb is safe, where the state licenses tree work | The licensed tree expert or arborist |
| Anything the company promises a customer or says in an advertisement | The person who signs the contract or approves the advertisement |
Which of these work types your state licenses varies. Check the rules that apply to you with your own state's licensing and pesticide authorities.
Now turn the four answers into a route.
| Route | When it applies | What happens next |
|---|---|---|
| A: proceed, standard review | Internal consequence; data cleared; fully verifiable; owner named | Run the Lesson 2 workflow; the preparer checks every item |
| B: proceed, enhanced review | Output reaches a customer, a proposal, a work order or the field; data cleared; fully verifiable; owner named | Run the workflow; the named owner checks every measurement, date, plant name, price and quoted label line |
| C: redesign, then triage again | Data not cleared, or part of the output checkable only by redoing it | Remove the data, split the checkable part from the judgement, or get the source first |
| D: no AI for the decision | The output would choose a product, rate or timing, judge a tree, design or program irrigation, authorise a dig, or declare a hazard controlled | Route the decision to its qualified owner; AI may later help record or explain what that person decided |
Why write it down. Best practice: the NIST AI Risk Management Framework (AI RMF 1.0, January 2023; voluntary, and being revised) states in Appendix C that human roles in decision making and oversight "need to be clearly defined and differentiated." A one-line triage entry is where a small landscaping company makes that true for one task. Nothing in the framework binds a landscaping company.
Remember this
- Triage before any tool opens: consequence, data class, verifiability, who is qualified to decide.
- Rate consequence by where an uncaught error could land; anything a crew could apply, cut, dig or program from is a field consequence.
- A product, rate, timing, tree-safety, irrigation-design or dig decision is Route D, whatever inbox it arrived in.
- The owner changes with the output, and which work a state licenses varies: check your own state.
- The label is the binding instruction; no model can authorise a departure from it.
Landscaping example
Fernbrook Grounds Care is a fictional lawn and landscape maintenance company with eleven employees and an AI assistant the owner has approved for company records and customer correspondence. On a Monday in April the office manager, Tamsin Oduya, triages four tasks before opening it. Task one: type Saturday's handwritten site-walk notes for the fictional Larkfield Commons HOA into the visit summary template. Consequence: an internal record the account manager will rely on. Data: company records, cleared. Verifiable: fully, against the notes. Owner: the account manager, Rafael Quint. Route A. Task two: a property manager asks for a written update on why the front beds are late. Consequence: a customer-facing statement. Verifiable against the crew schedule, the nursery's delivery email and the signed scope. Owner: Rafael. Route B. Task three: the crew leader, Dmitri Havel, texts a photo of brown patches and asks which fungicide to put down and at what rate. Tamsin routes it D: the photo and notes go to the company's certified applicator, Grace Mbeki, who will decide whether anything is applied and will read the label herself. Task four: the owner wants the assistant to "work out a new zone schedule" for a customer's controller. The company operates in a fictional state that licenses irrigation work, and no licensed irrigator is on staff this week. Route D for the programming decision; Tamsin notes it for the owner. The four triage lines take six minutes and are written before any task starts.
Responsible-use warning
Do not let an approved tool make triage feel unnecessary. Approval means the company cleared a system for certain data and uses. It does not mean every task suits it, that every output can be checked, or that someone qualified owns the result. The two costliest mistakes in this trade are asking a model which product or rate to use, and letting a tree-safety, irrigation or dig question pass as office work because it arrived by text. Triage the task, not the tool, and write the route down first.
Human-review requirement
The person who will own the output confirms the triage before work starts. The account manager confirms it for customer updates and complaint replies, the estimator and the person who signs the proposal for anything priced or promised, and the certified pesticide applicator for anything that touches whether, what, how much, when or where a product is applied. Irrigation and tree-safety decisions go to the licensed person where the state licenses that work. The owner checks that consequence was followed to the field, the data is cleared, a real source exists to check against, and a name is on the task. Anyone may move a task to a stricter route; AI never assigns the route.
Practical exercise
Use a fictional company, fictional properties and fictional customers only.
- List eight tasks from a typical week in your role, such as a visit summary, a customer update, a proposal paragraph, a crew work order, a pest question, a tree question, a controller question and a toolbox talk.
- Replace every real name, address, gate code and price with a fictional one.
- For each task, answer the consequence question by following the output to its final destination, and mark any task a crew could apply, cut, dig or program from.
- Record the most sensitive item each task would expose and whether an approved system covers it.
- Name what you would check each output against: notes, measurements, photographs, the signed scope or the label.
- Mark each task fully, partly, only-by-redoing or not verifiable.
- Name the qualified owner of each output from the owner table, and note which ones depend on what your state licenses.
- Assign a route, A to D.
- For each Route C task, write what change would move it to Route A or B.
- For each Route D task, write which answer ruled AI out and who receives the decision instead.
- Ask a colleague to triage the same eight tasks blind, compare, and note which question caused each disagreement. Record the results in your workbook.
Knowledge check
-
What are the four triage questions, and when are they answered?
Answer guidance: Consequence if wrong, data class, verifiability, and who is qualified to decide. They are answered and written down before any tool is opened.
-
Why is "what should we put down on this lawn?" a Route D task?
Answer guidance: It asks for a pesticide or fertilizer decision. Whether to apply, and the product, rate, timing and site, belong to the certified applicator reading the label, and federal law makes use inconsistent with the labeling unlawful.
-
What did the consultant experiment find on a task outside the AI capability frontier, and how far does it apply to landscaping?
Answer guidance: People using AI were 19% less likely to produce correct solutions than those without it. The subjects were consultants using a 2023 model, so it applies to landscaping only by analogy.
-
Why does this programme not open with a landscaping AI adoption or productivity figure?
Answer guidance: No verified source provides one for landscaping, lawn care or tree care, so none is given.
Sources and further reading
- AI Coalition Network — Certification Educational Standard: Human Review of AI-Assisted Professional Work. Accessed 2026-09-15.
- Fabrizio Dell'Acqua, Edward McFowland III, Ethan Mollick, Hila Lifshitz-Assaf, Katherine C. Kellogg, Saran Rajendran, Lisa Krayer, François Candelon and Karim R. Lakhani; Organization Science — Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of Artificial Intelligence on Knowledge Worker Productivity and Quality (2026-03-01). Accessed 2026-09-15.
- United States Congress (Federal Insecticide, Fungicide, and Rodenticide Act) — 7 U.S.C. 136j — Unlawful acts. Accessed 2026-09-15.
- National Institute of Standards and Technology — Artificial Intelligence Risk Management Framework (AI RMF 1.0), NIST AI 100-1 (2023-01-26). Accessed 2026-09-15.