How AI identification works
One Big Green can analyze a Display photo to recognize plant species, estimate quantities and sizes, and detect planters and relevant characteristics. The results provide a starting point for a structured Field Inventory without having to enter every detail manually.
AI identification can save a substantial amount of time compared with manually entering all plant and planter information, especially when you are working with a Service Site that has little or no existing inventory data. Instead of first building a detailed spreadsheet, you can use photos to help create the inventory and then review the proposed details.
AI identification requires a photo of the Display. One Big Green cannot identify the plants, planters, or other details of a Display without an image showing what is there.
AI identification supports several workflows. You can import photos first to create Displays, then identify them with AI. Alternatively, you can create Displays manually or use an existing inventory, provided the Displays have photos. You do not need to reimport photos already attached to existing Displays.
AI produces results for you to review. Although it does not remove the need to check the inventory, it can dramatically reduce the time required to create it compared with manually entering all the information. Photo quality and the way a Display is photographed can affect recognition accuracy.
Identify Displays with AI
In the Back Office, open the relevant Service’s Display list.
Click “Identify with AI”.
Select the Displays you want the AI to analyze. You can skip Displays that already have detailed, verified information.
Start the identification by clicking “Identify”.
Wait for the analysis to complete, then review the proposed plant and planter details for each Display.
Correct the results where necessary. Add missing items, adjust details, or remove items that the AI detected incorrectly.
When the inventory looks correct, click “Save”.
Taking photos for better recognition
AI can only work with what is visible in the image. These practices can help it recognize plants and planters, and estimate plant sizes, more accurately:
Photograph the Display clearly and make sure the plants and planters you want identified are visible.
Capture the whole Display, including the plants and the planter or planters. If a container is hidden by foliage or furniture, the AI may have difficulty identifying it.
Take the photo from a clear, reasonably straight-on angle. Avoid extreme angles that hide the planter, plant structure, or parts of the installation.
Use good, even lighting. Avoid very dark images, strong backlighting, harsh glare, or heavy shadows over plants or containers.
Keep the image sharp. Hold the camera steady and make sure foliage and containers are in focus rather than blurred.
Move close enough for leaves and containers to be visible, but keep the relevant parts of the Display in frame. For a large or dense installation, consider additional photos if the workflow allows them to be attached to the same Display.
Avoid obstructing the Display with people, bags, tools, furniture, or other objects that hide important details.
For more accurate size estimates, include familiar objects in the photo when possible, such as a chair, table, or phone. These objects provide a point of reference that can help the AI estimate the plant’s size relative to something of a known or recognizable scale.
For mixed arrangements, try to show the different plants clearly rather than photographing only the outer edge of the arrangement.
Use a recent photo that reflects the current Display. A photo taken before plants were replaced or the arrangement changed may produce outdated results.
These recommendations are not guarantees. Even a clear photo may lead to an incorrect species, quantity, size, or planter count, so check the result against the real installation.
Review and correct the results
Treat AI output as a draft of the inventory, not as a confirmed record. Compare proposed details with the photo and what is physically present at the Service Site.
Check that every detected plant and planter actually belongs to the Display.
Check plant and planter quantities, and remove duplicate or incorrectly detected items.
Correct plant names, size estimates, or other details when the suggestions are wrong.
Add any plants or planters that the AI missed.
Save only after you have reviewed the result and made the necessary corrections.
Using AI with existing Displays
AI identification also works with existing Displays that have photos. To enrich an existing Service Site’s Field Inventory, open the Display list, select the Displays that need identification, and follow the same AI identification steps. There is no need to import the same photos again.
You can create Displays manually and attach photos, then use AI to help populate plant and planter details. Choose the workflow that best fits the information you already have: photo import is useful for quickly creating Displays, while manual entry may be more suitable when verified inventory details are available elsewhere.
Practical tips
Start with a small batch so you can get familiar with the results and review process.
Prioritize clear, well-lit photos that show both the plants and their containers. Include a familiar object, such as a chair or table, when practical to help with plant-size estimates.
Skip Displays that already contain complete, verified details unless you specifically want to review them again.
Always check the number of plants and planters. AI is trained to recognize mixed planting, but when a planter contains several individual plants of the same species, it may have difficulty determining how many separate plants there are. The photo does not show the root systems, so the AI may not be able to tell whether visible growth comes from one plant or several.
Keep photos up to date after major replacements or changes to the Display.
Use AI to speed up inventory creation, but review the information before treating it as accurate.