I am not a skilled gardener, but we have a large plot of land that requires ongoing maintenance. I enjoy physical work outside, yet I often hesitate because I do not know where to start.

Instead of pruning randomly or avoiding it altogether, I decided to treat each plant as an experiment. The goal is not just to improve the garden, but to improve my ability to make correct pruning decisions.

For this experiment, I selected five overgrown shrubs and set a clear objective: bring them back to a healthy, well-structured standard using guided input from ChatGPT.

Hypothesis

Given 3 to 5 structured photos and very minimal context, ChatGPT can function as a real time pruning advisor by identifying dead or crossing branches, recommending cut height ranges, and suggesting structural thinning, which should reduce decision friction and result in a visibly healthier structure within one growing season.

Setup Constraints

  • ChatGPT used via the app
  • Physical execution required in backyard
  • Clarifying questions allowed? yes/no
  • Follow-up iterations allowed? yes/no
  • Manual pruning decisions allowed? yes/no

Execution

  • Selected five overgrown shrubs as test subjects.
  • Captured 2 to 3 structured photos per plant from just the full structure
  • Provided seasonal context, location, height, and pruning goals.
  • Requested specific branch-level guidance and cut height ranges.
  • Applied pruning instructions in real time.
  • Uploaded follow-up photos where uncertainty remained.
  • Adjusted cuts based on iterative feedback.

Outcome

  • Five shrubs structurally reduced and reshaped.
  • Clearer canopy structure and improved airflow.
  • Reduced hesitation before aggressive cuts.
  • Repeatable workflow established for future pruning tasks.

Lessons for Applied AI Usage

  • Structured visual input improves decision quality.
  • Explicit goals produce more precise recommendations.
  • AI is effective for bounded physical systems with visible constraints.
  • Iteration reduces uncertainty more than a single static answer.

Next Iteration

Track regrowth quantitatively over one season and compare AI-guided pruning against unguided control shrubs.