Case Study

Keeping the Garden Alive

Drip irrigation design and procurement, two-instance AI collaboration.

Problem

An English homeowner with a small courtyard garden in southern Spain needed a reliable automatic watering system. She leaves the property unoccupied for weeks at a time in summer. Without intervention, the plants die. The brief was simple: bought once, installed once, runs while she’s away.

The complication was that her brother was handling the project remotely from a different part of Spain, to spare her the language barrier and technical coordination. He had the sketch and the photos. He did not have the specialist knowledge.

What happened

He uploaded a hand-drawn sketch with measurements and six photographs of the courtyard. Two AIs worked the problem in sequence.

The first designed the system: a battery-operated tap timer for automation, a Y-splitter to handle the courtyard’s sliding gate, 13mm polyethylene pipe routed along the planted beds, and twenty-five individually adjustable drippers, one per plant. The individual drippers matter because the courtyard mixes drought-tolerant olives with thirsty calla lilies; a single flow rate would kill half of them. Seasonal watering schedules calibrated for the local climate came with the specification.

The second AI was resident in the brother’s browser so could directly navigate to the local branch of a large DIY chain. The original design specified Gardena components at around 122 euros. The procurement AI found compatible Spanish alternatives at 82 euros, confirmed every item available for in-store collection at the local store, and placed the order. Home delivery was not reliable and would have stalled the project.

The brother just had to click and pay. He said: “It’s like having your own personal shopper”. Collection, neatly boxed, at the local store was ready within two hours.

Outcome

From initial brief to paid order: under one hour. Her brother collects the kit, travels down for a visit and installs it himself in an afternoon. No tradesman, no language barrier, no two-week wait for a quote.

Why it matters

The traditional route for this kind of project involves consulting a local irrigation specialist, waiting for a quote, negotiating, scheduling. Two to three weeks minimum, assuming the language barrier can be navigated.

The two-AI approach separated design from procurement cleanly. The handoff between them worked because the design specification was complete and precise enough to be acted on directly.

The brother came to the project with practical intent and no specialist knowledge. The AI closed that gap entirely. What he brought was the sketch, the photos, and the constraint about in-store pickup. Everything else was handled.

A detailed write-up of this case — including the full component list, the seasonal watering schedule, and the Spanish/Gardena equivalence notes — will be added in due course.

Tools
Claude (two instances)
Time to outcome
Under one hour, problem to paid order
Who
Property owner’s brother, acting remotely as project proxy

← All case studies

practicalwisdom.ai · April 2026