Before building the field version of CropSCOUT, we ran structured interviews with farmers in Jizzakh, Surkhandarya and Shakhrisabz — covering what they grow, what eats their time, what eats their money, and what they would actually pay for. This page is live: every figure below is computed from the responses as they arrive.
The sample runs from 5 sotix to 41 ha, with a median of 27.5 sotix. Reported experience has a median of 10 years across the 29 respondents who gave a number — some answer not in years but with "since I can remember."
Farm type is multiple-choice — the three counts sum to 37, not 31, because most respondents work more than one kind of ground.
| Holding size | Respondents | Share |
|---|---|---|
| Under 10 sotix | 5 | 16.1% |
| 10 – 20 sotix | 9 | 29.0% |
| 21 – 50 sotix | 5 | 16.1% |
| 51 – 100 sotix | 6 | 19.4% |
| Over 1 hectare | 5 | 16.1% |
n = 31 · 1 sotix = 100 m², 1 hectare = 100 sotix. Where a respondent gives a bare number we read it as sotix.
Single choice, one answer each.
n = 25 · single answer per respondent
Asked which job costs them the most in a season, 8 of the 24 who answered named spraying or plant protection. This is the clearest signal in the data, and it is the reason CropSCOUT launches as a precision sprayer.
n = 24 · free-text answers grouped into categories by us — that grouping is a judgement call, not a measurement
When we spray chemicals by hand, more than the correct dose lands on the crop. But if it were applied evenly and precisely — the same measured amount to every plant — the chemical would not be wasted, and it would not be dangerous for the health of whoever eats it.
— respondent, final commentsThat is our product thesis, arrived at independently by a farmer who had not heard it from us. It is why we frame CropSCOUT as a low-pesticide machine rather than a labour-saving one. The cost farmers name is not only money — it is the operator standing inside the spray cloud, and the residue that reaches the person eating the tomato.
Each respondent watches three short videos — an electric weed cutter, a spraying robot, and a universal multi-tool robot — and rates each from 1 to 5.
Mean rating, 1–5 scale · n = 29 (not every respondent completes the rating grid)
These averages have shifted every time respondents were added. With 17 people rating, the electric cutter averaged 3.94, the sprayer 4.18 and the universal robot 4.24. With 28, they sit at 4.00, 4.04 and 4.46. Nothing about the concepts changed — only who was asked. Treat the ordering as provisional.
Asked which concept would genuinely make their work easier — Spraying robot 20, Universal robot 14, Electric cutter 11 of 31. We weight this question more heavily than the star ratings: it asks about use, not appeal.
Seasonal rental leads at 51.6%, against 25.8% for outright purchase and 22.6% for instalments. This is the model we had already planned, now confirmed by the people who would pay for it.
Respondents raise service before they raise capability: spare parts that are easy to find, a workshop within reach, and a price a farmer can actually meet. None of them mention autonomy or software. For us that is a distribution problem, not an engineering one — and we do not yet have an answer to it.
Answers to "what would help you most today?" — published without names, and only from respondents who agreed to anonymous publication.
The first thirty interviews asked what eats a farmer's time. These ask what the chemistry costs — how often they spray, who advises them, and what has gone wrong. Every figure below counts only the farmers who were asked these questions, not all 31.
n = 6 · asked only in the second round
and how they fight it · 6 answered
product, dose, and the area one application covers · 6 answered
and what a visit costs · 6 answered
chemicals applied incorrectly · 6 answered
31 farmers across three areas is a starting point, not a market study. We publish it in full — including the answers that contradict our pitch — because the next round is designed around these gaps:
If you farm in Uzbekistan, the survey takes about five minutes and the numbers on this page will include you the moment you submit.
Take the survey → See the robot