Recipient Organization
GOOD AGRICULTURE INC
2025 ARLINGTON AVE NE
ATLANTA,GA 30324
Performing Department
(N/A)
Non Technical Summary
Imagine, as a farmer, speaking into your phone - "I just harvested 397 pounds of radishes from bed 19." Behind the scenes, your words are transferred to an AI that connects the new data to everything that's ever happened in bed 19 during the recorded history of the operation. It knows that these are French Breakfast Radishes, that they were planted 43 days ago, that the soil is a clay loam that was amended with potash prior to planting, that the field got two inches of rain on four occasions over that 43-day period, that irrigation was run on another 3 occasions, and that the yield of 397 pounds is about 20% lower than expected. By the time you get the radishes topped, washed and in the cooler, the AI has taken additional dictation to fill out necessary food safety forms and calculated that your weight loss from field harvest to washed product is 17% higher than expected. When you sit down that evening, you get a digest of insights from the day, including the suggestion that you may want to switch to a different source of potash with a lower nitrogen concentration for your next radish crop.We are building a flexible farm management system for small and mid-sized farmers that uses existing voice to text technology for data entry and artificial intelligence to analyze the data entered. The tool will be able to create a bespoke system for each farmer who uses it, essentially becoming a digital analyst highlighting potential problems in farm operations. This will reduce the resources needed for a farm to perform operational analysis, giving small and mid-sized farmers a more even playing field with larger farms.To build this we are collaborating with existing companies to allow for easy data collection and collaborating with small farmers to build a data set on which to train our AI. We plan to start with data from five annual crops for our Phase I project and expand from there. Our goal is to have a working prototype at the end of Phase I that can be useful to farmers even if it doesn't have all the features we hope to eventually offer. Rather than creating a stand alone product, our plan is to partner with existing software for small and mid-sized farms, thus allowing the innovations to quickly reach farmes across the US.Good Agriculture offers scalable back office services for small and mid-sized farmers.We've seen our farmers increase their profitability after using our financial services simply by knowing what's going on with their finances. We expect to see the same effect as farmers begin using this tool and are able to get a clear view of their operations and make improvements that result in more efficient production. By giving these farmers the tools to build more efficient businesses, the overall food production system in the US will become more resilient and able to withstand shocks from a changing climate, disease and natural disasters.
Animal Health Component
50%
Research Effort Categories
Basic
0%
Applied
50%
Developmental
50%
Goals / Objectives
Our goal in Phase I is to create a usable product for small and mid-sized farmers that allows them to easily collect their data and uses artificial intelligence to analyze it and present suggestions for improving operations.1. Objective 1: Create a usable product for small and mid-sized farmers that allows them to easily collect and query their data.2. Objective 2: Develop artificial intelligence algorithms to accurately analyze the collected data.3. Objective 3: Develop artificial intelligence algorithms to present suggestions in a meaningful way to improve farm operations.
Project Methods
Stage, Objective, and Technical Questions:Stage 1A (months 1-2); Objective 1; Technical Questions 1, 2Conduct tests on 10 different farms with 5 different annual crops and 10 farm operators.Source existing voice-to-text technology and test its accuracy.Ask 10 farmers to collect observations via phone app or voice memo app.Transcribe observations using the voice-to-text tool and assess accuracy.Proceed to Stage 2A if accuracy is good, otherwise explore alternative options.Stage 1B (months 1-2); Objective 2; Technical Question 5Conduct in-depth interviews with 25+ farmers and surveys with 50+ farmers.Refine hypotheses based on interviews and validate with surveys.Follow customer discovery best practices for qualitative and quantitative assessment.Stage 2 (months 3-4); Objective 1; Technical Questions 1, 3, 4, 6Engineer working with data scientist and farmers will lead this stage.Test data entry and quality control processes using existing solutions.Evaluate accuracy and assess errors due to technology or operations.Identify and incorporate additional data sources into the database.Stage 3 (month 5); Objective 1; Technical Questions 7, 8, 13Data scientist and product owner will lead the data work.Develop a front-end for the database to allow farmers to query and export data.Train farmers on using the front-end and collect feedback.Track data error, completeness, customer adoption, and usage rates.Stage 4 (months 6-7); Objective 2; Technical Questions 8, 9, 10, 11Use customer insights to identify statistically-valid tests and visualizations.Build an MVP front-end with pre-set reports and user-run correlations.Track usage of visualizations and tests run by farmers.Stage 5A (month 8); Objective 2; Technical Questions 12, 13, 14Focus on commercial activities outside the grant budget.Reach out to potential customers, gather feedback, and improve onboarding.Track marketing and scalability metrics.Stage 5B (month 8); Objective 3; Technical Questions 15, 16Evaluate machine learning models based on data volume and types