To learn more about Digital Green’s AI-powered platform and how it impacts small-scale farmers in India, Ethiopia, Kenya, Nigeria and Nepal, we encourage you to sign up for our newsletter.
INSIGHTS
Blog
Farmer.CHAT by Digital Green – Making Vetted Farmer Knowledge Accessible
October 16, 2023
Farmer.CHAT
An AI Assistant created by Digital Green to make vetted farmer knowledge accessible.
Experience the power of Farmer.CHAT!
Scan or click image to chat:
Background and Focus
Raising small-scale farmers’ productivity and incomes is critical to improving rural livelihoods. However, most farmers lack access to relevant and actionable information that could help them to improve their productivity and respond to climate changes.
Traditional farming relies on public extension systems, often delivered one-to-one through in-person training by frontline extension agents. Extension workers face many challenges in this system, such as reaching remote farming communities, providing timely information for critical farmer needs, and gender-specific guidance.
Since 2008, Digital Green has tackled these challenges of increasing the effectiveness of extension workers with our core approach – videos produced and for farmers – which leverages peer-to-peer, video based learning and existing social networks. We’ve ensured sustainability by partnering with Ministries of Agriculture in India, Kenya, and Ethiopia, securing co-funding and integrating our approach into daily operations of their extension systems.
Randomized controlled trials have found that, compared to those traditional extension services, our video-enabled extension approach: is 10 times more cost-effective than on a cost per adoption basis (from $35 to $3.50); reaches 30% more farmers; and delivers a 43% gain in practice adoption rates.
Core Approach
Digital Green aims to utilize our insights gained in agricultural extension and recent developments in artificial intelligence and large language models and achieve an even greater scale of strengthening the public extension system.
Our approach is based on our Theory of Change (TOC), which outlines three key elements. Firstly, we aim to create a user-friendly and low-cost AI chatbot to enable advisory delivery. Secondly, gender and climate-smart agricultural content will be collected and made available through the chatbot. Finally, we will build the capacities of extension agents (EAs), including women, to access and deploy the chatbot.
This will lead to EAs engaging with more farmers and delivering increasingly more relevant and customized advisory to them, EAs expanding their on-job knowledge, building their confidence, saving time, and agronomists and other experts using farmer data and famer/EA feedback to build gender-sensitive and CSA-models (i.e. fertilizer, seed, etc.) to continuously enrich advisory delivered by the chatbot.
Ultimately, this will enhance advisory quality as well as speed and scale of distribution, granting more farmers, especially women, access to knowledge. As a result, gender and climate-smart farming practices will see increased adoption, promoting higher productivity and incomes for farmers, and reducing gender disparities while enhancing climate resilience.
On-Demand Tailored Advisory
We have built a multi-lingual AI powered chatbot which acts as both a content retrieval and delivery mechanism to both receive and push messages for customized content and query the database on demand for EAs to triage issues they see in the field with location-specific nuance. This chatbot is multilingual and overcomes technological literacy barriers with multi-media inputs and response messages including text, audio and video.
Real-Time Content Improvements
Core to the platform is a protocol enabling seamless and secure data sharing which allows for ongoing improvement to localized advisory services, adapting to diverse agricultural needs. This protocol allows for external data sources such as vetted research, local weather, and soil databases to be dynamically updated, bringing the most up-to-date information to the field. Building off best practices in conversational UI we are creating an advisory delivery and feedback channel which actively collects and incorporates EA feedback to bot interactions, ensuring that the content it delivers remains relevant, up-to-date, and effective.
Reporting and Analysis
In support of our government partners’ data-driven decision-making processes, we have developed a system of automated dashboards which enable analysis and real-time tracking of usage behavior and other relevant metrics to improve the bot’s performance and design. To assess cost per farmer reached and practice adoption metrics we are developing a cost analysis component. Additionally, data dashboards and a farm-level report generator assist EAs to understand farmer needs and plan their activities, and to assist extension system managers to view reports of EA and service provider activities, and evaluate EA and content effectiveness.
Farmer.CHAT as featured at the
United Nations 77th General Assembly
Farmer.CHAT on Telegram