Gooey.AI Workflow Accelerator Supported by The Rockefeller Foundation
For frontline-focused Government & Non-Profit Organizations
Last updated
Was this helpful?
For frontline-focused Government & Non-Profit Organizations
Last updated
Was this helpful?
UPDATE:
Check out this wonderful post by Kevin O'Neil, Managing Director, New Frontiers, The Rockefeller Foundation about the Gooey.AI Workflow Accelerator:
AI offers incredible potential for social impact and government organizations.
With this accelerator, we will support a cohort of non-profit and government organizations in building expert AI assistants—informed by your organization’s knowledge, data, and internal systems—to measurably advance your mission and SDG goals.
At Gooey.AI, we have significant experience creating AI solutions used by thousands of farmers, nurses, educators, technicians, and other frontline workers. Over the last two years, we have gathered evidence that AI tools can positively impact lives, enhance productivity, and strengthen organizational capacity across a variety of languages, education levels, skills, and income groups.
We want to help many more organizations effectively use AI by making these solutions more affordable, reusable, and accessible. Our goal is to demonstrate how multiple organizations can leverage a common set of human-readable AI workflow recipes, learning from one another as they each build AI solutions tailored to their specific missions.
We are specifically looking to help organizations confronting the following challenges:
You need to quickly provide beneficiaries or frontline workers with expert advice or personalized information while ensuring data privacy and compliance with security standards and regulations.
This advice should be based on trusted sources of knowledge—such as videos, images, text, the web, audio, PDFs, Google Docs, or internal systems—that your organization already has access to.
You want to deliver this information cost-effectively, rapidly, at scale, and potentially across language barriers.
Your organization's experts or mentors already provide this advice to frontline workers or beneficiaries, but you aim to dramatically expand your reach by using AI to handle common requests, freeing up experts to focus on more complex challenges.
You want to learn best practices from AI experts and peer organizations.
Here are some examples of this general use case in specific settings, but many others are possible:
A community healthcare worker gets coaching advice via WhatsApp voice notes on how to handle an unfamiliar set of symptoms via your organization’s Android app, drawing on your diagnostic handbook and/or approved WHO guidelines.
A water infrastructure technician can snap a photo of a broken pump on their phone and get trustworthy advice on how to fix it, along with diagrams and page-level citations to relevant manuals.
A farmer gets personalized AI advice on planting strategies via SMS, drawing on existing agricultural extension documents, diagrams, training videos, real-time weather/market data and their location, crop, etc.
A social worker accesses a web portal to determine which safety net programs a client is eligible for. The portal draws on a legal corpus of eligibility guidelines, along with customized rules that align with your organization's existing IT systems.
AI in 2025 excels at summarizing information in a highly personalized way. Speech recognition across hundreds of languages, translation, information retrieval, summarization, visual and text analysis, and step-by-step reasoning are rapidly becoming solved AI challenges.
However, these systems work best when they draw from known, relevant, and trustworthy sources—unlike off-the-shelf chatbot services like ChatGPT, which may cite popular but inaccurate references. The key is to build and refine AI solutions tailored to your organization's mission, clients, and specific knowledge base—and that’s exactly what we will help you do.
If your organization is selected for this program, we will work with you to build, test, evaluate, and iterate an AI workflow that directly serves your use case. This could be a chat or voice bot that interacts with your clients or frontline workers, providing them with customized advice in their language, using your internal knowledge and other trusted resources—ensuring you can rely on it. If you're unsure whether this program is the right fit, feel free to reach out to us at accelerator@gooey.ai.
As part of this accelerator, you will collaborate with other organizations tackling similar challenges, gaining insights from their experiences and tools. You will have the opportunity to share AI workflows, prompts, success strategies, and evaluation methodologies. We will also work closely with you to test AI tools in real-world scenarios—ensuring that your beneficiaries, data, and mission remain protected.
Our goal is to pioneer a new model of AI development—one where organizations focus on content, user experience, impact measurement, and iterative improvement (via AI Workflows), rather than diverting resources toward technical execution or keeping pace with the relentless wave of AI advancements (leave that part to us).
What You Will Receive:
A deployed AI solution (e.g., a chatbot) designed to support your use case, along with key evaluation and training materials to help you deploy, manage, and assess its effectiveness—including cost estimates at scale.
Deployment of your AI copilot on the web, voice-based phones, SMS, in-app or WhatsApp, with GPU and token costs covered. This includes access to the latest private and open-source AI models from OpenAI, Google, Mistral, Meta, DeepSeek, Anthropic, and more.
Iteration support from pilot to production, scaling up to 1,000 users.
No-cost AI consulting and training during the accelerator period, plus discounted or at-cost usage of Gooey.AI afterward (up to $30,000 in value).
A secure team workspace on Gooey.AI for collaboration on workflows and content.
Full ownership of AI assets—including code, prompts, knowledge bases, evaluation dashboards, and analysis scripts. These assets will belong to your organization or be made publicly available when appropriate, ensuring flexibility in deployment across platforms.
Technical and business evaluation tools to align AI adoption with your organizational goals, such as:
Understanding how bot usage fits into your theory of change and connects to SDGs.
Assessing which AI models best understand or translate representative audio clips from your users.
Dashboards to track solution utilization and measure its impact.
Co-learning opportunities with other organizations going through the same process.
12 virtual training sessions with AI experts and other cohort members
A report on the performance of your AI intervention for your own decision-making, and for sharing with funders and/or other stakeholders at your option.
We are looking for organizations that meet the following criteria:
You have a use case resembling those above that is already part of how you advance your mission or is planned for the immediate future
Serve at least 1000+ people per year (either offline or digitally)
You are an accredited government agency or non-profit organization
You have beneficiaries or frontline workers (e.g. caseworkers) who can primarily interact with your services digitally via their phones or computers (vs in-person visits)
Your organization leverages a documented knowledge base (webpages, videos, PDFs, Google Docs, call center FAQs, 1000s of PDFs) and/or APIs (to access internal or external systems, databases, etc).
You have a small technical team (1-2 people) who can work with us to plug the AI system into your existing technology stack, but you would not otherwise have the capacity to build such a system right now (e.g. you do not already have your own NLP researchers on staff or an engineering team greater than 20 people).
You have at least one field-level expert who deeply understands the use case and will gather usability feedback from the frontline workers or beneficiaries of the system.
Have a client-benefit or health-related mission – Your organization directly serves people rather than focusing primarily on advocacy, research, or policy drafting.
Enable quickly measurable impact – The value of your intervention can be assessed within 30 days of a user's first interaction (e.g., increased connections to services, reduced time to resolve an issue, etc.).
Demonstrate a strong interest in scaling with technology – You are committed to leveraging AI and digital tools to expand your impact.
To participate in the Accelerator, your organization agrees to:
Dedicate a field-level domain expert AND technical PM in the organization to attend ~15 1-2 hour virtual working sessions over the course of 6 months to develop the solution and learn alongside other organizations. They will also need to devote 8-10 hours per month to develop the solution and iterate it with user feedback.
Agree that the software, AI workflows and other tools developed for you will be ultimately made publicly available under creative commons license and available to other organizations in your field.
This does NOT apply to your internal data, frontline worker or beneficiary data, conversations, or any personally identifiable information.
Work to make referenced knowledge base files accessible (e.g. shared Google or Microsoft365 documents) and/or systems securely available via APIs.
Provide 20-40 “golden” question and answer pairs ie. Common or difficult questions and the ideal answers that a human expert would ideally provide. We will also work with your expert to document the process, reasoning, knowledge sources and tools they would use to craft their answers.
You will test the AI solution with us
Advertise the AI solution to at least 100+ users.
At least 20 users are made available for feedback-related interviews.
You will report on how effective the AI solution is for your mission.
You will participate in evaluating the approach to building AI solutions as an intervention and the results of which will be made public. The goal of this evaluation is to assess our approach and the extent to which it helps the participating organizations.
For this round, we will be accepting 6 organizations into the Accelerator. Organizations that are not selected will be able to follow our progress and we will publish the cohort’s insights, AI workflows and techniques as we progress.
Feb 19, 2025 - Applications Open on https://gooey.ai/AcceleratorApply
March 14, 2025 - Applications Due
March 19, 2025 - 6 Winning Organization Informed
March 24 - May 9, 2025 - AI Training and Prototyping
Weekly 1-2 hour zoom training sessions covering:
Golden Evaluation & Common QnAs
Golden audio & language model selection (for non-English languages)
Knowledge base curation
LLM Prompting
Agentic fun: Connecting to tools, functions and external APIs
WhatsApp, Voice and Web integrations
Fine-tuned reasoning
Theory of Change and SDG measurement
Analytics, conversation categorization & dashboard creation
Privacy, security and PII compliance
Cost/benefit and ROI modeling - what’s the cost to successfully serve a user with AI?
Ongoing AI model optimization and knowledge base maintenance
Deliverable: Each org has a prototype to usability test with users
May 12 - June 6, 2025 - Iterate with Users
Weekly demos and best practice sharing sessions over zoom.
Identify gaps through user feedback and iterate.
Share best practices and prompts.
Revisit and update AI models
Deliverables: AI solutions improved with user feedback. Scale to >20 users. KPIs are being measured.
Jun 9 - July 3, 2025 - Deployment to Statistical Significance
Scale up deployment to reach statistical significance
Iterate dashboards and refine metrics
Understand ROI and expected costs at scale
Share workflows with the broader community
Deliverables: AI solutions deployed to 100s of users and workflows publicly shared
July 7 - July 21, 2025 - Accelerator Evaluation and Reflection
Gooey.AI is a secure, low-code AI workflow platform for frontline-worker focused organizations to quickly deploy and measure the impact of AI solutions. Our work has been demo’d at the UN's General Assembly, our partners include the Gates Foundation, Microsoft Research, Opportunity International and PeoplePlus.AI and over 1M people have run our AI workflows since 2023.
The Rockefeller Foundation’s mission — unchanged since 1913 — is to promote the well-being of humanity throughout the world. Today the Foundation uses advances in power, health, food, and finance sectors to ensure everyone has good jobs, good food, good health, and more at a time when climate change’s effects are taking lives and undermining livelihoods.
If I’m accepted into this program, does it cost money?
No. Rockefeller Foundation is supporting the Gooey.AI platform (which includes associated AI costs from Google, OpenAI, Anthropic, etc), team training and usage costs for the 6 accepted organizations, up to 1000 users in 2025. Thereafter, Gooey.AI will offer at cost support up to $30,000 per organization.
We’d love to hear about your use case, but we are unlikely to accept applications from 501(c)(5) organizations or for-profit social enterprises—unless you partner with an accredited NGO or government agency that serves as the primary participant in the Accelerator.
Feel free to reach out to us at accelerator@gooey.ai and we'd be happy to discuss further.
If I build a solution in the program, can I easily move it from Gooey.AI to other technology platforms?
Yes! This effort is about abstracting AI solutions into their component prompts, evaluations and models and as such, we’ll offer tools so that you can export and run your AI solution on other technology platforms such as MSFT PowerApps, OpenAI’s GPT Builder, etc.
What level of privacy and compliance is supported?
Currently, Gooey.AI is both GDPR and SOC2 Type II certified and we have agreements with Google, OpenAI, Microsoft and Anthropic that data shared will be private and NOT used for training purposes. For additional details, please visit https://gooey.ai/privacy.
I’m geeky. What’s the technology architecture of Gooey.AI and how do my knowledge docs and APIs integrate with it?
Here’s an example workflow of a typical AI workflow intended to give Malawi farmers personalized advice via WhatsApp, voice or an Android application. The green portions would be provided by you - the partner organization - while the white and gray boxes are provided by Gooey.AI. To learn more, please visit https://docs.gooey.ai
Plus, here’s an architecture diagram showing how AI Workflows act as middle-ware between AI models and communications platforms.
