The blog is put together by Tanu Prasad, from the Glific team
One of the two themes of this sprint was Prototyping New Ideas. In our conversations with partner NGOs, we often hear exciting suggestions from the brilliant teams managing or running chatbots on the ground—ideas that always keep the end-user at the center. This track was designed to give those ideas a space to take shape. With one-on-one collaboration from the Glific Team, NGO participants were able to move from concept to working prototype, turning their thoughts into tangible action.
Read more about the impact sprint here.
For the prototype track, the process began with NGOs identifying the specific problem they wanted to solve and mapping out how a prototype could address it. Most of them had already completed this prework before arriving at the sprint. This prework made it possible to fully utilize the two sprint days—dedicated to building, testing, and refining prototypes. By the end, NGOs walked away not just with a functional version of their idea, but also with clarity on the reasoning behind it, giving them the confidence to adapt and reuse the solution independently in the future.
A total of 8 NGOs were a part of the prototype track, all working in diverse areas within the social sector—ranging from education and livelihoods to ecological equity, gender equity, and safety. Alongside the common sessions held for both tracks, prototype participants were also encouraged to re-think their engagement metrics and reimagine the prototypes they were building. This created an opportunity not only to document their learnings within the prototypes but also to reflect on their use cases more broadly.
With over 20 participants working closely with the Glific team, the experience was one of true collaboration—co-creating solutions while ensuring that NGOs could take back the learnings, refine their prototypes further, and independently repurpose their implementations for the future.
Here’s a glimpse into what it was like for the Glific team to work hand-in-hand with the Prototype track NGOs.
- The Antara Foundation
The Antara Foundation is an Indian public health NGO dedicated to addressing maternal and child mortality and malnutrition. They partner with the government to improve public health systems by training frontline health workers and implementing data-driven solutions. Their mission is to ensure every mother and child has an equal chance at a healthy life.
Danish Kalim and Ravi Raj, joined at the sprint from the Antara Foundation.
As part of this sprint, the team explored an idea to collect form data through voice input, aiming to reduce the effort for users who would otherwise have to manually type into a form. Imagine the convenience of simply speaking out detailed responses instead of entering them line by line.Aha Moment! When the collected data was passed through an LLM. For example, when we said we were in Gopal, MP, it correctly suggested Bhopal. This highlighted the potential of LLMs to not only capture but also auto-correct and refine data, making the process far more reliable.
2. Save Missing Girls – Missing Link Trust
Save Missing Girls is a non-profit organization dedicated to child safety, focusing on preventing sex trafficking and abuse while promoting inclusivity and behavioral awareness during adolescence. With limited prior experience in managing chatbots, they were guided to prototype their first ideas for leveraging this technology. The goal was to develop an AI-powered chatbot that enables children and caregivers to access help, information, and support related to trafficking and sexual abuse.
Radhika N and Panchali Kar joined from the Missing Link Trust team
The AHA moment stands when they built a working prototype of their vision in just two days! Their limited prior experience with the platform turned into an advantage—bringing a fresh perspective that focused on what should be done and how, rather than being constrained by existing tools (with the consistent focus on keeping the end-user at the heart of the design, solving for them.
We also spent time brainstorming creative ways to take the prototype to users, pilot it in real scenarios, and gather meaningful feedback. This feedback would then guide the next iterations, ensuring the solution evolves in line with user needs.
3. Tiny Miracles
Working towards solving problems and bridging the poverty gap, Namrata and Meghna from Tiny Miracles joined us for the sprint (for the first time) to prototype a health chatbot that allows users (who are essentially factory workers at Tiny Miracles) to upload medical records, combines this with dynamic lifestyle data collected in KYC style, and provides behavior-change–based advice in response to their health queries.
Why pick this to prototype?
- Tiny Miracles observed that poverty-related health issues among their community—linked to diet, lifestyle, sanitation, and water—often lead to serious illnesses (TB, malaria, typhoid, diabetes-related complications). With limited resources, people are vulnerable to unnecessary procedures. The chatbot was envisioned as a way to give workers a reliable “second opinion” on their reports.
- Despite running for a while with a limited user base, engagement remained low. Noting that most workers are more comfortable listening and speaking than reading, they explored and tested voice-based messages to help users receive and understand information in their preferred format.
Seeing their first AI flow come to life on day one and a full prototype by the sprint’s end revealed that LLM flows aren’t as complicated to build as they imagined—The sprint made it clear: even “complex” AI flows become approachable when you dive in, experiment, and keep your users at the heart of the solution.
Through constant exploration and idea tweaks, Tiny Miracles created a prototype—by them, for their own workers!
4. ATECF
A.T.E. Chandra Foundation (ATECF) works on reviving water bodies to improve groundwater and agriculture, promotes regenerative farming for soil and climate resilience, and strengthens the social sector by supporting NGOs, governments, and tech-led solutions for systemic change.
ATECF has been using a chatbot for the past two years to train their Gram Sevaks on the Avni platform and RWB (Rejuvenation of Water Bodies) along with an AI-powered flow to answer their questions while using the Avni Gramin app. They initially joined the sprint to explore strategies for gamifying their training flows to increase engagement. However, the Impact Funnel workshop and conversations with the Glific team and peers made Aditya Jain (the POC from ATECF) reflect on the original purpose of the chatbot. This led him to design a six-month roadmap (before the next season) with improvements across various user touchpoints to achieve the chatbot’s proximal outcomes.
He also prototyped another LLM flow to collect data from Gram Sevaks during the post-season endline survey, which will then generate impact testimonials from the data.
Aditya also expressed his excitement to share this with his team, as it will significantly reduce the effort required by their communications team to gather stories on the ground.
5. The/Nudge Institute
The/Nudge Institute works on building a poverty-free India. They work closely with Governments and other NGOs to work towards this.They needed 2 different AI chatbots as part of one of their programs called the Economic Inclusion Program.
Changemakers (CMs) on the field interact with program participants, supporting livelihood initiatives and social and financial inclusion. They often need quick access to specific information about schemes, entitlements, and programs, but currently rely on their Block Leads for answers. A single Block Lead supports many Changemakers, which means this overwhelms the BLs and also not all the questions of CMs are answered on time. The organization aims to solve this by creating a chatbot that empowers CMs to find answers instantly.
The first bot was focused on reducing the mortality rates of domesticated goats and thus help improve the financial status of goatherds, while the second bot focuses on helping people know about various government schemes that they can get.
Since both the bots had similar functionality, the Glific Team only had to help them prototype the first one while they were able to build a barebone chatbot for the second one without much help. Special mention to Jaishil and Shruthi who joined us from Nudge. Shruthi was very enthusiastic about the bot she was building and eager to learn and experiment with Glific. She made sure she built the prototype herself, not the Glific team, by only asking questions when needed and saying, “Don’t do it, just tell me what to do so I don’t forget.
The Prompt Engineering session really helped them as they were able to identify fairly easily why the bot was not behaving correctly and was able to fix the issues quickly.
6. Leadership For Equity
Leadership for Equity is a non-profit organization in India that works in deep partnership with state and district governments to transform the public education system. They focus on building the capacity of government officials and teachers through co-created programs, policy advisory, and training. Their goal is to improve student learning and well-being at a large scale, ultimately ensuring every child has access to a quality education. Anjaly Gopi and Jesbin Skariah Thomas joined from the Leadership For Equity.
During the sprint, they developed a teaching-style assessment logic that, when paired with the questions teachers asked for support, could provide highly contextual responses.
The Aha moment was seeing how the combination of teaching style and the specific question allowed the system to generate tailored, context-sensitive answers that adapted to different teaching approaches.
7. Indian Partner Network (IPN)
IPN is a digital platform aimed at helping small-to-medium sized nonprofit organisations in India to build, strengthen and sustain their internal capacities to deliver impact more effectively. It is an initiative of Sattva — a social sector consulting / advisory organisation. IPN was created as a one-stop resource for nonprofits to get knowledge, visibility, networks, and connection opportunities. Puneeth JN was representing the Sattva team at the sprint.
We were exploring opportunities to automate the onboarding experience for NGOs on the India
Partner Network (IPN) platform to make the process smoother and more efficient. In addition, we also wanted to streamline engagement during events and webinars hosted on the platform. Specifically, since IPN already has event calendars, we considered an approach where collections can be automatically created based on upcoming events, and nudges or reminders can be sent to relevant NGOs, encouraging them to participate and engage more actively.
The objective behind this? It would help improve discoverability, reduce manual effort, and enhance overall participation in events and webinars.
The process of prototyping also gave the organization an opportunity to learn to use the Glific API, integrate Google Sheets and webhooks, and handle edge cases to scale his prototype effectively.
8. Hornbill
Hornbill has been supporting Fortify Health (FH) in running a program focused on ensuring that atta (wheat flour) contains sufficient iron. The program combines Glific with Hornbill’s image detection platform to provide mill workers with near-instant feedback on whether their atta batches need fortification. Traditionally, mill staff had to wait up to a week for lab results, but with this solution, they can get answers within minutes. Considering that a mill typically produces around five batches of atta a day, this shift translates into a significant boost in productivity.
Hornbill had already designed the chatbot flow to meet the program’s primary goals. The next steps were to refine the user experience and capture analytical data to further improve the platform. The specific requirements were:
- Simplify onboarding by asking mill staff to choose their language and enter their mill code only once, rather than every time they interact with the chatbot.
- Record details of conversations—such as image URLs and response times from Hornbill’s platform—into a Google Sheet, so the team can track and analyze performance.
- Reduce the time it takes to deliver iron content results from up to four minutes to under two minutes, ensuring a smoother user experience.
Sid from Hornbill was able to address the first two requirements with minimal guidance, making quick progress on those improvements. The third requirement proved more complex. It required collaboration with Hornbill’s development team, and while an online session with their developers helped, the process was slowed by an internal misunderstanding of one of our APIs. As a result, the time reduction goal wasn’t completed within the sprint. Work is actively underway with Hornbill to bring the turnaround time below two minutes.
Closing
In just two days, clarity met creativity—each team charted their own path towards the goal that they had set! NGO Sprints at Glific create a vibrant space for social sector organizations working with us to come together, share ideas, and collaborate with like-minded peers and fresh thinkers to tackle their goals. Since much of our work happens remotely, these sprints also give the Glific team a valuable chance to connect face-to-face with the NGOs we partner with, listening to their perspectives and learning how they approach their unique challenges.