Menu AI

Enabling nutritious, affordable and community-sensitive national school feeding programmes is a significant challenge. There are issues on budgetary constraints and nutritional requirements, while we need to ensure that ingredients are locally available (ideally, from smallholder farmers). On top of that, meals must be culturally acceptable. Challenges in balancing all this translates into fewer children covered, less nutritious meals and less economic impact. Moreover, communities are often excluded from the decision process.

Our Menu AI is an integrated school menu creation platform. It is a freely accessible web application using a powerful AI to optimize school menus. It incorporates a modality for integration of communities in the decision process, including data collection, following a process developed over 10 years of experience.

Menus designed with our tool are more affordable, more local and more nutritious. More children will benefit from school feeding, and more farmers will have access to a structured demand.

Poor diets are the cause of an estimated 1.1 million of the 3.1 million child deaths that occur each year as a result of undernutrition. One in two children worldwide receives school meals in over 161 countries, with an overall yearly estimated investment of 40B USD. However, 73 million children worldwide are in extreme poverty and are out of school feeding programmes, hindering their access to education in equal conditions. Thus, creating more nutrition and cost-optimized menus will allow a larger coverage and better meal quality. Local procurement will ensure supply chain resilience and benefit the local economy. The optimization performed by the AI allows the substitution of manual methods for menu design, achieving much more affordable, nutritious and local menus simultaneously. It includes a “community” mode, allowing the integration of local communities in the decision process,

More children receiving school meals translates not only into higher education access and performance and helps the closure of the gender gap in registration, attendance, dropout and child marriage. Beyond that, School Feeding is one of the best platforms known to tackle child malnutrition, and it has a deep impact on local economies when the supply chain includes local sourcing.

We created an online, free to use software that uses AI to calculate the best menu for national school feeding programmes while allowing data collection at ground level in challenging environments,  community integration and participatory processes. Based on local and seasonal food availability, nutritional requirements of the targeted children and local eating habits, the algorithm finds the best combination of ingredients to create a menu that is affordable, nutritious, local and culturally acceptable.

Governments worldwide can benefit from the power of the algorithm. The usability has been simplified over three years of user research and improvement, collecting inputs from over 20 countries. It can be accessed online at no cost, and the training time is just 4 to 8 hours. The algorithm can perform in minutes a task that requires up to four weeks for a human.

Moreover, the software includes a feature in which communities can, simultaneously, participate in the process and be educated on nutritious diets: By means of a tablet-friendly interface, school principals or government workers can collect food preferences, local food prices and provide training to communities. That data can then be used via AI to refine the menu.

The solution serves a wide set of primary and secondary beneficiaries. The primary beneficiaries include children, local farmers, food processors, traders etc. It particularly benefits women who are primarily responsible for local food processing and trading. It further benefits food security across the community, as school feeding impacts food availability at the household level.

By reducing the cost of the menus, the same budget can be used to feed more children, that right now are not receiving school feeding. Access to school feeding is known to improve registration, attendance and academic indicators; reduce the gender gap in access to education, and reduce the likelihood of child marriage. This will have a significant impact on lifelong productivity, especially for women. By increasing the nutritional content of the meals, school meals will be better placed to tackle local dietary deficiencies and improve the health of children. Healthier children will become healthier adults, living a longer and more productive life.

By increasing the amount of locally-produced food bought by the programme, we create a fantastic opportunity for local farmers to access a stable demand and fair prices. More income will allow them increased resilience and access to better inputs and technology, and the spillover effects give a strong economic push to the community. Studies from Ghana, Nigeria, Kenya and Nepal show how school feeding purchases can create significant opportunities for local farmers and farmer-based organization. The inclusion of certain food groups and individual foods in menus may strengthen their value chains and reduce postharvest losses, particularly those that are more commonly cultivated by smallholder farmers and women. There are studies in Kenya demonstrating that each dollar invested in school feeding purchases by smallholder farmers have an income multiplier of 2.27, which affects not only farmers but across the whole community.

By integrating communities in the decision process, we create an opportunity for local empowerment and the application of traditional knowledge. It also serves as a nutrition sensitization and behaviours change communication tool. Moreover, we have the opportunity to gather data at that level, including food prices, local eating habits and available products. Thus we can craft a menu that is totally adjusted to that specific community. The data collected there can be used for a number of studies and analysis; since at present, that granularity of information is rarely gathered.

Smallholder farmers struggle to access markets,

with sporadic sales at unfair prices. Integrating them into School Feeding’s
vast, stable demand, via adequately design menus, will give them access to a
stable demand, key for their resilience. Research in Kenya shows that each dollar
invested in locally-sourced school menus has a multiplier of 2.27, extending
across the community. On the other hand, budget savings means more children
within the community will be enjoying school meals. Beyond the AI, the tool
ensures that everyone in the community participates in the decision process,
trustworthy data is collected and the whole community becomes empowered.

There are already seven countries worldwide that are using the platform or are being onboarded. By September 2021, when most countries are bound to reopen their schools after the pandemic, we expect at least 2.5 million children will be receiving meals designed with the tool. Moreover, the current platform resulted from the union of the PLUS School Menus software (Existing since 2017, tested since 2019 in seven countries) and the PCD Menu Planner (Existing since 2011, implemented in 10 countries to date, designed menus for over 10M children).

There are four main areas in which the tool has become a game-changer:

The optimization provided by the AI is more advanced than previously seen. Most food optimization solutions are based on linear programming, which limits its applicability. Results are more often than not just a mathematical result, which is not gastronomically realistic. Our advanced algorithm integrates local-taste inputs, as well as worldwide food data (Price, nutritional content and origin) to provide the best menu. The tool also identifies the most cost-effective, nutritious crops that can be grown locally and integrated into local diets. One of our early adopters said: “The tool allows the transformation of nutritional demand into agricultural supply”.

Simultaneously, a very refined, simple interface was crafted over years to allow any user to tame the power of the AI. In four simple steps, the user provides all necessary inputs and obtains an optimal meal that can be then adjusted manually. It does in minutes a task that now takes weeks.

Nevertheless, the platform allows the integration of communities in the decision process. Following a procedure used and improved over 10 years by the Imperial College, and making use of a tablet-friendly modality, communities provide inputs and data, while training and sensitization activities take place.

Thus, bringing a new approach on how to design menus in social programmes can extend beyond schools. Emergency settings or hospitals can benefit. Moreover, agricultural development programmes can benefit from identifying those local crops bringing an equilibrium between price and nutrition.

The population served by the solution with the easiest measuring are children in developing countries and communities who are receiving the meals created with the tool. The first full version of the platform has been recently released, while most countries are currently in schools closures due to COVID.

However, in Dominican Republic and Bhutan, active users are crafting meals for a total of over 1.8M children. Moreover, meals have been already crafted for the programmes in Burundi and Angola – to begin in September. Therefore, current estimation is 2.5M children.

Programme managers in Nepal, Nigeria and Cabo Verde are being onboarded in the tool, for a total of over 10.5M potential school children. It is expected that they become regular users of the tool by the next school year, meaning 13M children served.

We have already received strong interest from the school feeding programme leads in Sierra Leone, Ghana, Kenya, Uganda and Namibia, programmes covering 5.7M children. Thus, for the next five years, we expect at least 18.7M children served by the tool. Nevertheless, we are receiving interest at an accelerated rate as the pandemic starts to recede, for which the actual estimation might be exponentially higher.

Beyond children, the rest of the community will be benefited by their integration in the decision process, as well as the increased numbers of local farmers that will be increasing their income thanks to meals crafted in a locally sensitive manner. The number of people within these two groups is, simply, unmeasurable.

As we are working to ensure that the maximum number of governments worldwide use this tool to develop their school menus, serving as many children as possible. Thus, our main indicator is the number of Governments using the tool, and the number of children that are supported by those programmes. Both indicators are easily measurable.

Nevertheless, the impact of the software is multidimensional, and each Government can look for different advantages. Focusing on cost reduction, or in a redistribution of the food sourcing to ensure that more local farmers are benefited from this demand opportunity. Therefore, there are many other different indicators that can be used at the National level, related to the improvement of the meal.

A sample list of possible/common indicators:

• Number of children benefiting from meals designed using the software.
• Number of schools and local government areas (for geographical coverage)
• Number of local cooperatives/farmer-based organizations linked.
• Dietary diversity of children.
• Production diversity

Core team:

• Full-time staff: 1
• Part-time and support staff: 8
• External Contractors: 4
• Total: 13

Beyond the core team, we count on a wide range of in-house experts providing ad-hoc support

The team has a wide range of expertise. We count on six School Feeding, Nutrition and Agriculture experts, with decades of combined programme delivery at field level around the world. We have renown world-class academics around child nutrition and school feeding programmes, with the vast international experience and prestige of the Partnership for Child Development, based at the Imperial College London.

Moreover, we have the World Food Programme’s TEC team leading the technological delivery of the software, which includes a Scrum Master, a UI expert and three developers; as well as a focal point at the WFP’s Innovation Accelerator, providing strategic inputs and connecting with partners.

Beyond the core team, within World Food Programme and Partnership for Child Development, we established a wide network of colleagues, with deep knowledge in different areas, that are providing continuous, ad-hoc contributions to the project.

Being a team merged from a United Nations agency and a top international research institution, diversity is granted.

We are collaborating with people from all over the world, while the core team comes from four different continents. Following internal regulations, any addition to the team follows a strict HR process, which takes into account diversity, equity and inclusion as core values in the hiring process.

Being selected will be key to overcome our current challenges and consolidating our current expansion process.

Firstly, having the support of MIT colleagues will allow us to review and enhance further our current software. Having access to some of the most brilliant academic minds in technology and beyond to support our solution will be a game-changer for us.

Then, the overall network will allow us to obtain fresher looks over our solution from experts in different areas, including communication and public relations, an area in which we are in need of knowledge. Mentorship and coaching will help us greatly in shaping our long term planning and managerial strategy.

Access to grants will be a fantastic push to our current funding strategy for 2021, and would be key in accelerating our process towards global expansion

And finally, exposure to media and the unparalleled prestige brought by this nomination would facilitate greatly our upcoming work on expanding the usage of the software by governments worldwide, apart from increasing the attention from donors and partners.

We would welcome ad-hoc expert support in areas such as communications and branding, in order to craft a strategy adapted to the particularities of this project. 

On the other hand, we are always aiming to walk an extra mile on our technology. We have a number of ideas to improve our AI and UI, for which we would require financing mainly – but technical expertise and advice from knowledgeable people outside the project would go a long way. We are looking forward to receiving fresh looks into our product, critique, suggestions and a bit of out-of-the-box thinking.

Finally, managerial advice on how to run the project in the long run, someone to challenge our current ideas, will be very beneficial. We would be glad to receive this kind of advice and trigger further brainstorming to ensure we are on the right path to financial and programmatic sustainability.

We would love to partner up with big TEC organizations, especially UX & AI experts. We have limited resources to continue the building up our technology, and as far as we have ideas and technical capacity, we would be able to achieve larger potential if we count on partners from the sector. Large companies that are able to provide expertise, funding and technology access are one of our main targets. Furthermore, academic expertise from the vast pool of MIT’s human capital around technology will be largely welcomed.

On the other hand, we are already partnering up with a number of governments, but we would be happy to expand this list (i.e.: Global North governments able to provide us support). Moreover, linking up with multilateral organizations and additional regional bodies will accelerate the adoption of the software.

Our software provides a simple, scalable and efficient solution to make nutrition and education more accessible while empowering local communities. By crafting optimal menus for national school feeding programmes, we ensure instant impact at a massive scale. The optimization AI algorithm can reduce food purchases costs by up to 20% in this multi-million government investments, allowing many more children to receive school meals - one of the most effective interventions leading to an immediate increase in access to education, attracting more children to schools.

Beyond just costs, the tool considers the variety of improvements in different areas that school meals bring along. The AI considers as well nutritional and food sourcing parameters. This means that we can safeguard or increase the nutritional value of the meal provided to children – improving their health status and future development - and at the same time, craft the menu in a way that food comes mainly from local sources, ideally from smallholder farmers. This provides a fantastic opportunity for vulnerable farmers to access a large, stable, fair-priced demand, a key step in their resilience building.

Moreover, the software considers the importance and key contribution of local communities to any intervention. We provide an easy technology to collect their inputs, educate them on the importance of nutritious diets and get to understand how is food understood culturally – everything integrated within the same platform.

Therefore, in a single activity, we are able to build the resilience and empowerment of a whole community. Increased access to education for better-nourished children, vast market opportunities provided to local food producers and communities empowered and involved in the whole decision-making process – all these benefits come from a better menu design.

We could benefit largely from the GM Prize award, in order to continue the improvement of our technology, and have access to more academic evidence about the impact and features of menu design in fragile contexts such as refugee-populated areas.

Our software provides a simple solution to make education more accessible while empowering local communities. By crafting optimal menus for national school meals programmes, we ensure having an immediate impact at large scale. The optimization AI algorithm is able to reduce food purchases costs by up to 20% in this multi-million investments, allowing many more children to receive school meals within the same budget. School meals is one of the most effective interventions leading to an immediate increase in access to education, attracting more children to schools.

Beyond just costs, the AI considers as well nutritional and food sourcing parameters. This means that we can improve the nutritional value of the meal provided to children – improving their health status and future development - and at the same time, craft the menu in a way that food comes mainly from local sources, ideally from smallholder farmers. This provides a fantastic opportunity for vulnerable farmers to access a large, stable, fair-priced demand, a key step in their resilience building.

Moreover, the software considers the importance and key contribution of local communities. We provide an easy technology to collect their inputs, educate them on the importance of nutritious diets, and get to understand how is food understood culturally for them – everything integrated within the same platform.

The prestige and financial support coming along with the AI for Humanity Prize will help us in continuing the global expansion of our solution, now supporting seven governments, and to keep on enhancing our technology.

Our software provides a simple solution to make education more accessible while empowering local communities. By crafting optimal menus for national school meals programmes, we ensure having an immediate impact at large scale. The optimization AI algorithm is able to reduce food purchases costs by up to 20% in this multi-million investments, allowing many more children to receive school meals within the same budget. School meals is one of the most effective interventions leading to an immediate increase in access to education, attracting more children to schools.

Beyond just costs, the AI considers as well nutritional and food sourcing parameters. This means that we can improve the nutritional value of the meal provided to children – improving their health status and future development - and at the same time, craft the menu in a way that food comes mainly from local sources, ideally from smallholder farmers. This provides a fantastic opportunity for vulnerable farmers to access a large, stable, fair-priced demand, a key step in their resilience building.

Moreover, the software considers the importance and key contribution of local communities. We provide an easy technology to collect their inputs, educate them on the importance of nutritious diets, and get to understand how is food understood culturally for them – everything integrated within the same platform.

The prestige and financial support coming along with the GSR Prize will help us in continuing the global expansion of our solution, now supporting seven governments, and to keep on enhancing our technology.