Disclaimer:
Please be aware that the content herein has not been peer reviewed. It consists of personal reflections, insights, and learnings of the contributor(s). It may not be exhaustive, nor does it aim to be authoritative knowledge.
Title
Please provide a name for your action learning plan.
Learning by Testing: Generative AI for Agile and Anticipatory Water Planning
Challenge statement
Challenge type: If you are working on multiple challenges, please indicate if this is your "big bet" or "exploratory" challenge.
Please note: we ask you to only submit a maximum of 3 challenges - 1x Big Bet, 2x Exploratory. Each challenge must be submitted individually.
BIG BET
Challenge statement: What is your challenge? (Please answer in specific terms: "Our challenge is that...”.)
Our challenge is how public officials can leverage generative artificial intelligence tools to review and analyze applications for new connections to the water network, in order to improve the agility and accuracy of the feasibility process with anticipatory capacity*.
*Note: Agility and accuracy refer, for example, to improving the cross-referencing of information, analysis, and decision-making between silos (i.e., the internal areas of the State Water Commission responsible for conducting technical analyses to determine the feasibility of a new application). Anticipatory capacity can be applied, for instance, to planning hydraulic infrastructure or adjusting requirements for new developments.
Background: What is the history of your challenge? What is causing or driving it? Who is involved? How does the current situation look like? What undesired effects does it produce?
NOTE: This Action Plan continues AccLab Mexico’s research and development efforts in water management and generative artificial intelligence, as first outlined in our previous Action Plan: “Promoting water sustainability with generative AI: Transforming the feasibilities procedure into a planning tool for water resources management” (go to: https://sdg-innovation-commons.org/pads/action%20plan/2349). In this new Action Plan, we focus on a deeper understanding of the feasibility procedure and on piloting a concrete solution.
The framing of our current challenge emerged from our previous efforts with the Water Commission to enhance the process for evaluating new water service connections in Querétaro. While the current feasibility analysis provides valuable information, several areas of opportunity for improving the analysis have been identified. Estimating water availability and future demand remains complex due to fragmented and outdated data, limited analytical tools, and coordination challenges across departments and institutions. Additionally, the ability to anticipate and respond to climate variability is constrained by the absence of integrated scenario planning tools. Internally, information silos and the lack of standardized mechanisms for data sharing can delay decision-making. These conditions contribute to a reactive planning environment, instead of a more strategic future-oriented one.
Quantitative evidence: What (official) data sources do you have on this challenge that better exemplifies the importance and urgency of this frontier challenge? You can add text, a link, or a picture.
Water scarcity is a big issue in Mexico, including the state of Queretaro. Regarding the total territory in the state of Queretaro, the percentages of areas affected by a particular type of drought are distributed as follows, as shown in the image of the map of Mexico:
0.0% Unaffected
3.8% D0 Abnormally dry
17.6% D1 Moderate drought
15.3% D2 Severe drought
27.2% D3 Extreme drought
36.1% D4 Exceptional drought
This means that a bit more than 60% of the state suffer from the highest types of droughts as of June 2024.
Furthermore, out of the 18 municipalities in Queretaro, one is experiencing moderate drought, two severe drought, four extreme drought, and eleven exceptional drought, which is the highest form of drought. These numbers show the vulnerability of the state with regards to access to water, illustrating the urgency of tackling this challenge. [1]
According to the Water Information System Monitoring of Mexico’s Major Dams, as of July 2024, Queretaro has 7 damns, 4 of them have reservoir levels of less than 10%, 2 damns have between 11 and 15% and only 1 has a reservoir level of 90%. [2]
Water scarcity in Queretaro has been exacerbated by population growth and an increase in housing. Queretaro stands out due to its economic and industrial development, ranking 5th place on a national level in terms of competitiveness [3]. Due to the competitive job market in the state, in the past years, many people have settled in Queretaro.
Comparing the Population and Housing Census of 2010 with that of 2020, the number of households increased by 63.45%, and the population by 47.9%. The number of households increased to a greater extent. One house was built for every 2.37 inhabitants, as shown in the image with the visualization of statistics. [4]
According to the 2022 National Survey of Household Income and Expenditures (ENIGH), 68.22% of the population in Queretaro has access to piped water and sanitation, and 68.46% have access to safely sourced drinking water and sanitation services.
In this context of water scarcity, population growth and urban development, it is important to address the issue of granting feasibilities for new connections to the hydraulic network in a more sustainable way and based on long-term planning.
Sources:
[1]https://smn.conagua.gob.mx/tools/DATA/Climatolog%C3%ADa/Sequ%C3%ADa/Monitor%20de%20sequ%C3%ADa%20en%20M%C3%A9xico/Seguimiento%20de%20Sequ%C3%ADa/MSM20240630.pdf
[2] https://sinav30.conagua.gob.mx:8080/Presas/
[3] https://www.liderempresarial.com/desafio-y-realidad-el-panorama-hidrico-y-energetico-en-queretaro/
[4] https://observatoriodeciudades.mx/blog/vivienda-y-expansion-urbana-el-caso-de-queretaro/
Qualitative evidence: What weak signals have you recently spotted that characterizes its urgency? Please provide qualitative information that better exemplifies the importance and urgency of this frontier challenge. You can add text, a link, or a picture.
We have identified signals in two high level domains: sustainable management of water resources and the use of generative AI in the public sector. We share the signals we identified below.
A) SUSTAINABLE MANAGEMENT OF WATER RESOURCES
Dams are alarmingly empty:
Due to the water crisis, Queretaro has become the third state in the country with the most dams that do not exceed 30% of their capacity. Back in May 2024, it was reported that El Centenario, La Llave, and La Venta dams were completely empty [1].
Two of the other four dams in the region barely exceeded this, as they were holding less than 5% of their total capacity.
This signal illustrates the urgency of contributing to the management of water resources in the state.
The privatization of water:
There are 23 concessions granted by the State Water Commission of Queretaro to private companies, including real estate firms for urban housing and commercial developments. Among the main concessionaires are Operadora Querétaro Moderno and Fraccionamientos Selectos I. This situation leads to discontent among the population, many of them complain about excessive prices in the water service [2] [3].
This highlights the dynamics of water in the state between private and public sectors, inviting us to think how the feasibilities process could contribute to a better planning of water resources in the future, and better decision making when it comes to deciding if the provision of water should be a private or public enterprise.
The impact of water scarcity in productive activities:
Small and medium-sized producers working in the primary sector, such as agriculture, livestock, or fishing, have seen their productivity decline due to the effects of the drought. Aquaculture producers blame low water levels for limiting fish farming. Aquaculturists in the region point out that several of their colleagues have migrated to the United States due to the income drop caused by these conditions.
Additionally, agricultural and livestock producers have also experienced reduced production due to low reservoir levels. Even if it rains, there is inadequate infrastructure for water capturing. [4] [5]
This situation showcases how various sectors are being affected by the water crisis and invites us to think how the feasibility procedure for new water connections for housing and commercial use could be managed in a more sustainable way in order to ensure access to water in the future not only for new constructions, but also for productive activities.
Another signal that illustrates this situation is related to the fact that Queretaro has become a hub for investment with the arrival of several data centers. However, it is anticipated that in the long term, energy demands may not be met. This is exacerbated by bureaucratic procedures, as companies often lack information about how the energy they use is generated. And in addition to energy, access to water to operate the data centers will also put great pressure on the water resources of the state.
Social response to government’s actions around water management:
Grassroots organizations and indigenous collectives in the region are actively involved in addressing various issues, which are highlighted below:
-Due to the "land use change voted by the council allowing 66,641 square meters settled in the area, 409 apartments will be built in four-story buildings, where a mall and services will be integrated" in the El Batán area, favoring real estate interests, environmental groups seek to designate several areas as protected zones [6].
-The Museo del Agua Bajo Tierra has published reports on the privatization of water in Querétaro and criticizes the "Ley Kuri," or concessions law, which regulates the provision of drinking water, sewage, and sanitation services, as well as unplanned real estate growth and overexploitation of water in urban areas.
-In 2022, the Caravana por la Vida mobilized several of these groups to the Otomí region in Santiago Mexquititlan and took over a well as a form of protest against excessive charges and the criminalization of activists. The Indigenous Council of Santiago Mexquititlán "also commemorated their own struggle, which began on March 31, 2021, against what they described as an attempt to dispossess their community of their well with water pipes owned by a private companies protected by state police," expressing concerns about the management of a well by the State’s Water Commission, which they affirm was built by their ancestors [7].
The examples presented above show some of the dynamics and challenges that surround the management of water resources and its linkage with land use. This highlights the involvement of various sectors in the issue, which invites us to think about possible links with their involvement in the creation of a feasibilities procedure focused on planning.
B) USE OF GENERATIVE AI IN THE PUBLIC SECTOR
Since the surge of generative AI, its adoption in the public sector has generated examples of how this technology can improve the way governments operate. And while the examples are relatively new and scarce, this technology has the potential to enhance —and even transform— the way governments operate. As shown in the image, some of the trends of using generative AI include improve speed and efficiency of public services and engagement, optimize employee time by reducing manual processes, draw insights to empower decision-makers, etc. [8].
On the other hand, the huge impact that AI systems can have in various domains of our lives call for a critical, thoughtful, imaginative and cautious approach to designing AI strategies.
As Vaugh Tan frames it: “To have good AI strategy, we need to make reasoned decisions about:
1. What humans can do which AI systems cannot — these are tasks which should be done by humans.
2. What AI systems can do much better than humans — these are tasks which AI systems should be built to do.” [9]
This becomes particularly urgent in the context of public services because these systems need to ensure human-rights, inclusion, ethics, transparency. Thus, it becomes imperative that we define pathways to design good AI strategies for deploying these systems in the public sector.
Sources:
[1] https://www.eleconomista.com.mx/estados/Principales-presas-de-Queretaro-estan-a-3.1-de-almacenamiento-20240514-0057.html
[2] https://www.jornada.com.mx/noticia/2024/01/26/estados/oligopolio-del-agua-socio-de-inmobiliarias-en-queretaro-3791
[3] https://tribunadequeretaro.com/informacion/investigaciones/ley-de-aguas-dan-concesion-disfrazada-corregidora-ya-serian-23-en-queretaro/
[4] https://www.eleconomista.com.mx/estados/Productores-agropecuarios-de-Queretaro-alertan-por-grave-crisis-de-sequia-y-escasez-de-agua-20240311-0083.html
[5] https://www.eleconomista.com.mx/estados/Sequia-afecta-a-96-de-las-perdidas-de-produccion-agropecuaria-en-Queretaro-20240319-0076.html
[6] https://noticiasdequeretaro.com.mx/2020/07/15/lamentan-activistas-cambio-de-uso-de-suelo-en-el-batan/
[7] https://losangelespress.org/mexico-de-conciencia/indigenas-en-defensa-del-agua-contra-represion-en-queretaro-20240610-8677.html
[8] https://wwps.microsoft.com/blog/services-ai-apolitical
[9] https://uncertaintymindset.substack.com/p/where-ai-wins
Value proposition: What added value or unique value proposition is your Accelerator Lab bringing to solving this challenge? Why is it your Lab that needs to work on this challenge and not other actors within UNDP, other stakeholders in the country respectively? Why is it worth investing resources to this challenge?
The Accelerator Lab adds unique value by operating in a practice-oriented, experimental mode—designing and testing proof-of-concept solutions in real-world conditions. Building on prior systemic analysis, the Lab will apply agile methodologies to prototype a AI tool and generate rapid learnings to inform future iterations. This approach enables low-risk exploration of innovative pathways and supports public institutions in making more informed, data-driven decisions.
At the same time, the Lab’s convening power and participatory methods foster collaboration across institutional and sectoral boundaries. In a context where UNDP Mexico is still building its profile in AI and digitalization, this work positions the Country Office as a credible partner for inclusive and sustainable tech adoption. The Lab’s connection to a global learning network further amplifies its impact, allowing it to share insights and draw from international experiences.
Short “tweet” summary: We would like to tweet what you are working on, can you summarize your challenge in a maximum of 280 characters?
AccLab Mexico is exploring how generative AI can help public officials review water connection requests more efficiently—boosting accuracy, breaking silos, and enabling anticipatory planning for sustainable water infrastructure.
Partners
Who are your top 5 partners for this challenge? Please submit from MOST to LEAST important and state Name, Sector and a brief description of the (intended) collaboration.
Please state the name of the partner:
State Water Commission of Queretaro (Comisión Estatal de Aguas de Querétaro)
What sector does our partner belong to?
Government (&related)
Please provide a brief description of the collaboration.
We will test a solution based on generative AI in one of the processes implemented by the Commission to manage access to water resources in the state. They are providing us with access to their information, data and personnel to conduct the learning cycle. Furthermore, it is expected that they will adopt the solution we test and scale it afterwards on their own.
Is this a new and unusual partner for UNDP?
Yes
Who are your top 5 partners for this challenge? Please submit from MOST to LEAST important and state Name, Sector and a brief description of the (intended) collaboration.
Please state the name of the partner:
Ministry of Finance of Queretaro (Secretaría de Finanzas de Querétaro)
What sector does our partner belong to?
Government (&related)
Please provide a brief description of the collaboration.
They are legally responsible for the Digital Strategy of the State of Queretaro. They oversee the design and roll out of digital solutions in all the ministries and government entities of the state. They are our key liason in the government and help us reach out to local actors, access resources to implement our learning cycle (e.g. venues, data, events) and connect our pilot to the state’s strategy to ensure its adoption.
Is this a new and unusual partner for UNDP?
Yes
Who are your top 5 partners for this challenge? Please submit from MOST to LEAST important and state Name, Sector and a brief description of the (intended) collaboration.
Please state the name of the partner:
National Autonomous University of Mexico Juriquilla Campus (Universidad Nacional Autónoma de México Campus Juriquilla)
What sector does our partner belong to?
Academia
Please provide a brief description of the collaboration.
They act as a knowledge partner and may provide their input along the process of our learning cycle.
Is this a new and unusual partner for UNDP?
Yes
Who are your top 5 partners for this challenge? Please submit from MOST to LEAST important and state Name, Sector and a brief description of the (intended) collaboration.
Please state the name of the partner:
Mottum
What sector does our partner belong to?
Private Sector
Please provide a brief description of the collaboration.
They are our partner for developing the technological solution we will test.
Is this a new and unusual partner for UNDP?
Yes
Learning questions
Learning question: What is your learning question for this challenge? What do you need to know or understand to work on your challenge statement?
What are the faced by public officials in reviewing requests and determining the feasibility of new connections to the water network that could be addressed through a generative artificial intelligence tool?
To what stage(s) in the learning cycle does your learning question relate?
Explore
Usage of methods: Relating to your choice above, how will you use your methods & tools for this learning question? What value do these add in answering your learning question?
By using the selected method, we will gain a clearer understanding of the challenges faced by public officials across the State Water Commission when analyzing the feasibility of new water connection requests. Rapid ethnography will help us identify specific pain points within the current process, enabling us to develop a targeted theory of change that will guide the design, development, and testing of an AI tool.
Existing data gaps: Relating to your choice above, what existing gaps in data or information do these new sources of data addressing? What value do these add in answering your learning question?
These data sources will enable us to clearly identify and analyze the core challenges faced by public officials when assessing applications for new water connections. This thorough understanding will help us establish a robust baseline against which we can measure the impact of the AI tool in the test stage.
Learning questions
Learning question: What is your learning question for this challenge? What do you need to know or understand to work on your challenge statement?
How can a generative artificial intelligence tool improve agility, accuracy, and anticipatory capacity in the technical analysis of requests for new connections to the water network?
Which activities within the water balance estimation analysis for granting new connections to the water network—such as report writing, detection of inconsistencies, or data cross-checking—can be complemented by generative AI systems, and which require human intervention?
What ethical considerations should be taken into account when designing and testing a generative artificial intelligence solution in the provision of public services for water management?
What effects does the use of a generative artificial intelligence tool have on reducing review times, decreasing errors, and/or generating more comprehensive analyses across areas?
*Note: At the conclusion of the final exploration question, the learning questions for the testing stage will be refined.
To what stage(s) in the learning cycle does your learning question relate?
Test
Usage of methods: Relating to your choice above, how will you use your methods & tools for this learning question? What value do these add in answering your learning question?
The chosen methods will allow us to: (1) design and test a solution that includes the perspectives of the various actors involved throughout the water feasibility process, thus increasing the probability of creating a relevant solution that can then be adopted after the pilot; (2) test an AI proof of concept in order to learn how to implement human-centered AI systems that can support public officials in their work; (3) collect data to learn how the proof of concept works and incorporate the feedback of public officials into a future iteration.
Existing data gaps: Relating to your choice above, what existing gaps in data or information do these new sources of data addressing? What value do these add in answering your learning question?
The data sources will help us to: (1) create a proof of concept based on generative AI so we can test it; (2) create a baseline to kick off our test and identify the changes produced by the solution after the test stage, (3) provide recommendations to iterate the solution based on users’ feedback. The data sources will help us to derive learnings from the pilot that can help the State Water Commission in Queretaro scale the solution we test. Furthermore, the data we generate will be helpful for other State Water Commissions who want to undertake similar efforts.
Learning questions
Learning question: What is your learning question for this challenge? What do you need to know or understand to work on your challenge statement?
What learnings does the proof of concept generate that allow scaling a solution based on generative AI to strengthen water management sustainably in Querétaro?
What institutional conditions facilitate or limit the adoption of the artificial intelligence tool within the State Water Commission?
To what stage(s) in the learning cycle does your learning question relate?
Grow
Usage of methods: Relating to your choice above, how will you use your methods & tools for this learning question? What value do these add in answering your learning question?
By employing this method, we will gain deeper insights into stakeholders' perspectives, enabling us to develop a robust and scalable plan for implementing an artificial intelligence tool.
Existing data gaps: Relating to your choice above, what existing gaps in data or information do these new sources of data addressing? What value do these add in answering your learning question?
These data sources will support the development of a clear scaling plan, enabling the State Water Commission in Querétaro to refine and expand an AI tool that strengthens their technical teams’ capabilities.
Closing
Early leads to grow: Think about the possible grow phase for this challenge - who might benefit from your work on this challenge or who might be the champions in your country that you should inform or collaborate with early on to help you grow this challenge?
Governments currently implementing or interested in applying AI to public service delivery—especially agencies responsible for water resource management—stand to benefit from this work. Additionally, regulatory bodies overseeing AI and those involved in establishing regulatory sandboxes may find value in these insights. NGOs and academic institutions dedicated to advancing inclusive AI technologies for public sector innovation, as well as private sector organizations focused on AI applications in government, are also key stakeholders to engage early in the growth phase.
END OF ACTION LEARNING PLAN: Thank you! The form saves automatically and your submission has been recorded. You may now exit this window.
6Clean water and sanitation
11Sustainable cities and communities
Comments
Log in to add a comment or reply.