Menu Close

Extending A2A: A Human-Aligned Proposal for Civic Collaboration

How can we learn from the tools emerging in AI—especially structured protocols like A2A—and apply those lessons to strengthen the systems we humans already run: our civic programs, public institutions, mutual aid networks, and local services?

What follows is an overview of the A2A Human-Aligned Extension, a proposal to extend Google’s Agent-to-Agent (A2A) protocol by introducing structures that reflect the roles, responsibilities, and character of human organization agents—since many of the same coordination challenges appear across both human and digital ecosystems.

We see value in applying lessons from and contributing to the evolution of protocols like A2A to support shared understanding and structured engagement among human organizations—without prescribing how they must operate.

We invite feedback from technologists, civic leaders, policy researchers, and community members.


What This Extension Proposes to Add

The A2A Human-Aligned Extension GitHub repository proposes to add structured metadata and semantic framing elements designed to make human programs and institutions more legible, discoverable, and interoperable within broader agent-based systems:

  • ProgramCard — a structure for representing real-world programs with goals, offerings, and contextual metadata.
  • ParticipationTemplate — a structured way to describe how people or organizations can engage, contribute, or replicate a program.
  • operating_character — a metadata block to declare values, governance tendencies, and communication styles, using frameworks like Spiral Dynamics, Hofstede’s cultural dimensions theory, or the Discovery Insights communication framework.

These proposed additions build on A2A’s foundation, extending its utility beyond digital agents by offering structured, interpretable descriptions of human-centered work.


Why This Matters

There remains a widespread absence of shared infrastructure for describing community-benefiting efforts—such as civic services, nonprofit initiatives, and mutual aid programs—in ways that support cross-organizational understanding, volunteer engagement, and clarity for funders and donors. By adopting a common structure:

  • Programs become more discoverable and transparent.
  • Participation becomes easier to invite, structure, and sustain.
  • Diverse actors can self-align around shared goals without requiring centralized orchestration.

The aim is to draw on the lessons of structured interoperability from the AI world—where protocols are explicit and interfaces are designed for clarity—to support civic and social initiatives in representing themselves to peers, participants, funders, and digital systems with similar precision and accessibility, without reducing their human character.


Read and Respond

You can read the full draft specification here:
👉 Extending-A2A-Human.md

We welcome feedback, critique, and discussion.


How to Contribute

We are currently seeking:

  • Feedback on draft conceptual models and proposed metadata structures
  • Suggestions for civic or nonprofit pilot applications
  • Collaborations with public interest technologists and community orgs

This is a versioned, evolving draft (v0.1.3 as of June 2025). Comments and issues can be submitted via GitHub:
👉 github.com/team-earth/a2a-human-extension


Kevin Kells
Project Lead, Team Earth, GOSR
team.earth/contact

Unsolvable: Think Again! Now Open — Tools and Data to Map Complex Problems

We’re excited to announce that the code used to create the systems maps in the Unsolvable: Think Again! book series is now available as an open-source toolkit.

👉 Explore the code here: github.com/team-earth/gosr-ai-workflow
👉 Browse the datasets here: github.com/team-earth/data


🧭 What This Is

Over the past few years, we’ve developed GOSR, a structured method to map complex societal problems — and the solutions already underway to address them. These maps have powered more than 10 published books across domains like:

  • Urban and rural loneliness
  • Mental health and addiction
  • Education reform
  • Climate change adaptation
  • Food insecurity
  • Civic resilience against extremism
  • Peacebuilding in entrenched conflicts

Now, the AI-assisted workflows behind those datasets are publicly available for anyone to reuse, extend, or build on.


🛠 What the Code Does – GOSR

The gosr-ai-workflow repo provides tools for building structured, modular solution maps from source content — using a flexible, repeatable schema:

  1. 🎯 Goal – What success looks like
  2. 🪨 Obstacles – High-level systemic barriers
  3. ⛓️ Sub-Obstacles – Detailed, grounded challenges
  4. 💡 Solutions – Concrete actions or ideas
  5. 🧭 Resources – Real-world organizations or initiatives

This schema allows you to create structured maps of civic problems — ready to feed into search engines, AI agents, or collaborative planning tools.


📘 10-Book Series: Now Fully Indexed

Each dataset aligns with a specific book in the Unsolvable: Think Again! series. You can now find structured metadata for all 10 books, including tags, summaries, publication dates, and direct dataset links.

The full list is indexed in metadata.yaml and embedded on kevinkells.com.


👥 Who This Is For

We designed the workflow and data to serve a range of audiences:

  • 🔧 Civic tech developers – Build search tools, visualizations, or AI interfaces on top of structured civic maps
  • 📊 Policy researchers & systems thinkers – Study patterns across domains, or contribute new use cases
  • 🌱 Social innovation labs – Use maps to surface local assets and align collaborative efforts
  • 🎓 Educators – Use books + maps in courses on systems thinking, design, or social innovation
  • 🧭 Community organizers – Create maps that help locals find “who’s doing what”
  • 🤖 AI builders – Leverage structured knowledge to guide people through complex challenges
  • 🏛️ Cities & funders – See the landscape of active work and where gaps or leverage points exist
  • 📚 Readers – Understand how each book was built, and how you can use or extend the ideas

🚀 What You Can Do Now

  1. Explore the data:
    Start with the Un-Lonely New York City map or browse the full dataset repo.
  2. Run the workflow:
    Clone the gosr-ai-workflow repo to generate your own maps from new topics or documents.
  3. Read the books:
    Unsolvable: Think Again! on Amazon gives story-driven context behind each dataset.
  4. Join the process:
    Suggest topics, contribute maps, or adapt the tools for your own work. We welcome collaborators across all fields and sectors.

Un-Lonely New York City: A Systems-Based Framework for Addressing Urban Isolation

Urban loneliness is a multifaceted challenge, influenced by factors ranging from architectural design to socioeconomic disparities. Recognizing the complexity of this issue, Un-Lonely New York City offers a comprehensive, systems-thinking approach to understand and mitigate social isolation in urban environments.

A Structured Approach to a Complex Problem

The initiative employs a radial hierarchy framework to dissect the problem systematically called GOSR:

  • Goal: Alleviate loneliness in New York City.
  • Primary Obstacles: Identifying major themes contributing to loneliness, such as lack of community spaces or digital exclusion.
  • Sub-Obstacles: Delving deeper into specific issues like the scarcity of “third places” or limited access to technology.
  • Proposed Solutions: Suggesting actionable strategies to address each sub-obstacle.
  • Existing Resources: Highlighting organizations and programs already working toward these solutions.

This structured methodology ensures a comprehensive understanding of the problem and facilitates targeted interventions.

Resources for Engagement and Implementation

The Un-Lonely New York City project provides several tools at its github page to support stakeholders:

  • PDF Guide – A detailed outline of the problem’s structure and potential solutions.
  • JSON Dataset – A machine-readable format designed for integration into AI applications, enabling dynamic analysis and application development.
  • Interactive Google Map – A visual representation of over 10,000 community resources across NYC.

Why the JSON File Matters

The .json file isn’t just for developers—it’s also designed to be readable by modern artificial intelligence systems. You can upload the file to an AI assistant (such as ChatGPT with file browsing enabled) to explore the structure interactively, ask specific questions, or generate new visualizations and summaries.

This feature empowers both humans and machines to navigate complex, systems-level data in a way that’s contextual, adaptive, and collaborative.

Collaborative Effort and Broader Implications

This initiative aligns with the principles advocated by Kevin Kells and The Unsolvable Project, emphasizing the importance of structured, collaborative approaches to complex societal issues. By providing a clear framework and accessible resources, Un-Lonely New York City serves as a model for addressing similar challenges in other urban contexts.

Engage and Contribute

Stakeholders are encouraged to utilize these resources, provide feedback, and contribute to the ongoing refinement of the project. For collaboration or inquiries, please use our contact form.



By adopting a systems-thinking approach, Un-Lonely New York City offers a replicable model for understanding and addressing the complex issue of urban loneliness, fostering collaboration and informed action across sectors.

Computer-Aided Design (CAD) of Complex, Human Problem-solving Models

There exists a class of complex, human problems (“CP2”) that manifest in the social, political, and environmental domains. These problems are unique in that they have multiple owners spread across multiple organizations; trust among stakeholders is required – and often absent – for the solution of these problems. The costs, subtleties, and sheer breadth of tasks needed to resolve these problems is unlikely to fit into a single human mind, and unlikely to be clearly represented by modern media – including multimedia documents like Google Docs, brainstorming applications like MindJet, and project-management software like Asana.

Three examples of CP2 problems follow:

  1. Race relations between the citizens and police in cities like Baltimore, Maryland;
  2. High rates of violent crime in Chicago; and
  3. Poverty and opioid misuse in the US Rust Belt.

Unlike class-one complex problems (CP1) that came before them and found adequate solution approaches, such as operations optimization, financial engineering, and battlefield command and logistics, CP2 problems require a sophisticated visual representation meeting the following criteria (C2), suited to aligning multiple stakeholders:

  1. The superordinate goal is clear;
  2. All participants, irrespective of their initial trust in one another, believe that their angle on the problem has been represented and will be acted upon;
  3. All obstacles to the superordinate goal have been captured, decomposed into actionable units, and suitable next actions have been proposed.

In short, CP2 represent the critical problems of our time, problems that will remain unsolved with current frameworks and systems because the need to inspire and align and convince stakeholders to have trust in a solution model is not possible with texts and static images… and Post-It notes. If we are to overcome the current challenges in race relations, violence, political polarization, environmental adaptation, and poverty, we will require a fresh set of tools, techniques, and systems that meet C2.

We are proposing a new branch of Computer-Aided Design (CAD) for the software-assisted design and implementation of complex, human problem-solving models.

Link to the working paper…
A Proposed Practical Problem-Solving Framework
for Multi-Stakeholder Initiatives in Socio-Ecological Systems
Based on a Model of the Human Cognitive Problem-Solving Process

Importance of Visualization Technology in Solving Large-scale, Complex Human Challenges

Omegalib: a Multi-View Application Framework for Hybrid Reality Display Environments (https://www.evl.uic.edu/entry.php?id=1644)

The key to a successful approach to large scale, complex human challenges is to present all stakeholders a data-driven model visualization. This is a visualization not just of historical data, and not even of real-time present data. What does that leave, predictive data visualization? It’s more…

It is a combination that starts with an aspirational visualization — the superordinate goal or future picture — a way to represent what stakeholders and community have agreed upon is the future state of their system, the future they wish to achieve, and to capture the emotion and the longing to achieve it, and the common purpose that has been revealed in identifying that future picture. It is what they agree things will will look like once the problem-solving work is complete.

In addition, the visualization also represents the hindrances and obstacles the stakeholders and community have agreed are hindering progress towards that future picture. It visualizes the dissection of these obstacles areas into many, many individual components that have simple solutions. And it visualizes the progress being made on each and every simple solution being implemented by the myriad resources who have joined in to contribute towards achieving the future picture.

It is the visualization of the future picture, of the obstacles, of the solutions to the simplified problems, of the engagement of resources, and of the performance of the solution implementations.

As facilitators, we call this the visualization of the “problem-solving model.”

(CAVE) technology at INL's Center for Advanced Energy StudiesThis visualization technology has the following characteristics:

Immersive

More powerful than viewing the model — vision, hindrances, solutions, resource engagement, and implementation performance — is to step into it and experience it from all sides, and with all senses in 3D — vision, sound, tactile.

AlloSphere Research Facility (http://www.allosphere.ucsb.edu/)

Enduring

The model, represented at exhibits in the city and on the web, remains throughout the lifetime of the effort, in order to maintain continual alignment, keep alive the memory of the energy expended by the stakeholders to define the model, and to keep the confidence of all stakeholders and resources.

Dynamic

CAVE2 at Monash University

The model exhibit and all its representations — on the web, in classrooms, at the town hall, in the central park, in the libraries — should adjust to represent actual data values. This means graphical representation of progress made, as well as adaptation to changes over time in the resource engagement, solutions, hindrances, and even vision.

Impactful

The visualization and exhibition of the model is stunning in size and beauty, but also in symbolism and in its power garnered by its placement in locations of authority.

Publicly Accessible

CAVE2 electronic visualization laboratory

Anybody can access the visualization at anytime, aided by representations of the same model via the web and in larger scale through different exhibits.

Navigable and Interactive

The model is interactive and navigable via mouse, touch screen, and handicap accessibility.

Pricing the Units of Philanthropic Work

In his 2018 book, “Bloomberg by Bloomberg,” former New York City Mayor and founder/CEO of Bloomberg LP, Michael Bloomberg, explains how his user-friendly invention of accurate, technology-assisted valuation of financial instruments became a breakthrough insight that made markets:

“When it came to knowing the relative value of one security versus another, most of Wall Street in 1981 had pretty much remained where it was when I began as a clerk back in the mid-1960s: a bunch of guys using No. 2 pencils, chronicling the seat-of-the-pants guesses of too many bored traders.”

Michael Bloomberg saw the opportunity to leverage technology to generate valuations in a way that nobody had previously conceived.

What the Bloomberg terminal delivered to finance, we propose to do for  charitable and humanitarian work  through the multi-step, technology-assisted approach to solving large-scale, complex human challenges we describe in these pages.

This requires re-framing charitable effort in terms of “what we, the stakeholders, want the outcome to look like” rather than in terms of which outcomes charitable entities are capable of providing. With the assistance of compelling visualization technology and a disciplined and indefatigable facilitation effort, the establishment of the superordinate goal, and then the meticulous stakeholder-driven analysis of the hindrances and obstacles to that goal, leads to the straightforward definition of “what needs to be done,” in a way that engages the stakeholders at every step in its construction.

By guiding the stakeholders to formulate their challenge from this perspective, the resulting problem description provides a natural framework for valuating effort on a per-dollar basis. Stakeholders describe their vision, their understanding of the problems broken down into finer and finer simplicity, and their weightings of the importance of the myriad solutions they’ve proposed to solve those more numerous but simplified problems. The simple performance measurement of each solution effort readily provides a measurement of how much the “needle” is moved — and in the right direction — for each philanthropic dollar spent.

When an accurate and unbiased valuation can be made of a commercial security, markets open up, investors act with confidence, commerce grows, and people prosper financially.

When an accurate and unbiased valuation of charitable effort becomes available, philanthropic marketplaces open up and become democratized, both donors and non-profit actors align themselves by “market forces” to do what works, and for so many of today’s seemingly intractable problems, the horizon of solutions comes into view.

Human Problem Solving

Human Problem Solving…

The human mind follows these steps when solving problems individually. With some help from technology, we can facilitate these steps, too, when we solve problem collectively.

Step 1: Future Pull

We know where things stand now, and we have a vision for how things should be — how we want things to be. That vision of how things should be can be a very powerful driving force.

Things should be like…

Step 2: Hindrances

We form an idea of what issues, problems, or obstacles we think are getting in the way of achieving the future state we envisions in Step 1.

The problem is…

Step 3: Solutions

We identify solutions to the hindrances in Step 2.

What we need to do is…

Step 4: Resources

We concern ourselves with how to obtain the needed resources (e.g., time, money, people) so that the solutions we’ve identified in Step 3 can be implemented.

How to pay for this?

Step 5: Performance

We monitor and improve the implementation of solutions from Step 3 by resources from Step 4.

How to work better?

Key is: Don’t Skip Steps

Mayoral Toolkit for large scale, complex human challenges at city level

City-level challenges often share the characteristics that the team.earth practice and technology tools aim to target:

  • Widespread public opinion exists that there is a problem that needs to be solved;
  • There are so many interconnected moving parts that there is otherwise no clear solution or even an agreement on “where to start” to solve the problem;
  • There is a critical mass of stakeholders willing to participate in a process if they trust it will lead to an improvement from their perspective;
  • Sufficient resources could be mobilized towards the solution if only a credible, fact-based plan existed how to employ each resource, endorsed by all the stakeholders.

We propose a Mayoral Toolkit for engaging large-scale, complex human problems in our cities.

This includes a five step process (1 2 3 4 5) that combines the psychology of “future pull” and the analytic approach of “value engineering” to engage all stakeholders in a process to develop a plan for what we need to do to move the needle in the right direction. It includes visually-oriented technology proposed by team.earth to engage the imagination of all stakeholders, helping each perspective contribute to the practical construction of a practical plan that offers specific guidance about what to do and where to start.

The team.earth mayoral toolkit. Visual problem solving technology and practice—to empower leaders at the city level to engage their citizens to sustainably solve large scale, complex human problems.

Social ROI and the Philanthropy Brokerage

You have to define the needle in
order to move it. There’s more.
You have to define both
“where we are now” and also
“where we want the needle to be.”
The per-dollar movement of the
needle is called the Social Return
on Investment or sROI.

In today’s world of nonprofit work—with philanthropy and volunteering making it possible—NGOs need to make the self-interested case to donors that their cause is worthy, and that they are effective and trustworthy in implementing the work. NGOs find themselves in need of “tooting their own horn” by putting the best photos on their websites and making the best possible annual reports of their results. Though there do exist some third-party efforts to audit the efficiency of NGO use of donations versus their overhead, there’s no clear way to audit effectiveness. Measuring effectiveness requires a definition of the goal state the NGO is working toward, and this definition cannot be defined by the NGO alone, but by all the stakeholders. Furthermore, for large-scale, complex challenges, the measurement of the contribution of a single NGO in an uncoordinated pool of efforts would not lead to a meaningful measure.

Meanwhile, in the world of for-profit investing, there is:

  1. An industry whose purpose is to rate and present investment opportunities to customers, and
  2. Government regulation to ensure sufficient honesty and transparency that an efficient marketplace can exist, in the name of the public good, i.e., for facilitating economic growth, but also to prevent fraud committed against individual investors.

Being the realm of for-profit, this industry is financed by brokerage fees.

There doesn’t exist a similar, thriving industry for nonprofit work, and there is poor financial incentive for there to be one. How would a brokerage house for nonprofits benefit problem solving efforts by receiving a fee paid from the donations meant to fund the solutions? Donors would rather all of their donation go to the charity in need, and not partly to pay a middleman whose ability to provide a benefit for the add-on service is uncertain.

The solution in the team.earth proposal works by creating a measure of the social good for each dollar of input applied to solving the challenge. We call this the Social Return on Investment or sROI. Instead of being focused on the NGOs, the problem solvers, we propose to define the needle and where it needs to move, working with stakeholders to ensure an accurate, authentic, and native understanding of how the denizens and participants in that system see their future.

Philanthropists are looking to
donate wisely. Worthy non-profits
are looking for donors.
Challenges are looking for
effective solutions.
Team Earth wants to make sure
Challenges—for lack of clarity
and definition—are no longer
under-represented in this equation.

Such a nonprofit or philanthropic brokerage would not list in its portfolio each nonprofit separately with its social benefits; it is hard to quantify in terms of social good in isolation from the challenge they are helping solve. Instead, the portfolio is a listing of different superordinate goals and the team.earth description of the challenge, functional solutions, and the resources that have come to the table to participate. In this way, philanthropists can decide which challenges they wish to help support solving. In the listing of resource participants, a nonprofit organization can appear under more than one superordinate goal, as part of many different problem descriptions.

The participation of the NGO in a particular challenge solution is guaranteed by construction to be effective. And the operational efficiency can be more easily and clearly defined because the work to be done is better and more simply defined.

In previous blog posts, we’ve described a process that begins with the stakeholders and delivers a set of functional solutions:

  1. Facilitating the superordinate goal statement
  2. Categorizing and scrutinizing the hindrances to the superordinate goal
  3. Functional solutions

The contribution of each functional solution lends itself to an estimate of how much that solution contributes to moving the needle in the right direction. It is possible for stakeholders to estimate the proportional contribution, how much each leaf node contributes towards needle movement as an approximation.

The rating or assessment of an NGO can be done on a per-challenge or per-NGO basis. It is judged by an independent assessment and not by the NGO itself in the context of the given challenge and the known functional solutions. The progress of the functional solution implementation, multiplied by the contribution of that functional solution to the solution component of that leaf node, multiplied by the contribution of that leaf node to the overall movement of the needle forms the numerator of the sROI. The denominator is comprised of the funding used by the NGO or resource to implement its contribution to the solution. The overhead needed by the entity is included in this calculation to reach an overall, gross sROI. The efficient use of funds, i.e., the ratio of overhead to effectively deployed donation capital, can still be assessed separately.

In the same way that a financial brokerage offers different packages to meet the needs of different investors, the philanthropy brokerage that team.earth conceptualizes can offer different packages for social investment, categorized by

  • Challenge type
  • Challenge specifics
  • Resource type
  • Solution type
  • NGO
  • Geographical location
  • etc.

Unlike the application of a brokerage concept to a single or small handful of NGOs, in the team.earth concept, the large number of challenges and their corresponding solutions and resource providers makes it viable to enlist the paid help of a brokerage service. This would defray the cost of:

  • Carrying out the team.earth process to obtain a superordinate goal statement and problem description visualization.
  • Development of technology for coordination and visual communication.
  • Management and measurement and coordination of resource entities implementing functional solutions.
  • Providing a single point contact for the solution efforts.
  • Single point of reliable and trustworthy information on the project and measurement of the current needle and maintenance of the goal.
  • Public relations and necessary communication for non participants for the good of the solution efforts.
  • Communication among participants.
  • Hiring HR and staff to carry out these functions.
  • Research into improving team.earth methodology between different challenge solutions.

The donor or philanthropist benefits from a reliable and trustworthy point of contact where they can obtain independent assessment of what is working. They no longer have to rely on self-reporting, e.g., on each NGO’s website.

For the NGO, it gives an opportunity to come into contact with donors, as the brokerage becomes a main funnel for qualified donors and philanthropists. The public relations work is done for the challenge, not for the NGO, but the participation of the NGO is a part of that public relations work, and a rating of that NGO’s work is included—their sROI with respect to that challenge.

This sets up a system of incentives:

  • NGOs align themselves to what is needed to move the needle based on a broader, more reliable definition of what needs to be done, a definition that is achieved through active participation of the stakeholders. Instead of saying, “We’re doing X because X is what we do best,” the NGO can now say, “We’re doing X because the problem description called for X as a functional solution that was needed to move the needle in the right direction.”
  • The NGOs want to increase their sROI to attract more funding now that their sROI, which includes the effective and efficient use of money to move the needle, is presented plainly and transparently and without risk of bias.
  • Philanthropists have the incentive to seek out brokerages to help them find the best way to make a difference with their contributions—how most effectively to move the needle in the right direction.

Philanthropy Brokerage Benefits

  • If you have a reliable measure of sROI, you can catalog and present these efforts as social investments or challenge-solving investments just as you would a financial investment.
  • You can offer donors and philanthropists different portfolios to invest in.
  • You can specify what the historical sROI has been as a number, and can thus easily present past performance.
  • You can tailor investment based on risk, e.g., high-risk investment for new efforts, and low risk for challenges with a proven track record of moving the needle.

Managing solutions—coordinating, measuring, and visualizing (Step 5 of 5)

In the previous steps of the team.earth stakeholder process, we transformed a large-scale challenge made up of a small number of massively complex components into a massive number of small, simple components. We have a list of leaf node issues and their straightforward functional solutions, so a list of things that need to be done. This list by construction, by design: If you implement any one of those functional solutions, you will be contributing to moving the needle, and moving it in the right direction, in the direction all the stakeholders agreed on that they wanted to achieve at the outset.