1 of 33

EOS Responsible Use AI Strategy

North Star Framework

2 of 33

What informs the work

After analyzing EOS’ existing value propositions, mission, vision, and internal focus group data, the following north star framework is designed to guide EOS in continuing to ground its work in increasing student participation in advanced courses,

improving school community, and setting up better post-secondary outcomes.

This framework serves as both a goal and a guardrail, ensuring that EOS’s innovations remain firmly focused on creating transformative, student-centered change.

3 of 33

Key Assumptions of EOS’ Work

The Role of Data and Building a Holistic Data Culture in Schools

Collecting and highlighting data provides both social capital and measurable accountability within complex systems. EOS supports schools by collecting and offering access to holistic data alongside district-provided data, enabling data-informed practices and policies that create student-centered environments, address critical barriers, shift adult mindsets, and expand rigorous academic opportunities.

The Role of Relationships and Culture Change in Fostering Student Belonging�& Agency

Relationships are essential to student success and culture change that benefits all students. For students to realize their potential, they need a supportive environment with trusted adults invested in their success both in and outside the classroom. With these data supports, adults play a crucial role in fostering this environment, and EOS equips schools with data, practices, and tools that systematically enable student potential, agency, and voice.

4 of 33

Essential Elements of EOS’ Work

The tree contains three essential elements that characterize EOS’ work. Fruit, branches, and roots should not only be present but also be aligned with each other when future work is considered.

MOST IMPORTANT

MOST VISIBLE

Fruit

What people can easily observe about EOS’ work and what may initially draw partners and funders to EOS including the outcomes, marketable features, and incentivized metrics of the work.

Branches

The means and methods of how EOS does its work.

Roots

What’s above the surface and easily visible is only a small part�of EOS’s work. EOS’ strives to create systems that acknowledge, support, nurture, and grow the talent and genius�of students.

5 of 33

What EOS Does Currently

Examples of EOS’ current work and how it maps to the essential elements of the tree, which offers a framework for understanding EOS’s work

MOST IMPORTANT

MOST VISIBLE

Fruit

  • Advanced Course Enrollment
  • Identification of Trusted Adults
  • School & District Policy Changes
  • Actionable & Holistic Student Data

Roots

  • Institutional Culture Change
  • Adult Mindset Changes
  • Systemic Accountability
  • Equitable Data Infrastructure
  • Recognition of Student Potential
  • Elimination of Systemic Barriers
  • Increased Student Sense of Belonging

Branches

  • Data Collection and Analysis
  • Student Insight Cards
  • EOS Data Portal
  • Professional Learning & Coaching
  • Partnerships with Schools

6 of 33

Care and Maintenance Considerations for Growth

As EOS grows to expand its work outside of advanced course enrollment, there will need to be intentional thought and care to ensure healthy growth that is aligned to the north star of EOS. The following slides explore the ways EOS can think about different types of growth and key questions for EOS to consider as the weigh new growth opportunities.

Deep Roots: Mission-Driven Growth

Branching Out: Tools-Driven Growth

Bearing Fruit: Stakeholder-Driven Growth

Questions for Deciding How & When to Grow

7 of 33

Deep Roots: Mission-Driven Growth

At the core of EOS’ mission-driven work is a commitment to address longstanding and emerging factors in the institutional systems that inhibit student access to rigorous opportunities. Deep roots work specifically addresses misconceptions and barriers to student potential.

  • How does this potential work enable institutional sufficiency rather than institutional dependency?
  • In what ways does this innovation disrupt prevailing narratives about students' potential to create increased academic opportunities?
  • How does this innovation address systemic or institutional factors rather than focusing solely on individual perspectives or values?

Intended to drive action that impacts student trajectories, EOS leveraged their mission to create a data-driven method for understanding the conditions that create—or inhibit—belonging for students, and what they can do to design classroom environments that enable academic belonging and success.

Key Questions

In Practice: Measures that Matter

8 of 33

Branching Out: Tools-Driven Growth

Data collection, analysis, and accessibility are the hallmarks of EOS’ unique offerings. By increasing access to data and analysis tools, EOS can illuminate new dimensions of the student experience as well as deeper insights into institutional and instructional performance.

  • How if at all is EOS uniquely equipped to add value in this problem space?
  • Does the addition of our tools in this problem space deter us from our core mission’s work?
  • What, if any additional capacity or skills will be needed in order to address this problem space effectively?

With access to a uncommonly large collection of student data, EOS has the potential to partner with other organizations to leverage its data expertise and uncover systemic issues driving chronic absenteeism.

Key Questions

In Practice: Chronic Absenteeism

9 of 33

Bearing Fruit: Stakeholder-Driven Growth

Given the varying layers of EOS’s work, it’s important to consider what is most visible and attractive as a value-add to school districts. Meeting the evolving and shifting interests of students, instructors, districts, and funders presents a valuable opportunity for innovation.

  • What is potentially gained by meeting these emerging or new needs? What is lost or diluted?
  • How, if at all, are we uniquely positioned to do this work?
  • How and for whom’s benefit is this innovation?

As the interest and momentum grows around leveraging EOS’ tools earlier in a student’s learning journey, EOS has begun to discuss how they can engage with middle schools as another partner offering.

Key Questions

In Practice: Middle School Acceleration

10 of 33

North Star Alignment

  • Is there alignment with the north star?
  • Are all three elements (fruit, branches, and roots) present already in the work, or will there be a�need to further growth to ensure there is?
  • Is there alignment between the new proposed work and the existing fruit, branches, and roots?

Questions for Deciding How & When to Grow

11 of 33

Questions for Deciding How & When to Grow

Organizational Coherence

  • Does this innovation align with or shift EOS’ core mission?
  • Will this work strain or alleviate organizational capacity?
  • Does the outcome align with our tools, resources,�and strengths?
  • Does this enhance or intersect with other priorities?
  • What are the tangible results, and is the effort�worth the impact?
  • What new needs or dependencies does this�create, and are they sustainable?
  • What are the potential impacts, both intended�and unintended, on our goals?
  • How does this coincide with stakeholders�and partners’ priorities and goals?

12 of 33

EOS Mission-Driven

AI Strategy

Appendix

13 of 33

North Star Alignment Scoresheet: Evaluating Futures

Work being considered

Type of growth: Root, Branches, or Fruit

High, Middle, or Low alignment with “Deep Roots,” “Branching Out,” or “Bearing Fruit” key questions

High, Middle, Low alignment with north star and organizational coherence questions

Recommendation to pursue further: Yes or No

Ex: Exploring inclusion of data around chronic absenteeism

Fruit

Middle

High

Yes, has alignment with EOS north star

tktk

14 of 33

EOS

Mission-Driven

AI Strategy Guide

15 of 33

The EOS AI strategy guide lays out a process to ensure that future technology and AI innovations are intentionally designed and advance the mission of EOS. Intentional Futures and EOS designed the guide for use primarily by EOS Tech & Strategy leadership, specifically for the process of building technology or technological integrations from conception to launch. The principles herein, however, have implications for larger programmatic or organizational choices, or development of future public goods.

Too often, organizations develop technology with the plan to mitigate bias on the backend, or to engage students and educators once it’s already time for launch. At EOS, we are committed to starting with students, practitioners, and those who will be most affected by the technology to help shape the product throughout its development.

This guide is intended to provide both scaffolding and flexibility so that EOS staff and stakeholders can engage, identify risks and opportunities, and weave in mission-driven practices from the start. It also provides tools, processes, bias risks and actions, and off-ramp moments to support a transparent technology development process and ensure that AI and technology ultimately centers schools and students.

Welcome to the EOS AI Strategy Guide!

16 of 33

The AI Strategy Guide was developed as just that: �a guide. Product and program development is always a living, iterative process. Your organization, the educational systems you serve, and the world itself will inevitably change, and your process should be flexible enough to accommodate those changes. Use your discretion as you proceed through the guide as you, at times, go deeper and add more testing or stakeholder insight, or engage a lighter touch for smaller changes to the platform. As you do so, ensure that you are not skipping steps in the process, and use the EOS North Star as your guide forward.

Mission-driven design doesn’t happen without rigorous commitment to the process, even when – and especially when – it gets hard. While this guide incorporates intentional moments to pause and take action to ensure mission-driven outcomes, the process assumes designing with intention always happens from the inception of the process. The bias risks and actions in this guide serve as key moments to pause and reflect on how well our organization is continuing to center its mission in the development process, rather than points where we consider it for the first time.

How to use this guide

Starting with Mission-Centered Design

17 of 33

“I also think about bias in technology...racial bias or any other kind of bias in AI technology that would impact our work and what AI is generating for us.”

“I think the reality that we've been through a lot of change and are going through a lot of changes in the organization and the thought about buy-in and how do you really secure buy-in across the stakeholder groups that need to buy in for it to be successful? I think that that could be a challenge area too.”

“I think we have to be incredibly cognizant of what we can and can't do and what we should and shouldn't take on and have a much stronger project management structure around this kind of monumental shift. And a really thoughtful runway.

EOS Staff Voices, Focus Groups Fall 2024

18 of 33

Principles for Mission-Centered AI at EOS

Investments in AI should support EOS’s mission, which is focused on improving the experience of underserved students in secondary education and their ability to access advanced coursework.

The design, development, and rollout of AI systems should be informed by the unique characteristics of each context in which an AI system will be deployed, with risk assessments and bias evaluations performed on training data and model outputs to ensure an AI system does not undermine EOS’s mission.

AI systems should be continuously monitored and re-evaluated to ensure they continue to serve their intended purpose as school cultures change, as student bodies change, and as available data and technologies change. AI system’s datasets and model documentation (see Appendix for templates) should be updated as the system itself is updated.

Investments in AI should support EOS in transforming school cultures and societal structures in support of EOS’s mission to increase access, belonging, and success in advanced secondary education courses for underserved students.

Student-centered

Contextual

Responsive

Transformative

19 of 33

20 of 33

21 of 33

EOS Mission-Driven

AI Strategy in Action

The following slides provide expanded definitions, actionable steps, and tools for use in each of the six stages of the mission-driven development process.

22 of 33

23 of 33

Product Discovery

Goal:

Define a specific problem or opportunity to solve, understand mission-alignment and return on investment (ROI) for solving the problem, and propose a potential solution.

  • A problem to solve or an idea for a solution
  • The EOS North Star documents, as a reference
  • Feedback & input from internal advisory groups and stakeholders
  • A high-level product requirements document
  • Sign-off to move forward to the Ideate & Plan stage
  • What problem or opportunity will the product solve? Is it mission-aligned?
  • Should we build something to address this?

Responsible

Product Owner*, Strategy Team, Technical Team

Accountable

Product Owner*, Strategy Team, Technical Team

Consulted

Internal Advisory Council*, Executive Leadership, Partnership Team, Board of Directors

Informed

External Advisory Council*, End Users

Overview

Inputs

Outputs

Key Questions

Stakeholders

*Strategic Additions

Inputs

Outputs

Activities

Sequence

24 of 33

Product Discovery

Responsible AI Strategy

Responsible AI Use Risk

Mission-Driven Actions

Organizational Culture Action

Stop Sign or Off-Ramp: Key resources or stakeholders are not available

Solution-first over stakeholder-first approach: Starting with a solution rather than a problem to be solved, or centering key influencers like funders and org leaders too early can fast-track solutions that are misaligned with the core value, mission and vision of EOS and/or the core values and needs of the schools and students EOS serves.

  • Review the EOS North Star and complete the alignment scoresheet
  • Engage the EOS AI Rubric in order to fully consider implications of the proposed solution

Consult the internal advisory committee for early feedback and guidance in order to foster communication, trust, and buy-in.

If the internal resources or key stakeholders are not available or accessible to the team, then stop development of the project or return to product discovery.

Internal Stakeholders

External Stakeholders

NONCRITICAL

CRITICAL

Engagement with internal stakeholders is critical to long-term success of the project

NONCRITICAL

CRITICAL

Select engagement can be helpful as needed, especially with experts

25 of 33

Understand the Opportunity

  • Before settling on a solution, create a broad set of possible problem/solution sets through ideation and brainstorming
  • Use the North Star alignment tool to confirm the solution aligns with the vision of EOS
  • Perform initial market research to create a landscape of competing products and assess the market opportunity and ROI for EOS

Documentation

  • Create a high-level product requirements document (PRD) that could include: product vision, business case & expected ROI, and general product requirements & product roadmap

Presentation

  • Share the product vision and business case to get feedback from Internal stakeholders and internal and external EOS AI advisory boards
  • Obtain formal sign-off to move to the Ideate & Plan stage

Activities

Product Discovery

26 of 33

Product Discovery

Tools

The EOS North Star

Overview: A tool to clarify alignment between EOS' organizational strengths, values, infrastructure, and opportunities for innovation.

Why it’s useful: The North Star Alignment score sheet builds shared awareness of the mechanisms and rationale behind initiatives and product opportunities. It highlights mission misalignment, competing interests, and key drivers of innovation efforts.

27 of 33

Product Discovery

Tools

Mission-Driven AI Use

Case Rubric

Overview: A comprehensive tool guiding EOS stakeholders through key considerations for AI tech development, including organizational, technological, mission-alignment, and field/market landscape implications.

Why it’s useful: Ensures effective and sustainable innovation by addressing contextual, technological, and organizational factors. Acts as an arbiter for competing interests and commitments in EOS’ decision-making process.

28 of 33

29 of 33

Prototype & Build

  • Approved product roadmap and planning documents from the Ideate & Plan stage
  • Stakeholder alignment and defined success metrics
  • Resource commitments (team, budget, tools)
  • Customer feedback and priorities (for iteration)
  • A functional MVP or early version of the product ready for testing and validation
  • Are the development efforts aligned with the product vision and roadmap?
  • Is the product meeting its intended functionality and addressing usability concerns that arise from user feedback?

Responsible

Product Owner*, Technical Team

Accountable

Product Owner*, Strategy Team, Technical Team

Consulted

End Users (Administrators, Teachers, Students), Partnership Team

Informed

Internal & External Advisory Councils*, Executive Leadership, Board of Directors

Overview

Inputs

Outputs

Key Questions

Stakeholders

*Strategic Additions

Goal:

Bring the product vision to life by developing the MVP and iterating based on user feedback to ensure functionality, usability, and alignment with the original vision.

Inputs

Outputs

Activities

Sequence

30 of 33

Responsible AI Risk

Mission-Driven Actions

Organizational Culture Action

Internal Stakeholders

External Stakeholders

Not considering unintended uses of the product: If potential misuse, intentional or unintentional, are not considered now, bias outcomes are more likely to occur and can lead to longer term mistrust and reputational risk.

  • Pause to develop, implement, and document a process for “future-proofing” the product
  • Leverage the external AI Advisory Council to provide feedback both during and after the build, centered on alignment and identifying unintended harm

Make it a non-negotiable priority to engage internal practitioners (those closest to the schools and districts where the tool will be used) to get feedback on early product versions or MVP.

If the internal resources or key stakeholders are not available or accessible to the team, then revisit the roadmap and adjust priorities to ensure alignment. Alternatively, pause further iterations to address gaps in stakeholder input or future-proofing considerations.

NONCRITICAL

CRITICAL

Engagement at critical points & sharing updates regularly

NONCRITICAL

CRITICAL

Early feedback as needed & sharing major updates

Responsible AI Strategy

Prototype & Build

Stop Sign or Off-Ramp: Key resources or stakeholders are not available

31 of 33

Product Development

  • Communicate requirements to the development team
  • Build the MVP according to the product vision

Documentation

  • Update technical documentation to ensure all team members have clarity on product features and architecture
  • Create user documentation or guides to support early pilot programs

Stakeholder Updates

  • Regularly communicate progress with stakeholders
  • Share major milestones, risks, and mitigation plans

Activities

Prototype & Build

32 of 33

Tools

The Ethical Operating System

Overview: A guide to anticipating the future impact of today’s technology is a toolkit that suggests key strategies for developers to explore the implication for their technological innovations.

Why it’s useful: The toolkit has a component to consider the risk zones developers should consider in their design process to better ensure the ethical development of future technologies.

Click here to access the tool

Prototype & Build

33 of 33

Responsible AI: A Catalyst for Change

“...This is the most important work that we can do in public education. The impact that we had on students... I got to experience firsthand and I believe every, every high school in the country should be partnered with us in doing this work.”

The first thing I think of is AI helping us take in various data points, both qualitative and quantitative and helping us create a plan for how a school or district moves through the work with us…That would be invaluable... It would be lovely to have somebody who could give you the specialized plan really quickly…”

“…District leaders are really looking at AI solutions. They're trying to figure it out. Just the mention of AI with your organization tweaks interest with people.”

— EOS Staff Voices, Focus Groups Fall 2024