All instruments
District Leadership Self-Assessment
Review the 26 district-level actions that set conditions for staff AI empowerment and student AI agency. Skim the strands or expand to review the full text of each row.
Practice
Level 1
Not Discussing Yet
Level 2
Starting to Plan
Level 3
Executing
Level 4
Advanced Execution
Level 1 · Not Discussing Yet
We have not yet started a process to outline AI-specific values, vision, and/or strategy in a way that coheres with existing organizational values, vision, and/or strategy.
Level 2 · Starting to Plan
We have started a process to outline AI-specific values, vision, and/or strategy in a way that coheres with existing organizational values, vision, and/or strategy.
Level 3 · Executing
We have outlined and shared with employees AI-specific values, vision, and/or strategy in a way that coheres with existing organizational values, vision, and/or strategy.
Level 4 · Advanced Execution
Level 3 PLUS: Member(s) of the C-suite have explicitly communicated support for -- or led the initial communication about -- the AI-specific values, vision, and/or strategy, and we have taken intentional steps to build staff investment in it, including opportunities for staff to engage with the vision rather than only receive it.
Level 1 · Not Discussing Yet
We have not yet made plans to launch an AI task force.
Level 2 · Starting to Plan
We are making plans to launch an AI task force.
Level 3 · Executing
We have launched and regularly convene an AI task force that either includes a member of the C-suite or has a mechanism for regular reporting to the C-suite.
Level 4 · Advanced Execution
Level 3 PLUS: The work of the task force is guided by clear objectives, and the group has a successful track record of accomplishing its objectives. OR: Having accomplished its initial objectives, our task force's responsibilities have been intentionally transitioned into standing leadership structures with clear decision-rights (e.g., AI is a recurring workstream of the cross-functional senior leadership team), such that a stand-alone task force is no longer required.
Level 1 · Not Discussing Yet
We have not yet made plans for incorporating AI leadership into the job description(s) and scorecards for one or more senior leadership roles in our organization.
Level 2 · Starting to Plan
We are making plans for incorporating AI leadership into the job description(s) and scorecards for one or more senior leadership roles in our organization.
Level 3 · Executing
AI leadership is explicitly reflected in the job description(s) and scorecards for one or more senior leadership roles in our organization.
Level 4 · Advanced Execution
Level 3 PLUS: There are clear decision-rights for major AI workstreams.
What this item asks
Decision-rights are clear when, for each major AI workstream (e.g., platform adoption, usage policy, instructional programming), a named role can make the call without relitigating who decides. This item asks whether AI leadership is assigned, not just assumed.
Level 1 · Not Discussing Yet
We are not yet developing or drafting a staff AI usage policy.
Level 2 · Starting to Plan
We are developing or have drafted a staff AI usage policy.
Level 3 · Executing
We have shared a staff AI usage policy that sets clear guardrails for personally identifiable information (PII) and ensures compliance with FERPA, COPPA, and state privacy laws.
Level 4 · Advanced Execution
Level 3 PLUS: We have implemented mandatory training on the usage policy for all staff, and we have a documented protocol for reporting, escalating, and responding to AI-related incidents. We use what we learn from incidents to monitor policy effectiveness and refine the policy over time, and our policy addresses AI acting on staff's behalf (e.g., agents operating with delegated access or permissions), including expectations for human oversight.
What this item asks
A staff AI usage policy should address where the real risks live. The most common gap: staff using personal AI accounts for work involving student or staff information, which places PII outside your system's agreements and protections. Strong policies address personal-account use directly, alongside approved tools and data guardrails.
Level 1 · Not Discussing Yet
We have not yet made plans to collect data from staff on their usage of and perspectives on AI.
Level 2 · Starting to Plan
We are making plans to collect data from staff on their usage of and perspectives on AI.
Level 3 · Executing
We have collected data from staff on their usage of and perspectives on AI.
Level 4 · Advanced Execution
Level 3 PLUS: We have collected multiple rounds of data from staff on their usage of and perspectives on AI, and we use this data to inform decisions (e.g., adjusting training, tooling, or policy) and to progress-monitor the efficacy of our AI-related work.
Level 1 · Not Discussing Yet
We have not yet made plans to collect data from students on their usage of and perspectives on AI.
Level 2 · Starting to Plan
We are making plans to collect data from students on their usage of and perspectives on AI.
Level 3 · Executing
We have collected data from students on their usage of and perspectives on AI.
Level 4 · Advanced Execution
Level 3 PLUS: We have collected multiple rounds of data from students on their usage of and perspectives on AI, and we use this data to inform decisions (e.g., adjusting instructional programming or student-facing tools) and to progress-monitor the efficacy of our AI-related work.
Level 1 · Not Discussing Yet
We have not yet made plans to collect data from families on their perspectives on AI.
Level 2 · Starting to Plan
We are making plans to collect data from families on their perspectives on AI.
Level 3 · Executing
We have collected data from families on their perspectives on AI.
Level 4 · Advanced Execution
Level 3 PLUS: We have collected multiple rounds of data from families on their perspectives on AI, and we use this data to inform decisions (e.g., adjusting communications or family engagement) and to progress-monitor the efficacy of our AI-related work.
Level 1 · Not Discussing Yet
We have not yet made plans for proactive external communication about our approach to AI and student technology use.
Level 2 · Starting to Plan
We are developing a strategic external communications plan that will share with families and our broader community the rationale for our approach to AI and student technology use (e.g., key messages, audiences, channels, and spokespeople).
Level 3 · Executing
We are executing on a strategic external communications plan that proactively shares with families and our broader community the rationale for our approach to AI and student technology use.
Level 4 · Advanced Execution
Level 3 PLUS: Our communications are accessible to all families (e.g., available in our community's major home languages and in family-friendly formats), we have prepared our response to foreseeable concerns (e.g., screen time, data privacy, academic integrity) before they surface, and we maintain channels for community input and use what we hear to adapt our messaging and strategy.
What this item asks
School systems are entering a period of heightened public attention to AI, screen time, and student technology use. This item asks whether your system proactively shapes that conversation with families and the broader community, rather than responding to it after sentiment has formed.
Level 1 · Not Discussing Yet
We have not yet identified or organized the innovators and early adopters in our organization.
Level 2 · Starting to Plan
We are identifying our innovators and early adopters and making plans to convene and support them as a coalition (e.g., explicit permission to experiment, dedicated time, defined guardrails).
Level 3 · Executing
We have identified and convened a coalition of innovators and early adopters from across roles and levels, with explicit permission, time, and support to experiment within defined guardrails, and we have begun to learn from and spotlight the work of our AI power users.
Level 4 · Advanced Execution
Level 3 PLUS: The coalition operates through structured mechanisms (e.g., regular convenings, resourced exploration), we capture what coalition members learn and diffuse it across the organization so lessons become organizational knowledge rather than personal expertise, and we deliberately engage skeptics (e.g., inviting them onto testing teams).
Level 1 · Not Discussing Yet
We have not yet established structures for AI experimentation; pilots happen ad hoc, in silos, or under the radar.
Level 2 · Starting to Plan
We are developing structures for visible, bounded, and supported AI pilots (e.g., defined scope, named participants, documented learning).
Level 3 · Executing
AI experimentation in our system happens through visible, bounded, and supported pilots (i.e., defined scope, named participants, documented learning) rather than under the radar.
Level 4 · Advanced Execution
Level 3 PLUS: We make evidence-based decisions about which piloted practices expand, adjust, or sunset, based on documented outcomes rather than enthusiasm or inertia.
Level 1 · Not Discussing Yet
We have not yet planned for how to manage for quality control as staff increasingly use AI within key workstreams.
Level 2 · Starting to Plan
We are planning how to manage for quality control as staff increasingly use AI within key workstreams.
Level 3 · Executing
We have put in place intentional practices to manage for quality control as staff increasingly use AI within key workstreams.
Level 4 · Advanced Execution
Level 3 PLUS: Our managers calibrate oversight to the stakes of the work, with defined expectations for review and revision of AI-assisted output in high-impact or high-risk workstreams, and recurring quality issues are used to refine our guidance and training over time.
What this item asks
Managing for quality means having intentional practices that keep AI-assisted work meeting your organization's bar: a clear expectation that staff own and are accountable for their output (reviewing and revising AI-generated work rather than using it as-is), norms that make it normal to flag quality issues in AI-assisted work, and oversight calibrated to stakes, so a board-facing analysis gets more scrutiny than an internal summary.
Level 1 · Not Discussing Yet
We have not yet begun evaluating options for a primary enterprise AI platform (a general-purpose AI platform provided to all or most staff).
Level 2 · Starting to Plan
We are formally evaluating options for a primary enterprise AI platform (a general-purpose AI platform provided to all or most staff).
Level 3 · Executing
We have adopted a primary enterprise AI platform (a general-purpose AI platform provided to all or most staff) that ensures data privacy protection, provides employees equitable access to AI tools, and provides admin oversight and data insight.
Level 4 · Advanced Execution
Level 3 PLUS: We monitor key platform usage trends. This data is regularly reviewed and used to inform relevant systems-level decisions.
What this item asks
An enterprise AI platform is a general-purpose AI platform your system provides to all or most staff: one that can support work across many functions and content areas rather than serving a single use case. To qualify, the platform must (a) ensure data privacy protection under an agreement with your system, (b) provide employees broad, equitable access to AI capabilities, and (c) give administrators oversight and usage insight. An internally built platform meeting these criteria also qualifies. An enterprise plan or paid license for a single-purpose (vertical) tool does not by itself qualify.
Level 1 · Not Discussing Yet
We have not yet made plans to monitor and manage our AI-related costs beyond paying license or subscription fees.
Level 2 · Starting to Plan
We are making plans to monitor and manage our AI-related costs, including establishing visibility into our overall AI usage and spend.
Level 3 · Executing
We actively monitor AI usage and spend across our platforms, attribute costs to major use cases or departments, and use this data to inform budgeting, procurement, and renewal decisions.
Level 4 · Advanced Execution
Level 3 PLUS: We match model capability to task requirements through an intentional routing strategy: routine, high-volume tasks are directed to lower-cost models, while our most complex, highest-leverage, or highest-risk tasks are reserved for the most capable models. We forecast AI costs as usage scales, and we weigh what our major AI investments return (e.g., time saved, quality gained, or outcomes improved) against what they cost, so we can expand, renegotiate, or cut with evidence.
What this item asks
AI pricing is increasingly hybrid: platforms often charge flat per-seat licenses for baseline access, while advanced models, agentic workflows, and usage beyond included limits carry costs that scale with use. This item asks whether your system manages both what AI costs and what it returns, so that spending flows toward the uses that deliver real value.
Level 1 · Not Discussing Yet
We have not yet made plans to address AI-specific considerations, such as data privacy and platform interoperability, in our technology and software procurement practice.
Level 2 · Starting to Plan
We are making plans to address AI-specific considerations in our procurement practice, including how we will review vendor data terms, cover AI in our data privacy agreements, and weigh platform interoperability.
Level 3 · Executing
Our procurement practice addresses AI-specific considerations at adoption: we review vendor data retention and model-training terms before purchase, our data privacy agreements cover AI-specific data flows, and we weigh platform interoperability and exit costs to avoid lock-in.
Level 4 · Advanced Execution
Level 3 PLUS: Our practice extends across the contract lifecycle, with renewal reviews that catch AI features newly added to existing products. We vet student-facing AI tools against defined criteria (data privacy, safety guardrails, developmental appropriateness, and instructional fit) before they are deployed to students, and we have a defined protocol for escalating and responding to safety or wellbeing concerns that surface through student use of AI.
Level 1 · Not Discussing Yet
We have not yet established expectations or structures for AI tools and applications that staff build themselves.
Level 2 · Starting to Plan
We are developing expectations and structures for staff-built AI tools (e.g., approved build platforms, review based on data sensitivity, ownership and maintenance expectations).
Level 3 · Executing
We have established a sanctioned pathway for staff-built AI tools, including approved platforms for building, review at the time of creation proportional to data sensitivity (with any tool touching student or staff personal information receiving the highest scrutiny, including security review during development rather than only at completion), and clear ownership and maintenance expectations.
Level 4 · Advanced Execution
Level 3 PLUS: Our controls are enforced technically rather than by policy alone: staff build within guardrailed environments where data cannot leave our system even when a tool is flawed, we maintain a current inventory of what staff have built, and higher-risk tools are re-reviewed on a defined cycle.
What this item asks
AI tools now allow staff without programming backgrounds to build working software: custom chatbots, applications, and automations. This item asks whether your system has structures that keep what staff build safe, secure, and sustainable.
Level 1 · Not Discussing Yet
We have not yet made plans to intentionally develop the AI competency of our existing staff; learning is ad-hoc and self-directed.
Level 2 · Starting to Plan
We are making plans to intentionally develop staff AI competency (e.g., structured learning experiences, protected time to practice, role-relevant expectations for AI use).
Level 3 · Executing
We are executing on an intentional strategy to develop staff AI competency that reaches broadly across roles, not only early adopters, and includes structured learning experiences and protected time to practice.
Level 4 · Advanced Execution
Level 3 PLUS: Competency development is differentiated by role, extends to building with AI (e.g., creating custom tools, workflows, or applications) rather than only using AI tools, and we track participation and growth to inform our next investments.
Level 1 · Not Discussing Yet
We have not yet made plans to incorporate AI-related competencies into hiring criteria.
Level 2 · Starting to Plan
We are making plans to incorporate AI-related competencies into hiring criteria.
Level 3 · Executing
We incorporate AI-related competencies into our hiring criteria.
Level 4 · Advanced Execution
Level 3 PLUS: We have incorporated AI-related competencies into hiring criteria across a broad range of organizational roles.
Level 1 · Not Discussing Yet
We have not yet looked at our key non-instructional workflows step by step to decide where AI does and doesn't belong.
Level 2 · Starting to Plan
We are making plans to break down key non-instructional workflows into steps and identify which steps are best handled by software that follows set rules, which are best handled by AI, and which need to stay with people.
Level 3 · Executing
For some of our key non-instructional workflows, we have decided step by step which parts are best handled by software that follows set rules, which by AI, and which stay with people, and we are acting on those decisions. Each AI-supported step has a person who owns it and is accountable for its output.
Level 4 · Advanced Execution
Level 3 PLUS: We match the level of human review to the stakes of the work: routine internal work may be spot-checked, anything that reaches families is reviewed before it goes out, and work with legal, financial, or safety consequences stays with rule-following software or people. We keep a current picture of how our high-volume workflows are handled and revisit those decisions as AI capabilities change.
What this item asks
A single workflow rarely contains one kind of work. Some steps follow set rules and are best handled by regular software that gives the same output every time. Some steps are well suited to AI. And some steps carry decisions that need to stay with a person. This item asks whether your system is mapping and making intentional decisions within workflows, at the step level.
Level 1 · Not Discussing Yet
We have not yet considered whether our organizational documents and data are ready to be used by AI systems.
Level 2 · Starting to Plan
We are making plans to identify which documents and data sources are the authoritative version for key organizational information (e.g., policies, procedures, and operational data) and to assign owners responsible for keeping them current.
Level 3 · Executing
For key organizational information (e.g., policies, procedures, and operational data), we have identified the authoritative version and named an owner responsible for keeping it current, and staff know which sources to point AI tools toward.
Level 4 · Advanced Execution
Level 3 PLUS: Our AI tools are connected to these authoritative sources so answers are grounded in them rather than in whatever content appears most relevant, owned content is reviewed on a defined cycle and outdated versions are retired, and when AI gives a wrong answer, we correct the underlying source rather than only the response.
What this item asks
AI systems are only as reliable as the information they draw on. When organizational documents and data are fragmented, outdated, or conflicting, AI can return answers that cite real sources but are wrong, and a confident citation can make a wrong answer more convincing. This item asks whether your system manages its information so AI works from trusted, current sources.
Level 1 · Not Discussing Yet
We have not yet made plans to re-evaluate our graduate profile and ensure it reflects student outcomes and competencies that will position our graduates to flourish — personally and professionally — in a world of accelerating AI integration.
Level 2 · Starting to Plan
We are planning a process to re-evaluate our graduate profile and ensure it reflects student outcomes and competencies that will position our graduates to flourish — personally and professionally — in a world of accelerating AI integration.
Level 3 · Executing
We have re-evaluated our graduate profile and ensured it reflects competencies that will position our graduates to flourish — personally and professionally — in a world of accelerating AI integration.
Level 4 · Advanced Execution
Level 3 PLUS: The AI-relevant competencies in our portrait of a graduate are each mapped to an assessment measure.
What this item asks
A portrait of a graduate (sometimes called a graduate profile or learner profile) is a school system's statement of the outcomes and competencies every student should have by graduation. Systems that do not serve graduating grades should apply this item to the culminating outcomes they define for their students. This item asks whether yours has been re-examined for a world where AI is reshaping what those competencies need to be.
Level 1 · Not Discussing Yet
We have not yet made plans to address AI's impact on fields and career paths in our college and career advising.
Level 2 · Starting to Plan
We are making plans to address AI's impact on fields and career paths in our college and career advising.
Level 3 · Executing
Our college and career advising includes structured opportunities for students to examine how AI is changing the fields and career paths they are considering, so that pathway decisions are made with that understanding rather than without it.
Level 4 · Advanced Execution
Level 3 PLUS: Our counselors and advisors have training and current resources on how AI is reshaping postsecondary pathways, we refresh these on a regular basis given how quickly conditions change, and examining AI's impact is a standard part of the advising experience for all students, not an offering that only some encounter.
What this item asks
How AI will reshape fields, majors, and career paths is genuinely uncertain, and the picture is moving fast. This item does not ask your system to predict the future of work with confidence. It asks whether students are making postsecondary decisions with the best available current understanding, and whether they are learning to keep examining the question as it evolves.
Level 1 · Not Discussing Yet
We have not yet made plans to establish shared expectations for student AI usage.
Level 2 · Starting to Plan
We are developing shared expectations for student AI usage (e.g., adopting an AI usage framework, drafting student-facing expectations that support responsible use while protecting academic integrity).
Level 3 · Executing
We have adopted an AI usage framework and trained instructional staff on it, so our community shares common language for when students work with AI and when they work without it on given tasks and assessments. We have also communicated clear expectations to students, signaling support for AI usage while setting academic integrity conditions that can be reliably enforced.
Level 4 · Advanced Execution
Level 3 PLUS: Our framework distinguishes types of student AI interaction by their impact on cognitive engagement, from uses that collapse rigor to those that actively uphold it. These expectations are consistently applied at the task and assessment level across classrooms, students can articulate what is expected of them, and we revisit the expectations regularly as tools and practices evolve.
Level 1 · Not Discussing Yet
We have not yet made plans to teach students about AI (how it works, its capabilities and limits, and its impacts).
Level 2 · Starting to Plan
We are making plans to teach students about AI, whether through stand-alone programming or within existing coursework.
Level 3 · Executing
In at least some grade levels, we are actively teaching students about AI (e.g., how AI systems work, their capabilities and limits, and their ethical and societal impacts).
Level 4 · Advanced Execution
Level 3 PLUS: Learning about AI is systematically implemented across our relevant grade levels, with a defined scope and sequence, integration within core content where appropriate, and a means of assessing student learning.
What this item asks
Understanding AI (Item 23) covers how AI systems work, their capabilities and limits, and their impacts on people and society. Applied AI literacy (Item 24) is students' capability to use AI well in practice: exercising judgment and discernment about when and how AI helps or harms, and designing and building with AI to address real-world challenges. Either can be delivered through stand-alone programming (dedicated courses, units, modules) or integrated within core content and HQIM; most systems will use both pathways.
Level 1 · Not Discussing Yet
We have not yet made plans for intentional programming or curriculum that builds students' capability to use and build with AI.
Level 2 · Starting to Plan
We are making plans for intentional programming and curriculum that build (1) students' ability to use AI with judgment and discernment and (2) their capability to design and build with AI.
Level 3 · Executing
In at least some grade levels or content areas, we are executing on intentional programming and curriculum that build (1) students' ability to use AI with judgment and discernment and (2) their capability to design and build with AI to address real-world challenges.
Level 4 · Advanced Execution
Level 3 PLUS: This programming is systematically implemented across relevant grade levels and embedded within core content, and we measure students' applied AI literacy (e.g., through performance tasks) and use the results to improve instruction.
Level 1 · Not Discussing Yet
We have not yet made plans to evaluate the implementation quality or impact of the AI tools we deploy to students.
Level 2 · Starting to Plan
We are making plans to evaluate our student-facing AI tools, including defining intended use and the student outcomes each tool is meant to improve.
Level 3 · Executing
For our key student-facing AI tools, we have defined intended use and target student outcomes, and we monitor implementation quality (e.g., usage patterns and interaction quality against intended use).
Level 4 · Advanced Execution
Level 3 PLUS: We evaluate each key tool's impact on its target outcomes using evidence beyond vendor-provided data, and we use the findings to expand, adjust, or sunset tools.
What this item asks
When a system deploys AI tools to students (tutors, writing supports, curriculum-embedded tools), the value depends on the quality of implementation and on measured impact, not on adoption alone. This item asks whether your system evaluates both. Usage and engagement metrics (logins, time on task) indicate implementation, not impact; impact means evidence of progress on the student outcomes the tool targets.
Level 1 · Not Discussing Yet
We have not yet looked at our key instructional workflows step by step to decide where AI does and doesn't belong.
Level 2 · Starting to Plan
We are making plans to break down key instructional workflows into steps and identify which steps are best handled by software that follows set rules, which are best handled by AI, and which need to stay with people.
Level 3 · Executing
For some of our key instructional workflows, we have decided step by step which parts are best handled by software that follows set rules, which by AI, and which stay with people, and we are acting on those decisions. Each AI-supported step has a person who owns it and is accountable for its output.
Level 4 · Advanced Execution
Level 3 PLUS: We match the level of human review to the stakes of the work: routine preparation may be spot-checked, anything that reaches students or families is reviewed before it goes out (a teacher reviewing and adjusting an AI-drafted lesson plan before teaching it is this review), and decisions with consequences for individual students stay with people. We keep a current picture of how our high-volume instructional workflows are handled and revisit those decisions as AI capabilities change.
What this item asks
A single workflow rarely contains one kind of work. Some steps follow set rules and are best handled by regular software that gives the same output every time. Some steps are well suited to AI. And some steps carry decisions that need to stay with a person. This item asks whether your system is mapping and making intentional decisions within instructional workflows, at the step level.