Skip to main content
Peninsula School District · Gig Harbor, WashingtonOpen source · free to run
Peninsula AI
№ 03  ·  Fellowship Submission Google & GSV Ed Leader Fellowship · 2026–2027 Cohort

Transformational Opportunity
Action Plan

Peninsula School District · Office of the Superintendent

Part 01

Identifying the Transformational Opportunity

Opportunity A From the anchor chart · System change by design thinking

Building the Open Adaptive District

Description

Peninsula is building a new way of working — a district that can keep pace with AI as it changes. Right now, when a new tool comes out, school districts often take years to figure out what to do with it. By then, the next wave has already arrived, and educators are exhausted. We are going to change that here.

Every six weeks, small teams across the district — cabinet departments, principal teams by grade level, and later groups of teachers and operations staff — will choose one experiment to try. They will start by asking what has changed in the last six weeks and what problem they want to solve. They will write down what they expect to happen and how they will know. They will run the experiment. At the end of the cycle, they will share what they learned — including what did not work — and everything will be posted publicly on psd401.ai so families, staff, and other districts can see how we work. (We call this the Open Adaptive District cycle.)

Each quarter, I will write a short, three-sentence statement of what we are after, why it matters now, and what we will not give up along the way. Teams take that direction and choose their own path inside it. I do not tell teams how to do their work — I tell them what we are after. A template guides this quarterly statement.

Rather than adding more meetings to coordinate all of this, we are building an internal AI tool that quietly reads each team's weekly written updates and keeps everyone — including me — informed about what is working, what is stuck, and where teams could help each other. The point is simple: a district learning to use AI well should use AI to organize its own learning.

What we are building is not an AI plan or a technology rollout. It is the district's ability to keep adjusting, honestly and in public, at the pace the world is actually moving — so that what we offer our students next year can be very different from what we offered them last year, and we can show our families exactly how we got there.

Why this is important

AI capability is doubling on a roughly seven-month curve while traditional district change cycles run two to five years. Without an adaptive operating model, every new wave of technology creates fresh paralysis. Districts default to one of two failure modes: waiting for the state to act, or accepting vendor-driven adoption that bypasses public mission. Both deepen inequity and surrender district agency. The Open Adaptive District is the alternative.

Barriers
  • Traditional change models (Kotter, Fullan) assume stable end-states and are too slow for an exponential environment
  • Cabinet, principals, and staff prioritize what they hear most often; the operating model competes with forty other initiatives
  • Meeting culture struggles to carry cross-silo coordination
  • Some leaders treat complex problems as if they were complicated and try to expert-solve them
  • Education norms punish public failure, which discourages honest debriefs
Risks
  • Moving fast without governance erodes community trust
  • Moving slowly cedes the field to vendors and consultants
  • Capacity stays in four or five senior people instead of a system
  • Cycles become theater — teams go through the motions without actually shipping
  • Bets that don't work get quietly dropped instead of publicly debriefed
Resources needed
  • Superintendent ownership and cabinet alignment (in place)
  • The district agent (the CIO builds summer 2026)
  • Quarterly intent discipline from the superintendent
  • Dedicated time for cycle teams — not bolt-on
  • Communications capacity to make wins and decommissions visible
  • A published playbook on psd401.ai
How have you tried to address this

Four years of progressive AI work at PSD: a growing roster of production AI agents we have built ourselves, a custom AI tools platform at psd401.ai/aistudio where staff create their own assistants, the public psd401.ai website where we share our work openly, and national-level conference and press visibility.

What has worked: top-down permission paired with bottom-up tinkering across cabinet, principals, and staff; making our work visible early through psd401.ai, public talks, and press coverage; building tools that staff actually use.

What hasn't worked yet: assuming directors would prioritize without being told; assuming principals would notice and replicate without scaffolding; treating AI as a teaching-and-learning topic rather than a district operating-model question.

What remains underbuilt: the cadence — a published, repeatable adaptive cycle that other leaders can run without being us. The Open Adaptive District is that cadence.

Priority vs 2026–27 Goals
High Priority
Resourcing Difficulty
Resources available
Personal Passion
High personal passion
Opportunity B From the anchor chart · Innovation at scale

The Open Adaptive District
as a Public-Interest Model

Description

The work happening inside Peninsula is too important to keep to ourselves. Other districts — large and small, well-resourced and not — face the same questions we do, but very few have the staff, the partners, or the permission to work them out on their own. Without somewhere credible to look, those districts often end up buying whatever a vendor is selling. We want there to be a public alternative they can turn to — one built by districts, for districts.

Across the fellowship year, we will turn Peninsula's working model into something other districts can actually pick up and use. Our public website, psd401.ai, will grow from a place where we tell stories of our work into a working library — the six-week cycle we run, the templates we use, the rules we follow about what we share and what we keep private, the AI tools we build, and the honest debrief from every team, including the experiments that failed. If another district could copy it, adapt it, or run it themselves, it will be there.

In February 2027, Peninsula will host a regional gathering for superintendents and their teams from across the Pacific Northwest. The day will not be a conference. It will be a working meeting, where each district leaves with a draft of their own first six-week cycle, a partner district to compare notes with through the spring, and clear ways to keep contributing back to the shared library as they go. We hope at least three or four other districts are running their own cycle by the end of the school year.

Alongside the gathering, we will publish a longer written case study describing how Peninsula has done this work and what we have learned. A shorter, more practical whitepaper will give district leaders the tools they need to start. A short visual piece — a video or photo essay — will help families and staff see the spirit of the work, not just the structure.

We are not trying to put a flag on the moon. We are trying to make sure that, when the next wave of AI lands, public schools have a model of their own to point to — one that other districts had a hand in building.

Why this is important

Capacity-constrained districts will otherwise be absorbed into closed vendor stacks. A public, district-owned alternative protects public education's agency in deciding what AI in education looks like. Districts deserve a seat at the table, not a seat in the audience. Peninsula's local work becomes a public contribution.

Barriers
  • Regional districts have very different readiness levels
  • Sustained inter-district collaboration is hard without immediate value
  • PSD capacity is already stretched leading internally
  • No widely shared district-centered AI operating framework currently exists for the coalition to rally around
  • Other districts may pattern-match this to past "leader district shows off" efforts and disengage
Risks
  • Coalition becomes conversation without artifacts
  • PSD overextends and weakens the internal work
  • Equity concern: early-mover districts benefit faster than capacity-constrained ones
  • Framing becomes too oppositional toward vendors and burns relationships
Resources needed
  • Convening capacity for the February symposium
  • psd401.ai evolved into a two-way contribution platform
  • Writing and publication time for the long-form article and whitepaper
  • Higher education and workforce partnerships
  • Fellowship network connections
  • Modest funding for symposium and publication production
How have you tried to address this

The primary national platform for sharing this work has been the American Association of School Administrators (AASA), where the superintendent sits alongside peers in national conversations on AI, governance, and operations. Beyond AASA, Peninsula engages prominently through CoSN (the Consortium for School Networking) and Digital Promise's League of Innovative Schools, with smaller participation in ERDI and WASA. Public-facing sharing through psd401.ai, published articles, and national conference talks has reached district leaders well beyond those formal networks. Individual districts have reached out one at a time for informal mentoring.

What has worked: districts across the country have been inspired by Peninsula's example — including at least one that stood up a full AI department after seeing our work. Peninsula has earned a national platform.

What hasn't worked yet: sharing has been one-way; there is no coalition structure; no shared framework for other districts to fork.

Priority vs 2026–27 Goals
High Priority
Resourcing Difficulty
Resources might be available
Personal Passion
High personal passion
Part 02

Action Planning

Idea

Launch and document The Open Adaptive District — an operating model in which cross-silo teams run six-week cycles (Plan, Do, Study, Share) under a quarterly superintendent intent, coordinated through the district agent (an AI agent that reads team Spaces and synthesizes across the district), with every cycle's lessons published openly on psd401.ai. Convene a February 2027 regional symposium and co-publish a long-form article and whitepaper framing PSD's operating model as a portable, public-interest contribution other districts can adopt.

How this addresses both opportunities

A and B are two layers of the same work. A is the internal practice; B is the field-facing translation. Every internal artifact becomes a public artifact. The cycles produce the evidence that the operating model works; psd401.ai publishes that evidence as the field's reference implementation. The February symposium activates the coalition. The publication makes the model citable.

Metrics for success

Headline
Proof point. ≥2 AI capabilities released after Sept 2026 absorbed into PSD operations within 60 days of release.
M-01
≥4 cycles completed across ≥6 teams (Aug 2026–May 2027)
M-02
≥4 public cycle debriefs on psd401.ai, including ≥1 decommissioned bet
M-03
February 2027 PSD-hosted regional symposium delivered
M-04
≥3 regional districts running their own cycle by May 2027
M-05
1 long-form article + 1 whitepaper + multimedia playbook on psd401.ai
M-06
The district agent operational from cycle one and in routine use
M-07
Quarterly intent issued by superintendent four consecutive quarters

Timeline

Jun – Aug 2026
Doctrine finalized · psd401.ai v2 scoped · async agent prototyped with Tech · cabinet and principal August kickoff
Sep – Oct 2026
Cycle 01 across cabinet department teams and principal grade-band teams
Nov 2026
First debriefs published · teacher teams and operational staff layered in · Cycle 02 begins
Dec 2026 – Jan 2027
Cycle 03 · symposium logistics finalized · whitepaper drafted
Feb 2027
Regional symposium hosted at PSD
Mar – Apr 2027
Cycle 04 · long-form article finalized · coalition commitments collected
May 1, 2027
Fellowship deliverables submitted — publication, whitepaper, psd401.ai v2 live, coalition documented

Stakeholders

Internal core

  • Superintendent
  • Cabinet
  • Board (sponsor)
  • Principals

Builder layer

  • Teacher teams (in-building or cross-district)
  • Operational staff (Tech · ESS · Business)
  • Any group ready to run a cycle

External

  • Students & families
  • Regional superintendents
  • Higher-ed & workforce
  • Fellowship peers

Support needed

  • Executive coaching and thought partnership from the fellowship network
  • Legal and governance review for the publish-vs-redact rule
  • Symposium logistics and facilitation
  • Editorial support for publication
  • Documentation and case-study writing support

Required resources

Project management capacity (the CIO and the Director of Research and Assessment as primary stewards) · cabinet and principal time inside the cycles · psd401.ai v2 build resources · modest funding for symposium production and publication design · travel for one or two coalition pre-meetings.

Risks & wonders

  • Will teams treat the cycles as theater if leadership pressure is uneven?
  • Will principals feel observed or empowered by the async digest? The answer depends on tone, not mechanism.
  • Will the coalition want a charter and artifact before the symposium, or is informal coalition-building sufficient in year one?
  • Does the publication serve PSD's narrative or the field's adoption? They overlap but are not identical.
  • How honest can decommissioning be in public without political cost?

Immediate next steps

Finalize Open Adaptive District doctrinethe CIO & the superintendent·Within 2 weeks
Build the district agent prototypethe CIO · Tech department·By Jul 15
Cabinet alignment sessionthe superintendent·Late July
August all-administrator kickoff — Cycle 01the superintendent & the CIO·First week of August
psd401.ai v2 design specthe Director of Research and Assessment & Communications·6 weeks
Symposium save-the-date to regional superintendentsthe superintendent·8 weeks
Whitepaper outlinethe CIO & the superintendent·12 weeks
Appendix · Optional

TregoED SCAN Protocol

S — See the issues

The district is operating in a moment when AI is transforming learning, work, communication, and public expectations on a months-long curve. The issue is not whether AI will be used in schools; it is whether districts will build the institutional capacity to govern, integrate, and direct AI responsibly — and whether public education will define its own model before closed vendor stacks define one for it.

C — Clarify the issues

"Adaptive District" means a district whose decision and execution loops match the pace of the change it operates in. "Public-interest model" means an approach centered on equity, transparency, interoperability, and district agency — distinct from market dependence.

A — Assess priorities

Issue Clarify Sev Urg Grth
AI is outpacing district decision cycles Districts react more quickly than they design H H H
Vendor capture of public-ed AI direction Closed stacks harden faster than public alternatives H H H
Cross-silo info flow is broken Coordination depends on meetings that don't work H H M
Public failure is culturally punished Killing bets is hidden, so learning is lost H M H
Field has no shared adaptive framework No reference model for other districts to fork M M H
Trust and communication must be proactive Families and staff need clarity before scale H H H

N — Name next steps

Action By Whom By When
Finalize Open Adaptive District doctrine the superintendent & the CIO Within 2 weeks
Build the district agent prototype the CIO By July 15
Align cabinet on operating model the superintendent Late July
August all-administrator kickoff for cycle 01 the superintendent & the CIO First week of August
psd401.ai v2 design spec the Director of Research and Assessment & Communications Within 6 weeks
Symposium save-the-date to regional superintendents the superintendent Within 8 weeks
Whitepaper outline the CIO & the superintendent Within 12 weeks
№ 03   Peninsula School District  ·  the Superintendent, Superintendent Google & GSV Fellowship   Draft v1