AI-Enabled School Redesign Suite

Designing an AI-enabled platform so schools can lead their own redesigns

Visuals in this case study are limited due to NDAs and proprietary information.

From coach-led redesign to an AI-enabled platform
From coach-led redesign to an AI-enabled platform

Timeline

Mar 2026 – present

My Role

Fractional UX Leader

Skills Used

UX Strategy

Product Design Strategy

AI Interfaces

Design-Engineering Collaboration Leadership

Wireframing & High Fidelity Prototyping

Design Systems & Libraries

Front-end (HTML, React, Angular, Vue)

Style Sheets (CSS, SCSS, SASS)

Design to Code Workflows

Stakeholder & Executive Presentations

Goal

Transform Transcend’s school redesign process - a bespoke, coach-led process refined over many years - into an AI-enabled tech suite that allows schools to be trained to lead their own redesigns, rather than paying Transcend to lead them through the process. This product sets the direction for a new, scalable business model for the organization.

Constraints

  • New product team operating in startup mode within an established non-profit

  • Complex, highly interconnected domain (school models, components, measures and plans) that only existed in coaches’ expertise and documents

  • Fractional role with alpha launch deadlines

  • Executive leadership needed to see and react to working software early in the process

Summary

Partnering directly with executive leadership, I led the user experience for a greenfield, 0-to-1 product - from the initial interaction specifications and coded prototypes, to a multi-app demo ecosystem, to production alpha releases. Designing and coding in React and SCSS on a coded, token-based design system I built from zero allowed new alpha features to be delivered from initial design to front-end code on a 3-4 day turnaround.

6 Weeks

To a complete, working demo of a highly complex application

3 Days

To a working Google Workspace prototype with two-way sync and embedded AI

3-4 Days

Design and front-end code turnaround for new alpha features

0-to-1

Greenfield product taken from first prototype to alpha launch

Process

Making an Analog Process Concrete

Transcend’s school redesign process has been proven over many years, but it was almost entirely analog - it lived in the expertise of their coaches, in workshops and in documents. Before any software could be built, the team needed a shared understanding of what the product actually was and how all of the pieces fit together.

I started with the core object of the work: a school’s blueprint, which includes the school’s components, its portrait of a graduate and how the design changes from a baseline version to an aspirational version. I built a working prototype in React within the first few weeks and wrote a detailed interaction and behavior specification covering views, navigation, side-by-side comparison, comments, saved views and export, so that design decisions were documented and shared instead of living as tribal knowledge.

Transcend’s school redesign process has been proven over many years, but it was almost entirely analog - it lived in the expertise of their coaches, in workshops and in documents. Before any software could be built, the team needed a shared understanding of what the product actually was and how all of the pieces fit together.

I started with the core object of the work: a school’s blueprint, which includes the school’s components, its portrait of a graduate and how the design changes from a baseline version to an aspirational version. I built a working prototype in React within the first few weeks and wrote a detailed interaction and behavior specification covering views, navigation, side-by-side comparison, comments, saved views and export, so that design decisions were documented and shared instead of living as tribal knowledge.

Guiding Principle

Executive leaders and coaches can react to and give feedback on something they can click through far faster and more accurately than a slide deck or static mockups. Getting to working software early reduces the time spent debating abstract concepts and brings gaps in the process to the surface much sooner.

Guiding Principle

Executive leaders and coaches can react to and give feedback on something they can click through far faster and more accurately than a slide deck or static mockups. Getting to working software early reduces the time spent debating abstract concepts and brings gaps in the process to the surface much sooner.

Demonstrating the Full Vision

Leadership needed a demo that showed the full vision of the product: how a Transcend design partner, a school’s design team and the school’s blueprint all connect throughout the redesign process. I designed and co-built three connected applications - a design partner portal, a design team portal and a blueprint builder - that share one component library and one data layer, allowing a presenter to switch between user roles in the middle of a demo and see the work carry over.

AI was designed into the entire flow rather than added as a separate chatbot, guiding users through each step of the process, including a voice conversation mode. Alongside the applications, I wrote production specifications and feature PRDs so that the engineering team could pick up the features that tested well with stakeholders.

Three connected apps on a shared component library and state layer

Leadership needed a demo that showed the full vision of the product: how a Transcend design partner, a school’s design team and the school’s blueprint all connect throughout the redesign process. I designed and co-built three connected applications - a design partner portal, a design team portal and a blueprint builder - that share one component library and one data layer, allowing a presenter to switch between user roles in the middle of a demo and see the work carry over.

AI was designed into the entire flow rather than added as a separate chatbot, guiding users through each step of the process, including a voice conversation mode. Alongside the applications, I wrote production specifications and feature PRDs so that the engineering team could pick up the features that tested well with stakeholders.

Three connected apps on a shared component library and state layer

Impact

A complete, working demo prototype of a highly complex application was delivered in 6 weeks, giving executive leadership a concrete, working product to react to instead of a concept.

Impact

A complete, working demo prototype of a highly complex application was delivered in 6 weeks, giving executive leadership a concrete, working product to react to instead of a concept.

Integrating with the Tools Schools Already Use

School teams already do their work in tools they know - Google Workspace for documents, presentations and calendars, and meeting note-takers like Fathom and Granola for recordings and transcripts. Asking them to move all of that into a new system would add a large barrier to adoption. In 3 days, I built a working prototype that integrated with Google Drive, Docs, Slides and Calendar with two-way (push and pull) sync and AI working alongside the documents, proving the integration approach before engineering time was invested in building it. Meeting notes and transcripts from Fathom and Granola are pulled in as well, so the AI can work from the conversations a team has already had, too.

Integrations: Google Drive, Docs and Slides, Calendar, and Fathom and Granola, connected to the platform and AI

School teams already do their work in tools they know - Google Workspace for documents, presentations and calendars, and meeting note-takers like Fathom and Granola for recordings and transcripts. Asking them to move all of that into a new system would add a large barrier to adoption. In 3 days, I built a working prototype that integrated with Google Drive, Docs, Slides and Calendar with two-way (push and pull) sync and AI working alongside the documents, proving the integration approach before engineering time was invested in building it. Meeting notes and transcripts from Fathom and Granola are pulled in as well, so the AI can work from the conversations a team has already had, too.

Integrations: Google Drive, Docs and Slides, Calendar, and Fathom and Granola, connected to the platform and AI

Decision

Integrate with Google Workspace through two-way sync rather than asking schools to replace the tools they already use.

Decision

Integrate with Google Workspace through two-way sync rather than asking schools to replace the tools they already use.

An AI Partner for Running the Redesign

Much of the redesign process happens in meetings, and the Transcend coach carries the preparation, materials and follow-through for each of them. I built a standalone prototype of the coach’s workspace in about four weeks using Next.js, Claude and Google Workspace. The prototype mirrors the production data model table for table, so anything created in the prototype can be moved directly into the product.

In the workspace, coaches can lay out meetings in order before dates are set and then schedule them all at once, see participants by role and check the readiness of materials for each meeting, working from documents, decks, spreadsheets and recordings that are mirrored from Google Drive. The AI assistant can be opened on any meeting to help the coach prepare, debrief privately or plan next steps - tailoring decks, adapting activities and adding objectives - while only reading the documents the user has permission to open. It can also build a school’s interview questionnaire as a Google Doc from a question bank, with the most important questions first. A read-only team site can then be published to keep the whole team informed.

The coach workspace: meetings, materials, AI assistant and outputs

Much of the redesign process happens in meetings, and the Transcend coach carries the preparation, materials and follow-through for each of them. I built a standalone prototype of the coach’s workspace in about four weeks using Next.js, Claude and Google Workspace. The prototype mirrors the production data model table for table, so anything created in the prototype can be moved directly into the product.

In the workspace, coaches can lay out meetings in order before dates are set and then schedule them all at once, see participants by role and check the readiness of materials for each meeting, working from documents, decks, spreadsheets and recordings that are mirrored from Google Drive. The AI assistant can be opened on any meeting to help the coach prepare, debrief privately or plan next steps - tailoring decks, adapting activities and adding objectives - while only reading the documents the user has permission to open. It can also build a school’s interview questionnaire as a Google Doc from a question bank, with the most important questions first. A read-only team site can then be published to keep the whole team informed.

The coach workspace: meetings, materials, AI assistant and outputs

Guiding Principle

AI tools for coaches need to work inside the coach’s real materials and permissions. The AI prepares and adapts the work, but the coach always stays in control of it.

Guiding Principle

AI tools for coaches need to work inside the coach’s real materials and permissions. The AI prepares and adapts the work, but the coach always stays in control of it.

Shipping Alpha Features

For the production alpha, I work as both a designer and front-end developer alongside the engineering team, designing and coding new features from scratch in React, TypeScript and SCSS. These have included planning and calendar tools, a step-by-step component evaluation flow, a redesigned library of solutions, implementations and schools, and workspace patterns that are shared across tools. All features are built on the shared theme and component system rather than with one-off styling, keeping the product consistent as it grows.

Tools that are proven in the prototype move into production with handoff documentation for the engineers who pick them up, and production’s blueprint tools are synced back into the prototype so both stay current.

The 3-4 day feature cycle inside the lab prototype, and the flow to production

For the production alpha, I work as both a designer and front-end developer alongside the engineering team, designing and coding new features from scratch in React, TypeScript and SCSS. These have included planning and calendar tools, a step-by-step component evaluation flow, a redesigned library of solutions, implementations and schools, and workspace patterns that are shared across tools. All features are built on the shared theme and component system rather than with one-off styling, keeping the product consistent as it grows.

Tools that are proven in the prototype move into production with handoff documentation for the engineers who pick them up, and production’s blueprint tools are synced back into the prototype so both stay current.

The 3-4 day feature cycle inside the lab prototype, and the flow to production

Takeaway

New alpha features are delivered from initial design to front-end code on a 3-4 day turnaround.

Takeaway

New alpha features are delivered from initial design to front-end code on a 3-4 day turnaround.

Get in Touch

Get in Touch

Get in Touch

Subscribe to EdTech Insights