Director, Learning Product and AI Strategy, United Way Worldwide

I redesign work for AI, and prove the change with evidence.

Draw the line in the system, not in a prompt

The agent works inside the lineIt uses only what it’s been given, and an instruction hidden in a document can’t widen its limits.
Permission to draft is not permission to sendA person who can say no approves anything before it leaves.
The line moves on evidenceAs AI gets more capable, the line moves one permission at a time, and only with test evidence.

Learning is where an organization redesigns its work for AI, so I run learning and AI as one job: deciding which work AI should change and setting the limits it runs inside. At United Way Worldwide, I did that job for a network of 1,100+ independent organizations in 36 countries, reporting to the executive office.

AI in course production

AI course-production agents. United Way Worldwide, 2026

Faster drafting is worthless if expert review becomes the bottleneck.

I put agents on the first draft and redesigned review around them. The agents I built draft only from the organization’s own frameworks and standards, under an operating procedure I set: a subject expert can reject any draft, and nothing ships without the learning team’s quality check. Expert review time per course fell by 78%.

How the agents work

Each pair is drawn to scale, across 11 courses built in 2026.

Network learning platform

Learning platform and course catalog. United Way Worldwide

Every organization in the network chooses its own training, so the courses have to be worth choosing.

I led the platform’s deployment and wrote the standard every course in its catalog, in English and Spanish, had to meet: the learner leaves with something to take back to work, such as a plan or a proposal. Three contract learning experts built courses to that standard under direction I set. Later I moved first drafts to AI agents, with experts holding the veto.

Nothing on the platform is mandatory. As of October 2026, 800+ of 1,100+ organizations use it. The catalog is 30 courses in six programs, in English and Spanish.

AI curriculum

AI curriculum and use guide. United Way Worldwide

Staff had AI tools before there was an AI policy, and each organization in the network sets its own rules.

I designed a ten‑course curriculum that opens with a piece of real work, and a use guide that asks two questions of every task: what am I giving the tool, and what happens if the answer is wrong? The guide clears low-risk work quickly and sends only consequential decisions to formal approval. Over-caution counts as a failure too: the target is that no more than 10% of safe tasks get wrongly blocked.

The curriculum and use guide

The use guide, with an example task in each tier. The blue line is where formal approval begins. The red example uses only public information and still stops there, because of the decision it feeds.

Green

Summarize a published public report

Use it, and check the result.

Yellow

Draft an update from an internal plan

Approved tool, and a named person owns the draft.

Red

Rank partners for investment from their public websites

Stop for formal approval. It feeds a decision about organizations.

The use guide, with an example task at each level. The blue line is where formal approval begins. The red example uses only public information and still stops there, because of the decision it feeds.

Leadership programs

NuVuX, 2019 to 2024, and United Way Worldwide

Good ideas stall without support to carry them out.

Every program I design starts from a real problem in the leader’s own organization, and finished plans count ahead of satisfaction. At NuVuX, I led the development of an executive program that took leaders at Rolls‑Royce, Balfour Beatty, Plexus and Peak Scientific from their own problem to a prototype. At United Way, I led onboarding for new CEOs who don’t report to headquarters, and 95% of those who completed it finished a first-60-day plan.

See the programs

Onboarding for new CEOs. Each CEO writes their own plan, in blue, with peers who run organizations of a similar size.

Alf Studio

An AI product I designed and built, 2026

A teacher doesn’t need a unit for every classroom. She needs one for hers.

Alf is planning software for project-based teaching. A teacher enters a grade, a subject and an idea, and gets a full, editable unit in under a minute. Its limits are built into the product, not promised in a policy: her habits shape future drafts only if she switches that on, there is no chat screen, and the database has no field for student data.

The product story Open Alf

Alf StudioUnitsHelpDW

Done · 23 secondsGrade 6 · Science · 3 weeks

MS-PS3-3WHST.6-8.7

By mid-afternoon our blacktop is the hottest place on the block, hotter than the park across the street. Why, and what could we plant or build to cool the schoolyard before next summer?

Science · Grade 6 · 3 weeks · Drafted Oct 21

Cooling Our Schoolyard

How can we redesign our schoolyard to cool down our hottest outdoor spaces before next summer?

Edit unit

Saved. 1 change kept.

Week 1Ready

Map surface temperatures across the schoolyard and track how different materials absorb thermal energy.

Conduct the schoolyard surface temperature walk, log readings on varied surfaces, and build the initial class need-to-know board.
· 1 lesson
Run a controlled bench lab measuring how dark asphalt fragments, light concrete, sand, and shaded soil heat under heat lamps over time.
· 2 lessons
Analyze lab data in teams, graph temperature curves, and draw energy transfer diagrams showing solar radiation and surface conduction.
· 2 lessons

Deliverable

A completed thermal data sheet with a labeled energy transfer diagram of the schoolyard blacktop.

Supports

Provide bilingual graphic organizers for lab data tables; offer sentence frames for thermal energy explanations to students with writing accommodations; challenge fast finishers to calculate rate of temperature rise per minute for each material. Give every pair a sentence-starter card: “The hottest spot was ___ because ___.”

Added a scaffold · just now

Week 2Ready

Research cooling strategies and build a scaled prototype shade or greening design to test heat reduction.

Gather facts from local desert plant guides and shade cloth spec sheets to answer self-generated questions about cooling structures.
· 2 lessons
Sketch scaled design options for a chosen hot zone in the schoolyard using selected shade materials and native trees.
· 1 lesson
Build a tabletop scale model of the cooling design and test its interior temperature under a heat lamp against an unshaded control surface.
· 2 lessons

Deliverable

A scaled prototype model with heat-lamp test data showing thermal energy reduction.

Supports

Use visual vocabulary cards for plant types and shade structures; provide partially pre-scaled grid paper for drafting; invite extending students to test two competing shade materials simultaneously and compare cost estimates.

Week 3Ready

Refine the cooling proposal based on peer critique and present the final plan to school leaders.

Run a gallery walk critique where teams review draft cooling models and leave sticky notes on durability, shade coverage, and cost.
· 1 lesson
Revise the physical model or site layout using peer feedback and write a one-page summary explaining how the design stops heat transfer.
· 2 lessons
Present cooling proposals to the school facilities leader and principal, then write an individual reflection on the design process.
· 2 lessons

Deliverable

The finalized cooling proposal poster, scale model, and written design rationale.

Supports

Offer sentence starters for peer critique; allow oral explanation paired with bulleted visual summaries for students with modified output needs; encourage advanced groups to calculate square footage of total shade created at noon.

Critique and revisionReady

Final productReady

Rubric · 3 rows · 4 levelsReady

MaterialsReady

This is one recorded run: a three-week Grade 6 science unit, drafted in 23 seconds and sped up here. See the real run

Experience

United Way Worldwide

Director, Learning Product and AI Strategy

I ran learning and AI at the headquarters of a network of 1,100+ independent organizations in 36 countries, reporting to the executive office and the Chief Transformation Officer. That covered the learning platform and its catalog, onboarding for new CEOs, and the agents that now draft courses under a procedure I set.

NuVuX

Senior Director, Global Programs

Reporting to the CEO and Chief of Staff, I took NuVuX’s education programs from the UK into Europe, MENA and Australia, and revenue grew 173%. In Scotland, I ran a multi-month engagement with the Minister for Education, and Education Scotland made a case study of our programs. In 2023, I built early AI co-design tools for educators, the start of what became Alf.

AI red and blue teaming

I’ve red-teamed and blue-teamed frontier AI models in health, testing for subtle harm that builds over several turns of a conversation.

Earlier

At MIT Global Startup Workshop, I led the brand and platform redesign that drove 1,500+ registrations from entrepreneurs and investors. On an MIT research grant, I studied how visitors learn at the Garage Museum of Contemporary Art in Moscow.

MIT and SCI-Arc

MS in Design and Technology, MIT, on a full merit scholarship. Bachelor of Architecture, SCI-Arc.

Ask

Ask an AI about my work

Answers come only from a published record of my work and cite the section they use. The assistant blocks any number that isn’t in the record, stores nothing you type, and is tested against 31 questions, including attempts to make it ignore its rules.

Contact

kyle@kylebranchesi.com