↑↓to navigate
A Proposal for Head of AI Innovation & Transformation, ACS Marketing

AI That Fights Cancer
at the
Speed of Mission

Building the AI capability that turns experimentation into measurable fundraising impact—and makes every marketing team permanently more capable.

Bill Shipman
Director, Business Engagement, Data & Analytics
Newell Brands • $8B global consumer goods
Proposed for
Head of AI Innovation & Transformation
American Cancer Society, Marketing
The Opportunity

AI could dramatically expand ACS’s ability to reach the right person at the right moment.

ACS has invested $5.4 billion in cancer research since 1946. More than 1.9 million Americans will be diagnosed this year. The gap between those two numbers closes faster when the right donor hears the right message at the right moment. That’s where AI comes in.

More of the right donors give more
Lapsed Relay captains, event participants who plateau, one-time givers who never convert to recurring—each gap has a data signal. The hypothesis: identifying and re-engaging the right people at the right moment could directly increase dollars raised and research funded.
Cancer.org traffic becomes mission fuel
People arrive at cancer.org in a moment of vulnerability or purpose. The hypothesis: contextually matched fundraising prompts—tied to the content they came to read—could convert that intent into mission support far more effectively than generic placement.
AI advantage compounds
Every successful experiment creates reusable workflows, better data, stronger internal skills, and faster execution. The organizations building that capability now will be harder to catch later.

The Missing Piece

The gap isn’t AI access.
It’s ownership.

ACS already has the ingredients. What’s missing is a role accountable for turning them into measurable marketing capability.

AI Innovation & Transformation owns this pipeline
Business
Problem
Prototype
Prove
Value
Scale &
Hand Off
Partners: Marketing · IT · Legal / Privacy · Digital Product / Data

Someone has to own the space between a good AI idea and a capability the organization can actually use.

The Operating Model

Problem → Workflow → Capability → Technology

Start with the problem, not the platform. Define the outcome, redesign the workflow, identify the capability required—then choose the technology that enables it.

01
Outcomes first, tools second
Every use case starts with a value hypothesis tied to a fundraising or marketing KPI. If we can’t articulate the mission outcome before we build, we don’t build.
02
Prove before you scale
Rapid pilots, real measurement, honest results. A use case that saves 10 hours a week beats a 60-slide strategy that never ships. Speed and credibility compound.
03
Transition, don’t hoard
Every proven solution ships with documentation, training, and a named owning team. The goal isn’t to be indispensable. It’s to leave every team more capable than I found it.
Proof of Concept

I’ve built this operating model before.

Over the past two years, I’ve helped build and scale AI capability inside Newell Brands, a 25,000-person global company. The playbook is proven. Now I’d adapt it to ACS’s mission.

0 to 7K
AI users activated in 18 months
Helped take Newell from zero active AI users to 7,000 in a year and a half. Built the enablement infrastructure that made it stick: training programs, adoption tracking, and a champion network across every major function.
1,000+
Employees trained in a single live event
Designed and hosted large-scale AI training events reaching over 1,000 employees in a single session. Built Quantum Academy, a self-serve learning platform that keeps skill-building going between live events.
Quantum
Academy
AI learning platform, built from zero
Designed and launched a full self-serve AI learning platform for a 25,000-person organization: structured curriculum, live events, and on-demand content. The kind of infrastructure ACS would need to make AI capability stick beyond any single role.
10+
AI tools and agents in production
Helped stand up more than ten live AI agents and data tools—from natural-language data assistants to automated workflow agents—alongside the governance needed to deploy them responsibly.
The Plan

What I’d build in the first 90 days

No sweeping strategy decks before week four. Three concrete phases: discover, pilot, scale.

Illustrative — Final priorities set during discovery

Days 1–30
Discover
  • Meet every ACS marketing team lead: Relay, Making Strides, Digital, Helpline, Campaigns
  • Pull event retention data: captain reactivation history, P2P fundraiser lift curves
  • Review site analytics for highest-traffic pages with no donation CTA
  • Shadow a Relay captain call and a Making Strides event briefing
  • Define the mandate with CMO, IT, Legal and Digital Product — decision rights, handoff points, budget and success measures
  • Score top 10 use cases by mission impact, data readiness, and lift potential
  • Present prioritized pipeline to CMO with dollar-raised estimates
Days 31–60
Pilot
  • Run the Relay captain reactivation model on one region with real lift measurement
  • A/B test a P2P fundraising coach for Making Strides participants stuck under $150
  • Launch one cancer.org CTA experiment on a high-traffic, high-intent page
  • Partner with IT and Legal throughout, not after the fact
  • Identify first AI champion cohort from across marketing teams
  • Document every model, result, and handoff spec for team ownership
Days 61–90
Scale
  • First value ledger delivered to CMO: donors reactivated, dollars raised, research funded
  • Relay captain model handed to the Relay team, documented and fully owned by them
  • Champion network launched across Relay, Strides, Digital, Helpline, and Campaigns
  • Q2 use-case pipeline presented, framed in dollars raised and research funded
  • Operating cadence and quarterly business review template live
  • Every pilot documented for the next person who runs it
The Pipeline

Six hypotheses I’d test in year one

Starting points, not assumptions. Each would enter the pipeline with a value hypothesis and earn its way into a pilot.

Relay Captain Reactivation
Hypothesis: a propensity model trained on historical captain behavior could score lapsed captains and trigger personalized win-back sequences that outperform generic re-engagement.
Relay for LifeRevenue
Live Prototype →
Peer-to-Peer Fundraiser Coaching
Hypothesis: real-time coaching prompts triggered by fundraising behavior signals could meaningfully increase per-participant totals for active event participants.
Making StridesRelay
Digital Fundraising Intent
Hypothesis: contextually matched prompts tied to why someone visited cancer.org could convert informational intent into mission support more effectively than static placement.
New CapabilityHigh Volume
Email Personalization at Scale
Hypothesis: dynamic content models that adapt message, offer and timing by donor segment could lift engagement and giving rates over batch-and-blast campaigns.
RetentionRevenue
Making Strides Plateau Detection
Hypothesis: identifying participants who historically plateau below their potential and intervening at the right moment could recover dollars that currently go unrealized.
Making StridesNew Revenue
ACS ACTS Discovery
Hypothesis: context-aware experiences on cancer.org could help more appropriate patients discover and enter ACS’s existing clinical-trial support pathway.
Patient ImpactMission
Live Prototype →
Scaling Adoption

Building what survives me

The champion network is how AI capability outlasts any single role. Every proven solution becomes a repeatable pattern the team owns, not a dependency on the person who built it.

1
Identify 1–2 champions per function
Pick people who are already curious and influential. They don’t need to be the most senior. They need to be the ones others listen to.
2
Equip with mission-specific playbooks
Give champions pre-built tools for their specific context, not generic training. Practical wins in week one build lasting credibility and momentum.
3
Run a monthly AI wins forum
Visible social proof drives adoption faster than training. A short monthly forum where wins get shared and named moves behavior at scale.
4
Measure behavior, not training attendance
Track active weekly AI use per function. Seat counts and completion rates are vanity metrics. Behavior change is the goal.
Embedded
Champions
Across key marketing functions
AI Wins
Forum
Share what works
Adoption
Measurement
Track behavior, not attendance
Documented
Handoffs
Teams own what scales
Accountability

A value ledger the CMO can read

AI investment without measurement is a belief system. Every initiative carries a before/after measurement and posts actual results to a running ledger visible to leadership every month.

Year 1 Use Case Pipeline — Illustrative Ledger Framework
Hypothesis set before build; actuals tracked after. Baseline measured at launch.
Use Case Value Hypothesis KPI Baseline Actual Decision
Relay Reactivation Personalized win-back outperforms generic re-engagement % captains reactivated Measured at launch Updated monthly PENDING
P2P Coach AI-drafted next asks raise avg fundraised amount Avg $ raised per participant Measured at launch Updated monthly PENDING
cancer.org CTA Contextual prompts lift donation conversion from organic traffic Donation conversion rate Measured at launch Updated monthly PENDING
Monthly Upgrade Propensity model improves one-time to monthly conversion Monthly donor conv. rate Measured at launch Updated monthly PENDING
Helpline Intelligence NLP on transcripts surfaces content gaps for editorial and marketing Content gaps identified / actioned Measured at launch Updated monthly PENDING
Decision path:
PILOT → SCALE / ITERATE / STOP
The value ledger is updated monthly and reported to the CMO. Initiatives that don’t deliver against their hypothesis get deprioritized, and the learnings feed the next cycle. No zombie projects. Decision options: Pilot → Scale → Iterate → Stop.
Why Me. Why Now.

I’ve built this before.
Now I want to build it for a mission that matters.

bill.shipman@gmail.com
"Cancer information, answers, and hope. Available every minute of every day." - ACS