One team, three flags, five days.
We're here to win the Badger Conference Challenge. AU, NZ and RSA pulling in the same direction across the three themes of the conference. Pick a theme below to jump in, or head to Team & Files for the pre-read library and team uploads.
The three themes
In Brad's words: "in order of weight, on AI, breaking down silos, and execution."
AI
From experiments to default. The AI exercise drives cross-brand cooperation and forces execution, specifically around AI. Where we are on the curve, who's shipping, and how to lift the whole group.
Breaking Down Silos
Why the group plays smaller than the sum of its parts. Pixar case study, "3 Types of Silos" pre-read, Shawn Carr's exercises that turn theory into actions we can measure once we leave the room.
Execution
From decisions to delivery. Aviva AI Claims case study, Shawn Carr's exercises designed to translate the room's intent into measurable actions across our brands.
The Buildathon
Each table presents for 15 minutes. One real business constraint, the current workflow, the redesigned workflow, the AI application, the measurable outcome, and a concept or prototype. Prizes for top three tables. Adjudication is practical, not flashy. This is where we win or lose.
Monday opener (non-competitive)
Two hours on Road to 2030 (group results, mission, vision), then two hours on People 2030 (Badger Leadership principles, engagement survey). The themes kick in from the afternoon onwards.
AI
From experiments to default. The AI exercise sits alongside Silos and Execution but carries the most weight in Brad's framing.
What Brad has asked of us
Spend a minimum of four hours a week learning, playing with, and building real fluency with Claude. Treat it like any other strategic priority. The execs who lean in now will be the ones shaping how Badger operates in the years ahead.
Sandbox VMs with the full toolset (Claude Cowork, Code, Projects, Perplexity) were provisioned in early April. Enterprise access via ARC.
The maturity curve the table has landed on
Prototype freely with Claude. Prove the concept. Then hand to IT to industrialise properly. Burning tokens to find and understand data is the biggest cost; once you know the logic, the build becomes far leaner. This is shaping how we're thinking about the Buildathon build.
From the Lemonade thread (Ryan Passanah)
The mission Ryan has named is AI-first to AI-only: Corné's API layer into the policy admin system, Jo's all-data-for-decisioning layer, and digital transformation across call centre, marketing, web and supporting functions. Done as one project, with one brain trust, one accountability line.
Three moves we could argue for
- Default Claude / Cowork access for every knowledge worker, with skills installed.
- An "AI lead" inside every BU, not just centralised.
- One measurable productivity goal per function per quarter.
Working notes
Breaking Down Silos
Why the group plays smaller than the sum of its parts — and the moves that change that.
The problem, in one line
We have the right pieces in AU, NZ and RSA. They're often optimising for their own P&L instead of the group's. Shawn Carr will lead the room through exercises that turn theory into actions we can apply, execute and measure once we leave.
The pre-reads
- "3 Types of Silos That Stifle Collaboration — and How to Dismantle Them" (HBR-style PDF, attached to Brad's 18 April email)
- Pixar Silo Case Study + the conference group questions (Brad's 19 April email)
Both live in the Pre-read library.
What we're bringing to the table
- Cross-brand customer journeys — where do we hand off badly?
- Shared services that should exist but don't
- Decisions stuck because no one owns them at group level
Working notes
Execution
Translating the room's intent into measurable actions across our brands.
Why execution is its own theme
Brad's framing: "study the material, discuss it, dwell on it. Come prepared. Shawn Carr will lead us through exercises that turn the theory into actions we can apply, execute and measure once we leave the room."
This is the theme that makes the other two stick. The room can agree on AI and silos all it wants — without an execution layer, nothing changes by Monday.
The pre-read
Aviva AI Claims Case Study (PDF, attached to Brad's 18 April email). Use this as the worked example of getting AI from concept to production-grade execution. Lives in the Pre-read library.
The 2030 backdrop
Monday morning sets the frame: Road to 2030 on group results, mission and vision, then People 2030 on Badger Leadership principles and engagement survey results. Execution exercises follow that.
Working notes
The Buildathon
15 minutes on stage. One business constraint solved with practical AI. Adjudication is practical, looking for projects that add the best value and address a table-wide challenge or opportunity. Prizes for top three tables.
The project: Pet Community App
A brand-agnostic pet community app built around the three group objectives below. Universal use-case first; territorial iterations follow. No insurance branding, no early marketplace scope creep.
Starting point: Jo Ferolli's prototype (Dotsure) — embedded on the Artifacts page. Tiaan leading the build, Rick alongside. Locked in by NZ, AU and RSA on the table thread, 15 May 2026. Stays under wraps from other tables until the room on Friday.
The three objectives (Glen)
Attract
Pull pet owners into our orbit through value they can't get elsewhere — content, community, useful tools — before they're shopping for insurance.
Retain
Make the app the thing they open weekly, not yearly at renewal. Stickiness comes from utility plus identity, not from policy reminders.
Optimise
Every interaction is a signal. Use it to lower CAC, improve underwriting, surface upsell moments — and to learn faster than competitors.
Scope discipline
Glen's principle, agreed by the table: be specific about the universal use-case first, then get territorially specific in v2.
Tempting scope creep (parked, not dead): a marketplace for pet services, vet booking, breeder directories, classifieds. We can build out the niches over time. For the conference room, the muscle to demonstrate is restraint plus clarity, not breadth.
Inspiration set
Sites Glen flagged for the look-and-feel and feature thinking.
- petsly.co.za — RSA pet community
- loyal.com — utility-led pet wellness brand
- yapperapp.co.za — RSA community app pattern
Jo's prototypes
The three HTML files Glen shared on 15 May (Community App, Go-to-Market, original mockup) live on the Artifacts page, embedded inside the team site so they're easy to review without leaving and so Badger's email gateway can't strip them.
The 15-minute pitch structure (Dylan's brief)
The constraint
What business problem, friction point, bottleneck or silo are you solving?
The current workflow
Where does the process create delay, duplication, cost, risk or frustration today?
The redesigned workflow
How should the process work before AI is applied?
The AI application
Where do AI, automation, Claude or agents improve the workflow?
The business outcome
What measurable value does this create for the business?
The concept or prototype
Prototype, MVP, workflow model, or clearly articulated concept. Does not need to be production-ready.
The brief in Dylan's words
"Bring a real problem, not a theoretical one. Existing Claude work is encouraged. Simpler workflows usually win. Strong business outcomes matter more than flashy demos. ARC will capture key learnings across the tables so the work continues beyond the conference."
And Brad's adjudication note: practical, table-wide challenge or opportunity, best value. Not theatrical.
Our working detail
This is the team's working area for fleshing out each of the six pitch sections. Everything stays inside the team.
Pre-reads & references
- Dylan's full Buildathon brief (6 May)
- Anthropic enablement webinar (Goldcast)
- YouTube intro videos (x2)
- Aviva AI Claims case study (execution worked example)
- DBS AI Journey pre-read (AI worked example)
All linked from the Pre-read library.
Bonus point reminder
Dylan: "Bonus points if your workflow does not accidentally create Skynet 😉"
Artifacts
A site within the site. Embedded HTML artifacts — prototypes, inspirations, experiments — loaded inside this team frame so we can review them without leaving.
How this works
Click any card below to load it into the viewer at the bottom of the page. Each artifact stays inside this site — no tab switching, no email-gateway dramas. Newest at the top.
Today's set is Jo Ferolli's three Dotsure-branded files — our inspiration starting point. When the team's own brand-neutral build emerges, it'll land here too with a Team build tag.
New artifacts are added via the repo, not from here. Got something to add? Ping Tiaan or open a PR.
Team & Files
Who's on the table, what's already been sent to us, and where to drop material we create.
Roster
Click any name or role to edit it inline. Edits persist in your browser.
Curated list of conference material sent to date. Click any item to jump to the source (Gmail thread, SharePoint folder, or public URL). Add-to-library functionality is on the roadmap — for now, drop a note in Slack and we'll add it manually.
| Material | Theme | Type | Source | Received |
|---|
1. Pick the bucket this file belongs to. 2. Drop the file in. Allowed types only — anything else is rejected.
Uploaded files
| File | Bucket | Uploaded by | Date | Size | |
|---|---|---|---|---|---|
| No files yet. Drop one above. | |||||