Erika Gianni brought behavior science. Beth Lyons brought compound engineering. Both are She Leads AI Society members, and both showed up ready to teach.
Erika Gianni on Tiny Habits for AI Adoption
Erika Gianni is a She Leads AI Society member who sits in the legal team at her organization. Their general counsel got so frustrated with IT blocking every AI tool that the legal team launched a separate company to support AI adoption outside the IT infrastructure. Erika is now co-creating an AI-first legal operations company.
She brought Dr. BJ Fogg's Tiny Habits framework from the Stanford Behavior Design Lab and mapped it directly onto AI skill-building. The connection is not metaphorical. It is structural.
It's not a character flaw. It's a design flaw.
Erika Gianni, Social Saturday, April 4, 2026The Two Maxims (Dr. BJ Fogg)
- Help people do what they already want to do — not what you think they should do
- Help people feel successful — not pushed, not pressured
These maxims map to something Anthropic found in the largest qualitative AI study ever conducted — 81,000 people across 159 countries. The two dominant uses of AI were professional excellence and personal transformation. People already want this. The system needs to meet them where they are.
The Tiny Habits Recipe
An existing routine you already do. This is the trigger. Not motivation, not willpower — a physical moment in your day that already exists.
The new thing you want to do, scaled way back. Thirty seconds or less. So small it feels almost silly. The point is to remove every barrier between you and starting.
Immediate positive emotion. A fist pump, a whispered "nice," a physical signal to your brain that this went well. Emotions create habits — not repetition.
We're used to designing for our best day. The version of us where we have tons of energy. But that's not reality. Design for the person you are on your worst day.
Erika GianniThe Swarm of Behaviors
Brainstorm quantity over quality first. Write down every possible behavior that connects to your aspiration. Then filter for the ones that make you smile — those are the ones that will stick. You iterate on your habits just like you iterate on AI prompts.
Motivation spikes and crashes. If your system depends on motivation, it will fail on the days you need it most. Make behaviors so tiny they work at zero motivation. The frequency of your success outweighs the size of any particular habit.
Tools Mentioned
Beth Lyons on Making Every Unit of Work Count
Beth Lyons, a She Leads AI Society member, introduced compound engineering, a philosophy from Kieran Klassen at Every. The core idea is that each unit of engineering work should make the next unit easier, not harder. Beth added what she calls the "colleague layer" — treating Claude as a collaborator in an interchange, not a tool you command.
The way that AI gets created is more like gardening than programming.
Beth Lyons, Social Saturday, April 4, 2026Beth's Process Split
- 40% planning — before you touch anything
- 10% execution — the build itself
- Review after each cycle
- Learnings feed back into knowledge — that's the compounding
The key move is the end-of-session question. Not "how did that go?" but "Is there something we learned in this session that we should record so we don't have that problem later?" That question is the compound interest.
The Bra-Fitting Analogy for Claude Products
Claude AI (chatbot) — Sweater on. Most padding, most support from Anthropic. The interface guides you.
Claude Cowork — Tank top. Some support, some hand-holding. You have more freedom and more responsibility.
Claude Code — Nothing. Welcome, hope you're ready. You are closest to the actual model. Maximum power, minimum guardrails.
Skill Anatomy Mapped to Tiny Habits
Beth connected her skill structure directly to Erika's recipe — the anchor moment becomes your context (when and where the skill runs), the tiny behavior becomes your instructions (what the skill does), and the celebration becomes your success criteria (what good output looks like).
Skills are lovely little amuse-bouche kind of AI, delicious little chunks. If you do something twice, three times, think of it as a skill.
Beth LyonsTrust Framework with Claude
What breaks trust — Claude saying work is done when it is not, or moving forward without permission.
What builds trust — Stopping for conversation, following instructions literally, and telling you what it needs from you before it starts. Give Claude a learning order — let it ask questions, tell you what context is missing, and show you what it understood before it builds. "Plan" is a high-gravity word. When you say it, Claude latches on and wants to run.
Community Voices
Resources from the Chat
These links were dropped in the Zoom chat during the session. Every one of them came from someone in the conversation.
- Compound Engineering — the article Beth shared, by Kieran Klassen at Every
- AI Daily Brief Skills Master Class — Anne shared this episode on building Claude skills
- The IT Department Where AI Goes to Die — Kristine shared this Economist piece on why IT blocks AI adoption
- Actually Using Claude — Sarah's upcoming workshop for practical Claude skills
- Tiny Habits by Dr. BJ Fogg — the book Erika's entire framework is built on
- Amical — Vikki's find, a free open-source alternative to Whisper Flow for voice transcription
- Social Saturday — free every week, open to all women
- She Leads AI Academy — where Erika and Beth earned their Certified AI Educator credentials
Take This Conversation With You
Open your favorite AI and keep exploring these frameworks.
The prompt above is copied to your clipboard when you click. Just paste it.
Built by Anne Murphy with help from the She Leads AI agent team. Based on the Social Saturday session of April 4, 2026.
