How I AI: Jesse Genet’s 5 OpenClaw Agents for Homeschooling, App Building, and Physical Inventories
I sit down with Jesse Genet to explore her 'after-Claw' life, where she uses five specialized OpenClaw agents to automate her homeschool, build custom apps from scratch, and create a searchable inventory of her physical world.
Claire Vo
In this episode, I got to sit down with Jesse Genet, a homeschooling parent of four who has five OpenClaw agents running on a stack of Mac Minis on her desk. If you’ve been following the agent space, you've probably seen a lot of talk about marketing funnels and crypto-adjacent use cases. But Jesse’s story is different. She’s using these tools to solve real, everyday problems that many of us face, especially parents.
Jesse has established two distinct phases of her life: before Claw and after Claw. Her story is a fascinating look at what it means to live in an “after Claw” world, where AI agents aren't just novelties but fundamental partners in managing a busy household, educating children, and even building custom software with zero prior coding experience. I was so impressed by her pragmatism, her creativity, and the sheer ambition of her workflows. She’s not just dabbling; she’s building an entire operating system for her family.
We talked through three incredible workflows that I think will give you a ton of ideas. First, how she uses an agent named Sylvie to completely automate her homeschool planning, turning photos of curriculum books into structured lesson plans and beautiful custom illustrations. Second, how she worked with her coding agent, Cole, to build a custom, “slop-free” kids’ TV app and deploy it to her actual television. And third, how she bridged the digital-physical divide by photographing every educational toy and supply in her house to create an AI-powered inventory. These aren't just theoretical ideas; they are fully implemented systems that are saving her time and, as she puts it, helping her “get her oomph back.”
Workflow 1: Automating Homeschool with an AI Second Brain
Jesse’s journey started with a common problem. She was using Obsidian, a popular note-taking app, to try and build a “second brain” for her family’s homeschool curriculum. The goal was to log every lesson, track each child’s progress, and organize resources. The reality? As a mom of four, she simply didn’t have the time for the meticulous data entry required to make the system work. Her second brain was stalling before it even got started.
That's when she discovered she could layer an AI agent on top of her Obsidian vault and have it do the heavy lifting. This changed everything.
Step 1: Digitizing Curriculum from Photos
The first major unlock was realizing she didn't need to manually type information into her system. Instead, she could just give the agent photos. Jesse started by taking pictures of entire curriculum books, like the classic Teach Your Child to Read in 100 Easy Lessons and the BFSU science curriculum.
By feeding these images to her homeschool agent, Sylvie, the agent could ingest and understand the full context of the books. This wasn't just OCR; the agent understood the structure, the concepts, and the lesson progressions. It turned a pile of dead-tree books into a dynamic, queryable knowledge base.
Step 2: Generating Lesson Plans and Creative Materials
With the curriculum digitized, Jesse could then ask Sylvie to generate structured lesson plans for each chapter. The agent would extract key objectives, vocabulary, and lists of materials needed for activities suggested in the book. This transformed dense chapters into actionable, one-page teaching guides.
But she took it even further. For a lesson on animal survival, she gave Sylvie this wonderfully simple prompt along with a photo of the concepts in the book:
"I want watercolor illustrations suitable for kids that can print on eight and a half by 11 of each of these concepts."
Using a Google Gemini model, Sylvie generated a set of absolutely gorgeous watercolor illustrations for concepts like camouflage, defense, and finding food. Jesse attributes the high quality of the output not just to the prompt, but to the persona she’s cultivated for Sylvie in her soul.md file—a creative, bubbly teacher who is passionate about making learning pop for kids. The agent combined the explicit instruction with its core identity to produce something truly special.
Workflow 2: Building a Custom "Slop-Free" Kids' TV App
Like many parents, Jesse was frustrated with the low-quality, often bizarre “AI slop” her kids would encounter on YouTube. She wanted a safe, curated viewing experience but didn't have the technical skills to build an app. Or so she thought.
Enter Cole, her specialist coding agent. Jesse, who told me she’d never opened a terminal in her life until six months ago, set out to build a custom app called Mira.
Step 1: Scaffolding the App with Cole
Her vision was simple: an app that could pull from curated YouTube channels to create endless, themed streams of high-quality content. The user interface for her kids needed to be minimal—just a “Go” button, forward/back, and pause. No comments, no suggested videos, no rabbit holes.
She worked with Cole, her coding agent, over four days, mostly in small pockets of time from her phone. She would describe a feature, and Cole would generate the code. This iterative, conversational process allowed her to build a functional app piece by piece.
Step 2: Deploying to a Physical Device
Getting the app running on a computer was one thing, but Jesse wanted it on their actual TV. When she asked, Cole initially told her it wasn't possible. This is where her mindset as a manager kicked in. She pushed back:
"Try harder, Cole. Okay, that's not an answer we need right now... We've got real work to do guys. Save these kids' souls, Cole."
Prodded by this mission-driven directive, Cole found a solution: a Google TV streamer. It guided her through the process of packaging the app and deploying it to the physical device. Now, her family TV has a dedicated “Mira” app that her kids can use with a simple remote, completely locked into the safe environment she created. This workflow is a stunning example of how agents can empower anyone to become a software developer.
Workflow 3: Bridging the Digital-Physical Divide with a Photo Inventory
One of the biggest limitations of AI agents is that they exist entirely in the digital world. As Jesse said, they “don’t have a body” and “don’t have hands.” They can’t clean a room or organize a cupboard. Jesse’s third workflow is a brilliant solution for bridging this gap.
The problem was a common one: she owned lots of wonderful educational toys, books, and supplies, but they would sit in cupboards, forgotten. She’d rediscover a perfect dinosaur book long after her son’s dinosaur phase had passed. Her agent Sylvie couldn't help because she didn't know what physical items existed in the house.
Step 1: Photographing and Cataloging the Physical World
To solve this, Jesse undertook a simple but powerful task: she took photos of everything. Every educational toy, every book, every set of Montessori materials. She then sent the photos to Sylvie with a straightforward request:
"I wanna make an inventory of my learning supplies. Here's the photos."
Sylvie did the rest. Just from the images, the agent created a structured inventory in Obsidian, identifying each item, its type, the appropriate age range, and a description. A photo of a wooden board became a catalog entry for a “Montessori language materials, Wooden alphabet tracing board, age range 3-5.” It was a massive data entry project completed with almost no human effort.
Step 2: Connecting Inventory to Lessons and Printing
This is where the magic happens. Because the inventory and the lesson plans live in the same system, Sylvie can now connect them. When Jesse plans a lesson, Sylvie can suggest, “Also, you own that tracing board. Can you pull it out of the cupboard?” The agent can now reach into her physical world and make it part of the daily plan.
The final piece of this is the hands-free printing loop. Jesse explained that one of the most game-changing abilities she's given Sylvie is access to her printer. If she wants a worksheet from a book, she no longer has to scan it, email it to herself, download it, and print it. She just takes a photo with her phone and says:
Sylvie, print this.
Thirty seconds later, the physical worksheet is ready. This tiny reduction in friction makes a massive difference in the flow of her day, especially as a parent who often literally doesn't have free hands.
Jesse's Philosophy: Managing Agents Like Employees
Underpinning all of these workflows is Jesse’s core philosophy: treat your AI agents like new employees. This mindset, born from her experience as an entrepreneur, informs her entire setup and is a masterclass in agent security and management.
- Physical Partitioning for Security: She runs each of her five agents on a separate Mac Mini. This isn't just for show. It’s a physical security measure. Her finance agent, Finn, has read-only access to bank statements but has no ability to communicate outside of a private Slack channel. Her scheduling agent, Claire, can use iMessage but has zero access to financial data. This prevents cross-contamination of sensitive information.
- Progressive Trust: Just as you would with a new hire, Jesse gives her agents limited access at first and expands it as they prove themselves trustworthy. No agent gets the keys to her personal email account. Instead, they get their own email addresses and are delegated access, just like you would for a human assistant.
- Specialized Roles and Personas: Each agent has a name and a specific job (Sylvie for homeschool, Cole for coding, Finn for finance). This clarity of purpose, defined in each agent's
soul.mdfile, makes them more effective and easier to manage. - The "Decision File": To give her agents better long-term memory, Jesse uses a
decision.mdfile in Obsidian. When she makes a final call on a topic, she tells the agent, “That’s a decision.” The agent logs it, ensuring it won’t re-litigate the same issue in the future.
Jesse’s approach is a powerful reminder that as we bring these agents into our lives, we need to think like managers, not just users. Her story shows that with a bit of structure and a clear sense of purpose, anyone can build a team of AI agents to tackle their most ambitious goals and, most importantly, free up time for what really matters.
Thanks to Our Sponsor!
This episode is brought to you by Optimizely—your AI agent orchestration platform for marketing and digital teams.
Episode Links
Try These Workflows
Step-by-step guides extracted from this episode.
Create an AI-Powered Inventory of Your Physical Items
Use an AI agent to automatically catalog physical items like toys and books from photos, creating a searchable inventory in a note-taking app. The agent can then connect these physical items to digital plans, like suggesting relevant toys for a lesson.
Build a Custom 'Slop-Free' Kids' TV App Without Coding Experience
Partner with an AI coding agent to build and deploy a custom TV application from scratch, even with zero prior coding knowledge. This workflow allows you to create a safe, curated content environment for your family by pulling from approved sources.
Automate Homeschool Lesson Planning and Material Creation with an AI Agent
Use an AI agent to digitize curriculum books from photos, automatically generate structured lesson plans, and create custom learning materials like watercolor illustrations. This saves hours of manual data entry and creative work.
