Small AI apps
Focused tools that solve one annoying problem — like checking product availability. Built in days, not sprints.
- claude-code
- apps
~/senssense
I'm Sen. I build practical AI tools for real problems — a store inventory checker, a video automation flow, a quoting tool — and share what I learn on YouTube.
/about
I spent years shipping product in ecommerce, which taught me one thing: consistency beats brilliance. Now I apply that to content — building AI pipelines with Claude Code that produce websites, automated content workflows, and local apps for small businesses. Everything I share on YouTube comes from systems I actually run.
/videos
New builds and honest tests, walked through end to end. Watch here, or subscribe on YouTube so you catch the next one.
newest
Solo founder, PM, mom, MBA student — one toolbox holds all of it. The full daily AI stack: Gemini, Claude Code, GitHub/Netlify/Cloudflare, fal.ai, OpenRouter, Notion. No sponsors, no secrets.
subscribe on YouTube →/systems
Focused tools that solve one annoying problem — like checking product availability. Built in days, not sprints.
Pipelines that pull, filter, and publish on a schedule — locally-targeted content that shows up when AI assistants answer questions in your area, or a research digest that writes its own first draft before you sit down.
Idea → script → storyboard → finished video, with quality gates at every step. A repeatable line for short-form and long-form, built around your voice, not a template.
/services
01 · start here
A working session on your current process: where AI fits, what to automate first, and a concrete build plan you can execute with or without me.
02 · full build
I design, build, and hand over a working system for your business — an automation, an internal tool, a content pipeline — with documentation your team can actually run.
03 · self-serve
The exact prompts, workflows, and templates from my videos — free in the resource library. Start here if you want to build it yourself.
/work
real estate · local SEO/AEO
An automated Google Business Profile content system for a real estate brand covering 11 towns — locally-specific posts generated, reviewed, and published on schedule.
ecommerce · short-form
An AI-assisted short-form content strategy for a kitchen brand — planning, hooks, and batch production built around the products people actually search for.
/faq
Small businesses that need to publish constantly but don't have a marketing team or a content team — local services, real estate, and ecommerce brands. I'm based outside Boston and work remotely with clients across the US, plus English- and Chinese-speaking teams. If your bottleneck is publishing consistently, or the same manual task eats an afternoon every week, that's my lane.
Start with the task you repeat every week that has a clear input and a clear output: publishing local content, drafting product copy, turning research into a first draft, checking inventory. Then ask the harder question — in that workflow, what breaks first? Messy data, no approval step, a tool that won't connect to anything. That answer gives you your obstacle list. A pipeline audit covers both sides: what to automate, and what will stop you from automating it, ending in a build plan you can execute with or without me.
Yes, but not the way most people mean. AI doesn't cold-message strangers for you — it makes you findable, then makes follow-up consistent. For a Boston-area real estate and contractor team I built a pipeline that publishes locally specific Google Business Profile posts across 11 towns on a schedule, so the business shows up when people and AI assistants ask questions about those towns. On the outbound side, I build lead-qualification pipelines: pull everyone who engaged with a post, score them against your ideal customer profile, and draft outreach you approve before anything sends.
Yes — English, Traditional Chinese, and Simplified Chinese. I run my own YouTube channels in English and Traditional Chinese, and I've built pipelines that handle both languages end to end, including a dubbing system that turns a Chinese talking-head video into an English one with matched voice and lip-sync, with a human review gate before anything publishes. If you operate in both markets, you don't need two separate workflows.
AEO (answer engine optimization) and GEO (generative engine optimization) are about being the source an AI assistant cites when someone asks it a question — in ChatGPT, Claude, Gemini, or Google's AI Overviews. Traditional SEO competes for a click on a results page; AEO competes to be the answer itself. For a local business it usually means specific, factual, well-structured pages about the exact things customers ask, published on a schedule rather than once.
No. I build the pipeline and hand it over with documentation your team can actually run. Your job is review and approval — the system handles drafting, formatting, and scheduling, and every pipeline I build has a human approval gate before anything publishes. Most clients work through a spreadsheet or a Notion page, not a terminal.
Often you can, and I'll tell you when that's the right call — no point paying me to rebuild something you can buy. Off-the-shelf tools work well when your process is standard and your data is clean. They start failing when the work needs judgment: deciding which of 11 towns a post is really about, catching a claim that shouldn't publish, rewriting output that doesn't sound like you. That judgment layer is where a custom pipeline built with Claude Code earns its cost. The connectors are the easy part.
For a lot of workloads, yes. I run local models through Ollama, and I can build pipelines where the sensitive steps — customer records, internal documents, pricing — never leave your hardware, while the steps that genuinely need a frontier model still call out to one. It's a real tradeoff: local is cheaper and more private, cloud is smarter and faster. Part of my job is telling you which steps actually need which.
The audit is free. We spend a session walking your current process, decide what to automate first, and agree on one measurable target before any code exists — posts published per week, hours returned to your team, quote turnaround cut from days to hours. You leave with that plan whether or not you hire me. If you go ahead, builds are a single fixed project fee, typically $500–$5,000 depending on scope. No hourly billing, no retainers, and no token or usage costs passed through to you — one price for one project. Send me a description of the problem and I'll reply within two business days.
Yes. Every document from my videos — prompt libraries, workflow templates, checklists — is free. One email address unlocks the whole library, and you can unsubscribe anytime. Signup runs on my own site, not a third-party email platform, and I don't sell or share the list.
/contact
Describe the problem you want to solve — an automation, a small app, a content system — and what's in the way. I read everything and reply within two business days.