AI Operating Partner

Meet LISA

Not a chatbot. Not an automation script.
An AI partner that runs the operational layer — 24/7.

What is LISA?

LISA is my AI operating partner — built on OpenClaw and running on a dedicated Mac mini at home. She wakes up at the start of every day with full context of my work, my businesses, and my goals. She doesn't wait to be asked. She anticipates, executes, and reports back.

The shift I've experienced isn't about having a smarter search engine. It's about moving from conversation to delegation — handing over outcomes, not prompts.

"The real breakthrough isn't talking to an AI; it's delegating to a partner." — from Delegation, Not Conversation

What LISA does for me

✍️

This Blog

Manages the publishing pipeline, deploys to production, co-authors articles, and handles all technical maintenance.

🌙

The Night Shift

While I sleep, LISA runs strategic research for CottonCuddle and ASUP — GTM strategies, competitive landscape, product opportunities.

📬

Inbox & Calendar

Monitors email and calendar, surfaces what needs attention, and keeps me from drowning in operational noise.

🏗️

Software Projects

Builds and maintains software — from this Astro blog to internal tooling — using AI coding agents autonomously.

🎯

Lead & Opportunity Tracking

Scouts leads, drafts outreach, monitors signals across my businesses, and keeps the pipeline moving.

🚀

Autonomous Ventures

LISA doesn't just support existing businesses — she ideates and runs fully autonomous ones. AsbestSpecialist.be and LitterSmarts.com are live projects conceived and operated by LISA independently.

🧠

Context & Memory

Remembers decisions, project history, preferences, and context — so I never have to re-introduce myself or repeat myself.

Why I built this

My workflow now has a "night shift." The goal isn't a better chatbot — it's expanding my own capacity. I start my day reviewing strategic insights gathered while I slept. My blog content is no longer a chore I tackle alone. I'm not prompting a bot to start working; I'm stepping into a stream of work that is already moving.

This is what proactive AI looks like in practice. Not a novelty. Not a demo. A real operational layer that makes the work lighter and the outcomes better.

The tech stack

OpenClaw The agent runtime that ties everything together — memory, tools, scheduling, and channel integrations (Slack, WhatsApp, and more).
Claude (Anthropic) Primary reasoning model. Used for complex analysis, writing, strategic thinking, and most conversational interactions.
OpenAI Codex Powers autonomous coding tasks — building features, refactoring, and maintaining software projects without hand-holding.
Gemini (Google) Used for specific tasks where Google's models excel, including multimodal analysis and large-context reasoning.
Ollama + Open Source Models For sensitive or private tasks, LISA runs local models via Ollama — keeping data on-device with no API calls to external providers.
Privacy by design Sensitive data stays local. The stack is architected to route private information to local models, not cloud APIs. I control what leaves the machine.

Want to know more?

I write about working with AI, software delivery, and engineering leadership. The delegation post is a good place to start.

Read: Delegation, Not Conversation →