Set up a second brain
Turn scattered notes, files, projects, and decisions into one source of truth that AI assistants can use without inventing context.
See an operations brain in action →
Optimization Doc Start diagnosis → Two people who make AI useful.
Share this page with anyone who needs a second brain, an AI agent, a useful assistant, or a workflow that finally works.
Optimization Doc founder
Jethro Jones
AI adoption strategist, systems builder, educator, and OpenAI Select Partner. Jethro turns recurring work into practical systems people can operate and trust.
Build useful AI agents
Create research, content, monitoring, and operational agents with clear permissions, review points, fallbacks, and ownership.
View an approval-first agent →Create role-specific assistants
Build Custom GPTs and other focused assistants that help people make decisions, prepare work, and follow a repeatable method.
Try the design-thinking assistant →Redesign workflows before automating
Map the real work, remove unnecessary steps, protect human judgment, and automate only the parts that reduce meaningful friction.
Browse workflow case studies →Connect the systems you already use
Join websites, APIs, email, CRM, payments, calendars, databases, and dashboards so the AI fits the work instead of becoming another tab.
Try a live Kit analysis app →Train teams through real builds
Run practical workshops and build days where people choose a useful problem, build something real, and leave knowing what to do next.
See the Spokane build-day materials →Rescue stuck AI projects
Diagnose broken automations, brittle prompts, confusing tools, and half-finished builds, then leave behind a system the team can maintain.
See AI Tech Support →
Practical AI educator and builder
Aaron Makelky
Former Wyoming teacher and football coach who now works in tech at Descript. Aaron tests AI on real work and turns what survives into tools, notes, and workshops.
Make AI practical for a team
Turn broad AI curiosity into workshops, repeatable prompts, and small experiments people can use in real work on Monday.
Book Aaron for a practical first test →Build small tools that earn their keep
Create prompts, Codex skills, and utilities for jobs that keep coming back, then document what holds up the second time.
Explore Aaron's tools →Move from chat to agents
Help people understand when a chat assistant is enough, when an agent is useful, and what boundaries make agent work reliable.
Watch the agents walkthrough →Keep AI work portable
Design workflows that do not trap important context in one chat, one tool, or one laptop, with durable notes and handoffs.
Use Aaron's Build Week hub →Teach AI literacy
Help educators and teams give better context, ask for revision, verify outputs, and decide where human judgment still belongs.
Read Aaron's AI notes →Design assistants for everyday work
Shape safe first-pass assistants for meetings, family logistics, creator chores, school communication, and recurring planning.
See the assistant recipe ideas →Four fast ways to see how we think and build.
Jethro Jones
An AI automation dashboard anyone can run
A practical command center for schedules, webhooks, credentials, failures, and AI-assisted diagnosis.
Jethro Jones
Debugging a real AI automation workflow
A real-life example of finding a failure, inspecting the system, and using coding agents to repair it.
Aaron Makelky
GPT Realtime Slides from OpenAI Build Week
Aaron's Build Week demo shows what happens when a focused idea becomes a working AI experience.
Aaron Makelky
Stop using AI chat and start using agents
A short explanation of the shift from one-off conversations to AI that can carry out a defined job.
Send the useful introduction
Know someone who keeps saying, “We should be doing more with AI”?
Send them this page. If they need a system built, you'll probably start with Jethro. If they need a practical first experiment or workshop, you'll probably start with Aaron.