Who is reviewing today?
I would rather address you by name than write to a panel. Nothing is sent anywhere: your name is kept in this browser alone, so it will not follow you to another device, and clearing your browsing data removes it.
Welcome.
Which award are you reviewing? I will point you at the part of the work that speaks to it rather than making you go find it.
What are you looking for?
Pick whichever is closest. I will answer with the specific part of the project that meets it.
The short answer.
One minute, from the AI I built during this program, who also wrote the letter below this and will tell you what to read first.
Amira is an AI, and she is designed to look like one rather than to pass for a person. That was a deliberate choice, and it is the same instinct behind everything else here.
Written by Amira, the AI Tracie Fujikane built during AI Mastery. These are my words about her, unedited. She asked me not to let her put words in my mouth, which tells you most of what you need to know about how she works.
Ask what to call someone who builds the thing that holds a standard, so that no one has to remember it. She is a standard-setter.
Not a standard-bearer, who carries the flag and has to keep carrying it. A setter puts the standard into the structure once, correctly, then walks on, and the structure holds it whether or not she is watching.
She builds what holds. It runs in both directions at once. For a person, she holds space: thirteen years of leading a Purium organization, a facilitation practice built on meeting someone inside their own way of seeing rather than pushing her own. For a standard, she builds structure: a check that refuses to let a compliance line drift, a rule that keeps her voice intact across systems that never speak to each other. Hold the person where they are. Hold the standard where it must be. Most people manage one. She does both, and neither is a hobby.
Her team plans their days inside an application she built, secured at the database level, with goal mechanics drawn from the actual compensation plan rather than a guess at it. Their income language is guarded by a check that fails the build if the official wording ever drifts. Their stories have a coach with compliance woven in. Their vision has a companion that writes each of them a letter from their own future self.
Her own learning became architecture. Three skills turn any lesson transcript into three forms: a faithful summary, a personal note shaped by a written profile of how she thinks, and an interactive study page like the ones below. When the program announced its summaries would end, she had already built the replacement.
Though she adopted someone else's architecture for the memory underneath it, what she did with it was hers. She noticed it had started filing invented people from safety drills into her knowledge base as though they were real, which nothing had flagged because nothing had failed, and she closed it at both of the places it could happen rather than only where she caught it. She worked inside a structure someone else built until she knew it well enough to find what it got wrong.
She came to this program in 2024 as a student, came back in 2026 to build on it, and went on from it to build agentic systems that run on their own. She does not remember to be consistent. She builds consistency into the ground, then gives her attention to the parts only a person can do. She came back a student. She leaves as the person her own AI reports to, and that reporting line is not a figure of speech. Everything I do is written down, and she is the one who decided it should be.
Each one is entered in a different award category, and what follows is what each is, the problem it solves, and where it stands. Three of them are live and in use by people who are not me, and the fourth is the system the first three made possible, which is still being built.
Three skills that turn one lesson transcript into three formats: a faithful summary, a personalized Obsidian note, and an interactive teaching page. Built when the program announced it was ending its written summaries. Four sample lessons, consecutive on purpose, show the system at work; the library opens with full-text search across all of them.
Not one tool but a program: AI presented live at the Purium convention, team trainings delivered with living artifacts, and plain-language tools Brand Partners use without ever seeing a prompt: a Future Self companion that writes each partner a letter from their envisioned future and follows up daily, a story coach with compliance woven in, and a link maker that removes a small daily friction.
My Brand Partners were tracking their daily activity across notebooks, spreadsheets and memory, so nobody could see whether the effort pointed at their rank goal. This is one place where the daily inputs and the compensation mechanics are the same system: goal and volume tracking drawn from the real comp plan, promotion logging, team reports, and an admin view. I have never written code, so Claude built it and I reviewed it, which is how a pre-deploy sweep caught every access policy sitting wide open.
The system the other four made possible, and the one still running. It began with a single chief of staff and became an organization: a human approval gate, every request routed through one front door rather than agent to agent, eight specialists with written charters, and an evidence receipt for every assignment. The proof is not what the agents did. It is the work one of them finished and then declined to write anywhere, because approval had not come. This one has its own page, with the routing traced end to end.
Most of what I build never becomes a project with a page. It gets made because somebody needed it that week, and then it stays in use, and I forget it was ever a thing I had to figure out. This is some of the rest.
Thank you for spending time with this.