I was trying to train an AI system to understand how I work. What came out of it was a more detailed picture of how I actually think than I'd gotten from any formal assessment process. That wasn't the point, but it was the interesting part.
The first Field Note described how I onboarded Claude to an existing body of organizational context — treating it less like a tool and more like a new team member who needed to understand how I think before they could be genuinely useful. What I didn't fully describe was what happened in that process that I didn't plan for.
To transfer context effectively, I needed to make my thinking explicit. Not summarized — explicit. Not "here are my priorities" but here is how I reason through a tradeoff, here is what I've tried that didn't work, here is what quality looks like to me versus what merely adequate looks like, here is the question I run every major decision through before I commit to it.
That exercise produced something I've been thinking about since.
The methodology
The knowledge transfer used a structured four-pass analysis of a corpus of 297 professional conversations spanning roughly three and a half years — December 2022 through early 2026. The corpus covered workforce analytics, HRIS architecture, organizational design, governance modeling, business cases, executive communications, and leadership decisions I was working through in real time.
Pass 0 was a calibration pass: classify every conversation by type, signal strength, and analytical value. 51% came back high-signal. 79% fell into strategic analysis or outbound drafting — an unusually clean corpus for behavioral extraction, because most of the low-signal noise had been filtered out by how I use AI in the first place.
Pass 1 extracted voice and communication architecture from 89 outbound drafting conversations — emails, leadership memos, LinkedIn writing, executive messaging. The goal was a replication guide: not a description of how I write, but a model precise enough to produce output I wouldn't need to substantially rewrite.
Pass 2 mapped cognitive and decision frameworks. Not what I decided, but how I decided — the sequencing logic, the tradeoff patterns, the things I consistently prioritized and the things I consistently refused. Adversarial reasoning turned out to be undercounted by keyword detection; it appeared embedded throughout strategic threads, not just in the conversations explicitly framed as pressure-testing.
Pass 3 built the strategic domain map. HR Technology dominated at 72 conversations. Organizational design and governance followed at 29. Workforce analytics at 24. Financial modeling at 22. The concentration wasn't surprising — but seeing it enumerated made the shape of where I actually spend my thinking visible in a way that felt new.
Leadership profile · corpus-derived
Four passes.
One operating system.
Extracted from 297 professional conversations spanning 3.3 years. Not self-reported. Not a survey. Signals derived from observed behavior — writing, decisions, organizational design, and strategic reasoning under pressure.
01
Pass 1 · Voice & Style
Communication architecture
89 outbound drafts analyzed. Analytical, direct, structurally precise — consistent aversion to anything generic or performative.
→Contrast framing as primary rhetorical device
→Layered explanation: concept → implication → conclusion
→Evidence → insight → recommendation as default structure
→Sycophancy treated as analytical failure, not tone preference
02
Pass 2 · Cognitive frameworks
Decision architecture
Reasoning operates through systems thinking and infrastructure logic. Adversarial pressure-testing embedded throughout — not occasional.
→Infrastructure-before-intelligence sequencing in every major initiative
→Long-term system credibility over short-term wins — consistent pattern
→Compounding credibility as the organizing principle for decision ethics
→Adversarial reasoning embedded throughout — not just flagged threads
03
Pass 3 · Strategic domains
Domain concentration
79% of conversations fall into strategic analysis or outbound drafting. HR Technology and org design are the primary operating theaters.
→HR Technology / Systems: 72 conversations — deepest domain by volume
→Org Design / Governance: 29 conversations
→Workforce Analytics: 24 conversations — growing concentration
→Financial Modeling: 22 conversations — applied, not theoretical
04
Pass 4 · Leadership philosophy
Operating identity
The Orchestrator identity is the organizing frame. Six non-negotiables extracted from repeated corpus evidence. Accountability is structural before individual.
→The Orchestrator: builds systems that outlast the builder
→Climate-based culture — behavior follows conditions, not character
→Substance-before-visibility: consistent across all public communications
→Data honesty as non-negotiable — 6 distinct refusal patterns observed
Governing question — surfaces across every major decision in the corpus
"Does this make the system more trustworthy, more capable, and more connected to the mission — or does it just make it look that way?"
Statistically unvalidated — but behaviorally grounded.
Pass 4 extracted leadership philosophy. Management style. Decision ethics. Non-negotiables. The things I would do regardless of political cost, and the things I consistently refused regardless of how they were framed.
One governing question surfaced across every major decision in the corpus. Not stated once — observed repeatedly, in different contexts, over three years.
Governing question
"Does this make the system more trustworthy, more capable, and more connected to the mission — or does it just make it look that way?"
What the profile produced
The synthesis document that came out of the four passes was the most detailed and accurate leadership profile I have ever received. More specific than any 360 feedback I've been through. More behaviorally grounded than any coaching engagement. More honest — because it wasn't filtered through social desirability, hierarchy anxiety, or the natural tendency people have to soften what they observe about you when they know you'll read it.
It named things I recognized immediately: the infrastructure-before-intelligence sequencing, the compounding credibility principle, the Orchestrator identity, the six non-negotiables. But it also named things I hadn't consciously articulated — the specific rhetorical structures I use to build arguments, the precise tradeoffs I make under pressure, the places where my restraint is an active choice and the places where it might be a blind spot.
One finding I found genuinely clarifying: data honesty appeared as a non-negotiable across six distinct refusal patterns in the corpus. Not as a stated value — as an observed behavioral pattern. I had always thought of data honesty as a working principle. Seeing it documented as something I had actually enacted repeatedly, under different pressures, in different contexts, was different from believing it about myself.
The I/O psychology observation
I've spent most of my career adjacent to talent assessment — pre-employment testing, behavioral interviewing, 360 methodologies, engagement surveys. The fundamental challenge in all of it is the same: how do you get an accurate signal about how someone actually thinks and behaves, rather than how they present themselves or how observers perceive them through their own filters?
The traditional answer involves validated instruments, normed samples, structured rater training, and statistical controls for bias. All of that matters. None of it fully solves the problem.
What the four-pass analysis did differently was use behavioral evidence directly. Not self-report. Not observer ratings. The actual decisions, the actual writing, the actual reasoning patterns — extracted from the work itself over a sustained period of time.
That's not a validated assessment methodology. There's no normative sample. There's no inter-rater reliability. You can't replicate it with a 20-minute survey. I want to be honest about what it isn't.
But as a way of generating self-knowledge — of making the implicit explicit, of seeing patterns in your own behavior that you don't notice from inside the day-to-day — it produced something I haven't gotten from any formal process. The closest analog in the I/O literature might be structured behavioral interviewing, where the premise is that past behavior under specific conditions is the best predictor of future behavior. This was that, applied to a longitudinal record of real conditions rather than recalled examples.
I don't know what to do with that observation yet. But I suspect the most interesting question it opens up isn't about AI at all — it's about what it means to have a durable, longitudinal record of how you actually think, rather than a periodic snapshot of how you describe yourself thinking.
The most valuable thing the exercise produced wasn't the profile. It was the discipline of making my own thinking legible — and finding out what it looked like when someone else read it back.
