Executive operating profile and writing on technology operating models in the AI era, by Matt Fairchild.https://fairchild.worken-usWed, 29 Jul 2026 00:00:00 GMThttps://fairchild.work/writing/the-constraint-ai-was-supposed-to-removehttps://fairchild.work/writing/the-constraint-ai-was-supposed-to-removeSynthetic personas were sold as an escape from the research budget. Operating one taught me the opposite lesson: every layer of the system consumes human research.Wed, 29 Jul 2026 00:00:00 GMTapplied-aigovernancehttps://fairchild.work/writing/auditable-then-articulatehttps://fairchild.work/writing/auditable-then-articulateTrustworthy AI judgment must be auditable before it is articulate. Explanation comes free with the model; an evidence trail exists only if someone built it.Tue, 28 Jul 2026 00:00:00 GMTapplied-aigovernancehttps://fairchild.work/writing/the-honest-defense-of-an-ai-judgment-systemhttps://fairchild.work/writing/the-honest-defense-of-an-ai-judgment-systemThe strongest defense of an AI judgment system is agreement with the evidence against it: concede the findings, show the design answers, and claim only the scope the system can defend.Mon, 27 Jul 2026 00:00:00 GMTapplied-aigovernancehttps://fairchild.work/writing/when-optimism-replaces-leadershiphttps://fairchild.work/writing/when-optimism-replaces-leadershipA uniformly positive leadership register carries no information. Confidence comes from coherence between what leadership says and what the organization experiences.Fri, 24 Jul 2026 00:00:00 GMToperating-modelsleadershiphttps://fairchild.work/writing/failures-that-look-like-successhttps://fairchild.work/writing/failures-that-look-like-successThe failures that mattered in this build produced smooth prose and clean exit codes. Why a trustworthy agent system verifies its most convincing signals hardest.Thu, 23 Jul 2026 00:00:00 GMTapplied-aioperational-memorygovernancehttps://fairchild.work/writing/hard-decisions-are-where-a-strategy-becomes-realhttps://fairchild.work/writing/hard-decisions-are-where-a-strategy-becomes-realThe organization learns the actual strategy from hard decisions, and closing the gap between an event and its explanation is an operating function of the executive team.Thu, 23 Jul 2026 00:00:00 GMToperating-modelsstrategy-executionleadershiphttps://fairchild.work/writing/executive-attention-is-an-operating-model-problemhttps://fairchild.work/writing/executive-attention-is-an-operating-model-problemExecutive attention is scarcer than capital or headcount, and its allocation should be designed into the operating model the way capital allocation is, rather than left to habit.Mon, 20 Jul 2026 00:00:00 GMToperating-modelsstrategy-executiongovernancehttps://fairchild.work/writing/writing-to-production-with-no-undohttps://fairchild.work/writing/writing-to-production-with-no-undoA store with no undo makes every write permanent, so governance must sit ahead of the write. Explicit identity, verification in code, and a human at the irreversible step.Thu, 16 Jul 2026 00:00:00 GMTapplied-aioperational-memorygovernancehttps://fairchild.work/writing/trust-as-machineryhttps://fairchild.work/writing/trust-as-machineryTraceability cannot live in a prompt. The ledger machinery that carries provenance, resolves conflicts by recorded rules, and delivers citations the end user can check.Thu, 09 Jul 2026 00:00:00 GMTapplied-aioperational-memorygovernancehttps://fairchild.work/writing/the-orchestrator-that-never-reads-a-documenthttps://fairchild.work/writing/the-orchestrator-that-never-reads-a-documentThe binding constraint of an agent system is context rather than compute. How a deterministic orchestrator that never reads a document ran 244 sub-agents over a corpus no window can hold.Thu, 02 Jul 2026 00:00:00 GMTapplied-aioperational-memorygovernancehttps://fairchild.work/writing/synthesis-over-storagehttps://fairchild.work/writing/synthesis-over-storageAn agent is only as good as the knowledge engineering behind it. Why we distilled a 580 MB document pile into 41 governed, traceable knowledge articles instead of refreshing the store.Thu, 25 Jun 2026 00:00:00 GMTapplied-aioperational-memorygovernancehttps://fairchild.work/writing/desk-control-center-the-full-descriptionhttps://fairchild.work/writing/desk-control-center-the-full-descriptionThe full description of Desk Control Center itself: every widget on the touchscreen strip under my monitors, what each one does, and what it pulls from live.Mon, 15 Jun 2026 00:00:00 GMTapplied-aicoding-agentspersonal-softwarehttps://fairchild.work/writing/a-control-center-for-an-audience-of-onehttps://fairchild.work/writing/a-control-center-for-an-audience-of-oneThis is the audience of one up close: the control center I built for one desk, the widgets it ended up with, and the hardware that fought me the whole way.Thu, 11 Jun 2026 00:00:00 GMTapplied-aicoding-agentspersonal-softwarehttps://fairchild.work/writing/software-worth-building-for-one-personhttps://fairchild.work/writing/software-worth-building-for-one-personBuild cost used to set a floor under what software was worth making. A capable coding agent drops that floor low enough to make building for an audience of one rational.Thu, 04 Jun 2026 00:00:00 GMTapplied-aicoding-agentspersonal-softwarehttps://fairchild.work/writing/strategic-execution-operating-modelhttps://fairchild.work/writing/strategic-execution-operating-modelA complete system for closing the gap between strategy and what an executive team delivers: five layers tied by a traceability chain, with the monthly outcome commitment as its atomic unit.Thu, 28 May 2026 00:00:00 GMToperating-modelsstrategy-executiongovernancehttps://fairchild.work/writing/talk-to-the-person-firsthttps://fairchild.work/writing/talk-to-the-person-firstMost workplace conflict becomes unmanageable through delay and triangulation. Two disciplines keep it small: talk to the person first, and write the follow-up before the conversation.Thu, 28 May 2026 00:00:00 GMToperating-modelsleadershiphttps://fairchild.work/writing/installing-and-sustaining-the-operating-modelhttps://fairchild.work/writing/installing-and-sustaining-the-operating-modelInstalling an operating model is the easy part. Keeping it light enough that leaders operate it rather than route around it, re-earned each cycle, is what makes it last.Thu, 21 May 2026 00:00:00 GMToperating-modelsstrategy-executiongovernancehttps://fairchild.work/writing/measuring-execution-healthhttps://fairchild.work/writing/measuring-execution-healthCommitment accuracy alone is easy to misread. Execution health takes a small set of honest signals, read together on the sprint cadence, to show whether follow-through is holding or eroding.Thu, 14 May 2026 00:00:00 GMToperating-modelsstrategy-executiongovernancehttps://fairchild.work/writing/the-executive-sprinthttps://fairchild.work/writing/the-executive-sprintA list of goals and priorities gives an operating model nothing to run. The unit it needs is the monthly outcome commitment: a defined result one leader owns and finishes by a date.Thu, 07 May 2026 00:00:00 GMToperating-modelsstrategy-executiongovernancehttps://fairchild.work/writing/nested-planning-horizonshttps://fairchild.work/writing/nested-planning-horizonsExecutive teams already run every planning horizon they need. What is missing is the connection between them: nested horizons linking three-year strategy to the monthly sprint via a quarterly rebase.Thu, 30 Apr 2026 00:00:00 GMToperating-modelsstrategy-executiongovernancehttps://fairchild.work/writing/what-a-governed-chain-changes-for-an-organizationhttps://fairchild.work/writing/what-a-governed-chain-changes-for-an-organizationWhat building software under a governed chain changes for an organization: governance becomes a property the system carries, and AI velocity becomes something a leader can stand behind.Sun, 26 Apr 2026 00:00:00 GMToperating-modelsgovernanceapplied-aidelivery-governancehttps://fairchild.work/writing/strategy-fails-where-it-becomes-workhttps://fairchild.work/writing/strategy-fails-where-it-becomes-workMost strategy is sound enough to succeed. It leaks in the untracked distance between the decision and the work, because executive teams have no operating model for their own execution.Thu, 23 Apr 2026 00:00:00 GMToperating-modelsstrategy-executiongovernancehttps://fairchild.work/writing/building-with-ai-without-vibe-codinghttps://fairchild.work/writing/building-with-ai-without-vibe-codingBuilding software with AI without vibe coding: a coding model does the writing, and deterministic code around it decides whether that work is allowed to proceed.Sun, 19 Apr 2026 00:00:00 GMTapplied-aicoding-agentsgovernancedelivery-governancehttps://fairchild.work/writing/people-remain-the-differentiationhttps://fairchild.work/writing/people-remain-the-differentiationModel capability is becoming something any organization can buy. Durable advantage comes from the judgment held by people and the governed memory that preserves it.Thu, 16 Apr 2026 00:00:00 GMToperating-modelsapplied-aioperational-memorygovernancehttps://fairchild.work/writing/from-business-question-to-approved-requirementhttps://fairchild.work/writing/from-business-question-to-approved-requirementRequirements quality has always depended on who was in the room. Capturing expert judgment as governed, reusable assets makes it something an organization owns instead.Sun, 12 Apr 2026 00:00:00 GMTdelivery-governancegovernanceoperating-modelsapplied-aihttps://fairchild.work/writing/designing-the-boundary-between-deterministic-and-semantic-systemshttps://fairchild.work/writing/designing-the-boundary-between-deterministic-and-semantic-systemsKeeping deterministic systems and adding AI where it helps settles nothing on its own. The real design work is deciding where the boundary sits and engineering the seam between them.Thu, 09 Apr 2026 00:00:00 GMTapplied-aioperating-modelsgovernancehttps://fairchild.work/writing/from-requirement-to-running-infrastructurehttps://fairchild.work/writing/from-requirement-to-running-infrastructureFollowing one requirement from an approved requirement to running infrastructure, where deterministic contracts carry the intent a coding model cannot preserve on its own.Sun, 05 Apr 2026 00:00:00 GMTdelivery-governanceapplied-aicoding-agentsgovernancehttps://fairchild.work/writing/operational-memory-as-enterprise-ai-infrastructurehttps://fairchild.work/writing/operational-memory-as-enterprise-ai-infrastructureModel capability is becoming a commodity. Durable advantage comes from operational memory: the durable, governed capture of decisions and signals that both people and AI systems can reuse.Thu, 02 Apr 2026 00:00:00 GMToperational-memoryapplied-aigovernanceoperating-modelshttps://fairchild.work/writing/from-fragmented-signals-to-earlier-interventionhttps://fairchild.work/writing/from-fragmented-signals-to-earlier-interventionEnterprises drown in scattered execution signals and find out too late. Capturing decisions and concerns in a durable, governed form moves intervention earlier and builds institutional memory.Thu, 26 Mar 2026 00:00:00 GMToperating-modelsapplied-aioperational-memorygovernancehttps://fairchild.work/writing/ai-changes-the-tooling-not-the-operating-modelhttps://fairchild.work/writing/ai-changes-the-tooling-not-the-operating-modelEnterprises are adopting capable AI tools but seeing uneven results. Enterprise value comes from the operating model around the work: decision rights, workflows, governance, and durable context.Thu, 19 Mar 2026 00:00:00 GMToperating-modelsapplied-aigovernanceoperational-memoryhttps://fairchild.work/writing/captioning-a-product-catalog-with-a-local-vision-modelhttps://fairchild.work/writing/captioning-a-product-catalog-with-a-local-vision-modelRunning an open vision model over 23,000 product images on local hardware. The model was the easy part; the deterministic envelope around it did the real work.Thu, 12 Mar 2026 00:00:00 GMTapplied-aioperating-modelsgovernance