Taste applied at velocity.

Recognizing what is good, true, sharp, useful, and on-brand fast enough to keep AI-native work from becoming polished mediocrity at scale. In the agentic era, generation is cheap. The bottleneck is recognition: choosing the right brief, the right angle, the right proof, the right level of finish, and the right thing not to ship.

This is the umbrella skill. Judgment without slowness, speed without slop. I would rather pull a draft I have not sat with than ship one I cannot defend.

Defining what good looks like with enough precision that a human team, vendor, or AI agent can execute without inventing the missing strategy. Weak briefs create expensive ambiguity for humans and over-compliant slop from LLMs. Modern leaders need to specify goals, constraints, inputs, outputs, review gates, and non-goals.

Most of my work is turning messy GTM intent into schemas, intake forms, routing logic, briefs, proof standards, and review loops. The thing I write most often is the spec the AI agent or junior teammate can actually execute against.

Making useful decisions when data is incomplete, incentives are mixed, and waiting for certainty would itself be a bad decision. Executive work is rarely deterministic. AI sharpens this: outputs are probabilistic, source quality varies, and the leader has to decide when evidence is sufficient.

I build decision layers before making claims. Attribution hygiene, proof governance, pipeline inspection, source maps, executive reporting that separates signal from noise. Dashboards are inputs, never the final read.

Making ambiguity legible so an organization can act: what is happening, why it matters, what to believe, what not to claim, and what to do next. Executive work is mostly creating the shared mental model that lets teams coordinate in uncertainty. Communicating decisions is the surface.

Narrative is infrastructure for me. The bridge between writing and systems building. Buyer economics, proof, market context, channels, and adoption paths all have to connect inside one frame.

Holding a strong point of view while staying willing to update it when the evidence, source quality, or operating reality changes. AI-native executives need conviction to move and humility to avoid hallucinated certainty, vanity metrics, or overconfident strategy.

I put a stake in the ground while building claim governance, evidence tiers, public-safe language, and review gates around the claim. The thesis can be loud. The proof has to keep pace.

Thinking in finance, sales, product, compliance, engineering, and RevOps frames rather than translating marketing language at the surface. Modern GTM leaders have to reason across functions. Translation is not enough. The executive has to understand what each function optimizes for and what proof they trust.

Cross-functional communication is the surface. The work is building artifacts that finance, sales, executives, product, compliance, and technical teams can each use without translation.

Knowing when to trust an AI output, dashboard, vendor, team member, or source. When to sample, when to verify, and when to throw it away. Over-trusting destroys credibility. Under-trusting destroys the productivity gain. The scarce skill is calibration.

I design guardrails that preserve the productivity gain. Source-backed libraries, human review gates, audit trails, approval statuses, QA loops, pipeline inspection. The review loop is the unlock, not the brake.

Treating AI agents as operating participants that need context, goals, constraints, feedback, memory, review cycles, and ownership boundaries. Leaders who treat agents as chatboxes get tool demos. Leaders who manage agents like junior team members get a repeatable operating system.

I have firsthand muscle here. Agent rosters, handoffs, task-specific workflows, source libraries, QA gates, review cycles. Each agent has a role, an owner, and a defined output, the same way I would onboard a teammate.

Adding enough structure to make work safe, repeatable, and reviewable without turning the organization into a compliance swamp. AI-native GTM will fail if governance is either absent or paralyzing. The executive skill is designing lightweight control systems.

Strong regulated-market differentiator for me. Claim tiers, proof libraries, RFP gates, compliance-aware content systems, executive review loops. Light enough to keep velocity, structured enough that nobody ships an unsupported claim.

Finding the hidden signal in messy GTM data while questioning whether the system is measuring the right thing at all. AI and dashboards can amplify bad measurement. The better leader asks whether the signal path is broken before making people optimize the wrong metric.

My proof library is unusually strong here. Attribution false negatives, ghost pipeline, signal-to-touch SLAs, lead routing, lifecycle definitions. A lot of what looks like strategy failure is measurement failure wearing a strategy costume.

Designing systems people will actually use: clear ownership, low-friction handoffs, feedback loops, and enough training to make the new behavior stick. The best GTM architecture fails if sales, marketing, proposal, finance, or executives do not adopt it. AI workflows make adoption more important because the process is new and psychologically unfamiliar.

I turn systems into usable operating habits. Playbooks, cadence, team lanes, sales handoffs, dashboards, intake, QA. The system that nobody uses is a worse outcome than the messy process it replaced.