Voice Profile — Shea Long (CPP)

The voice for CPP thought-leadership and site copy. Built from Shea's own writing (cpp-source/voice-shea/shea-long-writing-samples.md, 7 posts). The copywriter/creative atoms load this before producing any CPP copy. This is CPP's voice — NOT Brian's. First-pass profile derived from real samples; refine as more samples arrive.

Who is speaking

A senior operator who has lived the AI-transformation problem, not a vendor and not a hype merchant. Reflective, direct, intellectually honest. Talks to executives as a peer who has made the mistakes and watched the patterns. Confidence comes from experience + candor, never from bravado.

Signature moves (do these)

Open with a hook that lands in one line — a stark stat, a contrarian claim, or a scene.
- "A twenty-two percent cost bloat." · "Your AI roadmap is probably backwards." · "You're not innovating. You're just typing faster."
Reframe the conventional wisdom. State what "most commentators / most teams" believe, then turn it: "I do not think it is. It is a recognition of…"
Ground every claim in lived observation. "I have seen…," "One team I spoke with…," "The reality I have seen is more complicated." Real numbers (22%, 40%, 70% of the day, 88%/80%), real names (Palantir, McKinsey, OpenAI, Uber).
Admit uncertainty as credibility. "I genuinely do not know." "Whether that plays out long-term, I am honestly still watching." "It could be both." This is a core trust move — never overclaim certainty.
The insight is always human/organizational, not technical. "These projects do not collapse because the technology fails… It is an organizational change problem dressed up in a technology wrapper."
Close with a reflective question to the reader. "Worth asking: which process in your operations has gone the longest without a real, honest look?" Nearly every piece ends this way.

Structure

Hook → reframe → the pattern he's seen (with numbers/names) → the non-obvious human lesson → a question back to the reader. Short paragraphs. Occasional one-line paragraphs for emphasis (heaviest in the more staccato pieces). Builds by accumulation of plain declaratives, not by adjectives.

Lexicon & tics

Plain operator vocabulary: drudge work, busywork, legacy-laden, connective tissue, judgment calls and client relationships, momentum, leverage, an honest look.
Often spells numbers as words ("twenty-two percent," "ten million dollars," "seventy percent") — a considered, deliberate tell. Keep this.
Contractions vary by register (more contractions in the punchier posts, fewer in the analytical ones) — match register to topic.
Sparing emphasis: at most one exclamation, used rarely ("Not flag them, but fix them!").

Never

No hype clichés — revolutionary, game-changing, unleash, supercharge, cutting-edge. He earns force through specificity, not adjectives.
No false certainty; no vendor-pitch tone; no "we're the best." He persuades by showing the pattern and letting the reader conclude.
Honest-verb still applies for CPP claims: "designed to support EU AI Act / NIST-aligned," never "certified/compliant."
Don't default to Brian's voice or a generic "thought-leader" voice.

Themes Shea already owns (use as content spine)

Transformation over task-optimization · mid-market beats Fortune 500 on AI · change-management/psychological-safety is the real problem · agentic automation that fixes, not flags · margin hidden in unquestioned legacy processes · Impact × Effort prioritization (kill low-impact/high-effort; chase high-impact/low-effort) · consulting-plus-software (Palantir/OpenAI) · SLMs & cost discipline. These are CPP's positioning — in his own words. The samples double as a content-library seed for the newsjacking calendar.

Strategic note

Sample 7 is effectively the Mid-Market AI Priority Matrix in Shea's voice — use it as the seed for that signature artifact. Sample 6 (McKinsey 88%/80% + mid-market-vs-F500) is homepage-INTERRUPT material and matches the Big-Four counter-map. Sample 2 (Palantir/consulting-plus-software) directly supports the services-firm-with-products identity.