Three directions for the CPP site. Here's how we got to them — and here they are.

Research first — a word of copy comes after the evidence, never before. Nothing on these sites is invented. Every claim and every design decision traces back to a real source, so the work holds up when a skeptical operator pushes on it. Here's what went in before we designed anything:

• 1Four Big-4 / analyst reports — the state-of-AI research your buyers already read, so our framing meets theirs.
• 2A competitor scan — how the firms in your lane actually sound and look, and where the gaps are.
• 3An aspirational-brand study — seven authority benchmarks (McKinsey QuantumBlack, Stripe, Palantir, Thoughtworks and others) for voice and credibility.
• 4ICP research — four real buyers modeled in depth: the CEO/COO, CFO, CISO, and CTO.
• 5Shea's voice profile — built from his own writing, so the site sounds like CPP, not like a generic consultant.

Written in Shea's voice, drawn straight from the voice profile — operator-to-operator, candid, no hype adjectives. It holds an honest-verb discipline: the site says "designed to support" and "aligned to," never "certify" or "compliant." Every line was tested against a four-buyer executive panel (CEO, CFO, CISO, CTO), and every claim is source-tagged — if it's on the page, we can show where it came from.

The aspirational-brand and competitor research pointed to one insight: for this buyer, trust is earned through evidence, not asserted through polish. So we ran a five-category study of the brands this buyer already trusts — audit & assurance, security, financial-data instruments, developer tools, and institutional research — and pulled the shared signals that make each read as credible. The three directions are each a distinct expression of that trust register — different bones, not recolors of one layout.

None of this is decoration. Before a single pixel, the team wrote the research, the strategy, the copy, and the creative direction — and it's all here to open and read. If you weren't in the room, this is the room.

19 research, strategy, copy & design documents · ~60,000 words total · ~40,000 words of research & strategy · 5 pages of site copy · 6 perspective pieces drafted, awaiting publication.

The four Big-4 / analyst reports (McKinsey's Seizing the Agentic AI Advantage, the State of AI 2026 global survey, and two others) were ingested as source but aren't re-hosted here — they're third-party reports, not ours to publish.

We didn't guess at the register. We took the four buyer personas CPP modeled — the mid-market CEO/COO, the CFO, the CISO, and the CTO — and the same rubric a four-buyer panel already used (the one that ruled the motion hero out), and had each score all three directions out of 25 on: sourced evidence in view, built-not-styled, restraint, operator credibility, and the "expects to be challenged" tell.

For contrast, the same four buyers scored the site as it stands today at 7 / 25 — and all four closed the tab. This is what they're reacting to:

What the same buyers reject on it

• A self-quoted testimonial attributed to "The Leadership Team" — the company quoting itself. ("That's not a receipt; that's a mirror.")
• An unsourced "20+ Years" floating over the hero — a number performing credibility.
• A Crawl / Walk / Run / Fly maturity curve — a slide, not a system; the exact consultant-cosplay the panel punishes.
• A stock team-at-a-whiteboard photo and inevitability taglines ("From Vision to Value").
• Not one cited source on the page.

What ours does instead

• Every figure carries a visible source — figure · claim · SOURCE, ruled like a ledger entry.
• Honest verbs — "designed to support," "not to certify," "report pending." Never "certified."
• Named operators with checkable histories, not a wall of borrowed titles.
• No motion, no stock, no maturity curve — restraint as the signal.

Same four buyers, same rubric — this time scored against the live homepages of the firms CPP actually competes with. Each persona (the mid-market CEO/COO, the CFO, the CISO, and the CTO) rated every competitor out of 25 on sourced evidence in view, built-not-styled, restraint, operator credibility, and the "expects to be challenged" tell. The panel punishes advertised-but-unbacked rigor hardest.

A full-bleed motion/video hero is the exact vendor-deck look these three directions were built to defeat — and for this buyer it reads as weaker, not stronger. This isn't taste. We tested it.

A four-buyer panel (skeptical CEO/COO, CFO, CISO, CTO) reviewed the register and returned unanimous for the dense, sourced "instrument" look — "operators talking to operators," "someone cited where the number came from." The soft and decorative registers were rejected by all four.

These scores weren't handed down by us. They came from a panel we built to argue with our own work — and we've put its limits on the page next to its verdicts, because a scorecard you can't check is just marketing wearing a lab coat.

The panel is your buyers, not invented critics. Four expert personas, built from Covington Place's own ICP research: the transformation-ready CEO/COO, the CFO watching AI spend accumulate outside their visibility, the compliance-cornered CISO, and the CTO trying to separate real velocity from vibe-coding. These are the exact people who sign — or kill — engagements like this one. We didn't imagine an audience; we modeled the one already in the pipeline.

The rubric was fixed before any verdict. One test — sourced evidence in view, instrument-grade over styled, restraint, operator-to-operator credibility, and the tell that a site expects to be challenged rather than believed. It's the same test that ruled out a motion-video hero, scored the current live site, scored all three redesign directions, and scored every competitor. Nobody moved the goalposts between the thing being judged and the score. The test punishes advertised-but-unbacked rigor hardest — the failure mode this whole page is built to avoid.

Every score cites something you can see. Each judgment points at an artifact actually on the rendered page — a sourced-figure row, an honest verb, a self-quoted testimonial. Competitors were scored from their live homepages as they ship today, not from reputation or what we assume they're like.

It says where we lose, out loud. The same panel flagged Covington Place's own exposure as loudly as its wins: our proof is self-provenanced — our own receipts and packets — and we surface none of our ten real clients as a named logo wall, the one thing a skeptic scans for in five seconds. All four buyers named it. A scorecard that only flatters is worthless; this one told us where we're beatable.

And it corrects itself. When one competitor's homepage lazy-loaded and rendered dark on first capture, we re-captured it against a full render and re-scored it before trusting the number — no capture artifact was allowed to distort the table.

The honest limit — and it's the point. This is a modeled expert-persona panel, not a survey of live humans. It models how these specific buyers, grounded in real buyer research and a documented rubric, read a site — a decision aid, not a market survey. We say that plainly, the same way none of these sites should ever say "certified" without the paper to back it. The candor is the trust signal: if we'll tell you what this scorecard can't do, you can believe what it says it can.