The four pillars, Ore, and Concentrate: how to read your GTM Diagnostic

Run the T2D3 GTM Diagnostic and two minutes later you're holding a scored report on your go-to-market. Then the report does something most AI tools don't: it asks you to review it. Each finding carries a thumbs-up and a thumbs-down. A banner talks about promoting observations from Ore to Concentrate. A pre-review panel proposes verdicts and waits for yours.

That's not ceremony. It's the whole philosophy of T2D3 OS in one screen, and once you see why it works this way, you'll get a lot more out of every report. This piece explains the two ideas that first-time users meet on that screen: the four pillars the diagnostic grades you on, and the mining language — Ore, Concentrate — that tracks what you've judged.

Why four pillars

Websites change weekly. Channels come and go. What doesn't change is what a B2B buyer — human or, increasingly, machine — needs to believe before they buy from you. The diagnostic reads all of your raw data (traffic, rankings, backlinks, AI-engine answers, your own words) and interprets it across four durable dimensions:

Positioning. Does your website say who you're for and why you win — or does it describe your product and hope the right people self-select? We measure the gap between the positioning you state (you tell us who it's for and what it does) and the positioning your website communicates. That gap is usually the single most expensive problem in B2B SaaS marketing, because everything downstream — ads, content, sales calls — inherits it.

Growth. How big is your real demand engine? Not vanity traffic: we look at how much of your search demand is people who already know your name versus buyers with a problem who haven't met you yet. A site with 98 visits a month, 99% of them branded, isn't a marketing engine — it's a business card. Naming that honestly is the starting point.

AI Visibility. When your buyers ask ChatGPT, Claude, Gemini, or Perplexity "what's the best tool for X," do you appear in the answer? Answer engines are becoming a primary B2B distribution channel, and they cite authoritative, machine-legible sources. This pillar is new to most marketing audits and it's the one moving fastest.

Authority. Do third parties vouch for you? Backlinks, referring domains, and domain strength — not because SEO is the goal, but because the same authority signals that rank you in Google are what AI engines weigh when they decide who to recommend.

Four pillars, one interpretive sentence each. Never a bare number — a number without an interpretation is noise wearing a suit.

Ore in, edge out

Here's the uncomfortable truth about AI-generated analysis, including ours: a finding fresh out of the model is not yet truth. It's ore.

We borrowed the language of mining deliberately. Ore is genuinely valuable material — but nobody ships ore to a customer. It gets graded, smelted, and refined, and most of the rock gets discarded. In T2D3 OS, every piece of raw material moves through that refinery: what you upload and what the machine reads is Ore; when the system takes a position on it, that's an Assay; when a human confirms it, it becomes Concentrate — and only concentrate feeds the strategy work downstream. What you cross out becomes Tailings, reviewed out on purpose. (The full nine-stage journey, from Haul to Lode, is on our Humans are the loop page.)

Your diagnostic report works exactly this way. The neutral observations — the four pillar readings, the "what is working" strengths — arrive as Ore: the machine surfaced them, nobody has judged them. The recommendations — priority moves, buyer briefs — arrive as Assays: the system took a position and is waiting for your call. When you confirm a finding (or edit it to keep your own wording), it's promoted to Concentrate: confirmed signal that the rest of the platform now treats as true about your business.

Why the report asks for your judgment

Frontier AI models learned nearly everything they know without supervision — absorbing the whole internet with nobody labeling what's right for your company. That's what makes them fluent in everyone's marketing and expert in no one's. The missing ingredient has a name in machine learning: supervision. Human judgment about what's true and what good looks like, fed back into the system.

Every thumbs-up and thumbs-down you cast on a diagnostic is exactly that — a labeled example. It does two jobs at once:

1. It grounds your strategy. Downstream modules — your ICP, personas, value props, content — build on Concentrate, not on unreviewed model output. You never end up with a messaging framework built on a finding you'd have rejected in five seconds.
2. It trains your next report. Your verdicts are remembered and weighted into future diagnostics. The report you run next quarter is pre-reviewed against your past judgments — it already knows what you consider signal and what you consider noise.

This is also why the report pre-reviews itself before you arrive. When a diagnostic finishes, an AI pass reviews every finding against your locked GTM foundation and your verdict history, and proposes Keep, Demote, or — where the call is genuinely contestable — "your call," each with a confidence level and a one-line reason. You can unfold the full list, veto anything, and accept the rest in one click. The AI leads with a recommendation; the decision stays yours. Nothing is saved until you say so.

We think this is what working with AI should feel like everywhere: the machine drafts, grades, and proposes with full initiative — and the human is reserved for the two things only humans can supply: judgment and conviction.

What to do with your next report

• Don't just read it — rule on it. Ten minutes of 👍/👎 on one report compounds into every future generation the platform makes for you.
• Edit findings into your own words. An edited finding is kept as your version — the strongest form of signal you can give.
• Watch the pillars over time. Re-run the diagnostic after you ship changes; the four pillars and score history tell you whether the needle actually moved.
• Mind the AI Visibility pillar. It's the channel your competitors are most likely ignoring this year.

Ore in, edge out. The AI does the digging; your judgment does the refining — and the refined signal is what compounds.

The thinking behind this loop — signal versus noise, and how human judgment turns AI's endless output into compounding advantage — is the subject of my book Syntropy. Run your own free diagnostic at app.t2d3.pro/website-compass.