AI changed how CPG founders access information. It didn't solve the harder problem: decision confidence.

By Gustavo Nader, founder of Aterra Studio

General AI can guess, blend, and hallucinate a plausible-sounding paragraph as easily as a well-grounded one — and sound equally confident either way. That's genuinely useful for early exploration. It is not enough once real capital, time, and risk are on the line. Ask a friend or an advisor instead, and you'll get a real, well-meant opinion shaped by whatever they happen to have seen. Neither one is a diagnostic.

Gravity was built to be the second thing. Here's what's actually underneath it.

The architecture: Aterra Intelligence OS → the Gravity System

Gravity isn't a single AI call wrapped in a nice interface. It runs on the Aterra Intelligence OS — a signal-collection and analysis layer that pulls consumer language, competitive shelf data, ingredient and claim intelligence, and category movement into a structured signal library. The Gravity System is the diagnostic framework built on top of that base, and it's what actually produces a brief.

Three ways to use it, depending on what you need:

Reading for pattern, not noise

Trends are loud. Patterns are quiet — and most of the real opportunity lives inside the pattern, not the trend. That isn't just how Gravity talks about a market. It's built into how the system reads one.

Every piece of evidence that feeds a brief gets weighed for what it actually is: a single spike of chatter, or something that shows up again and again across reviews, reorders, and real repeat purchase. A claim that's suddenly everywhere in launch decks this quarter is a trend. A claim that's still surfacing in five-star reviews eighteen months after launch is a pattern. Gravity is built to tell you which one you're looking at — because the founders who build for the pattern are usually a step ahead of the ones still chasing the trend.

The eight dimensions

That distinction plays out across eight dimensions, each one built because the same handful of assumptions keep hiding in plain sight in CPG: that people will understand the claim, that the ingredient will carry the story, that the format fits the occasion, that the price feels justified, that a buyer will see the opportunity, that the consumer will repeat, that the market will educate itself. Most of the time, it won't. The market is honest, but it isn't gentle. Every Gravity Scan Brief exists to check one of those assumptions before it's expensive to be wrong about:

1. Whitespace Validation — is there a real, unowned space for this concept, or is the "gap" already filled?
2. Consumer Language & Signals — does the way you talk about the product match the way real buyers are already searching, reviewing, and describing the need?
3. Format Fit — does the format match a purchase occasion Americans actually have, or one nobody has managed to make survive?
4. Claims & Language — is the claim current, or three years stale in this market even if it's category-defining at home?
5. Ingredient Logic — does the formulation story hold up, including anything that reads as "proprietary" but hasn't been verified as such?
6. Competitive Positioning — who are you actually pricing against on a real shelf, not just who surfaces in a Google search?
7. Business-Model Risk — do the unit economics survive real retail math, distributor, broker, and trade spend included?
8. Founder Readiness — a directional read on execution capacity. It isn't scored into the overall number; it's context, not a grade.

Each dimension carries its own confidence level (High / Medium / Low) and evidence grade (A / B / C), so you always know whether a finding rests on strong signal or on a reasoned, disclosed inference. Nothing is presented with more certainty than the evidence supports.

The discipline that keeps it honest

The easiest way for an AI system to sound impressive is to invent things with confidence — a specific competitor weakness, a specific dollar threshold, a specific patent claim. Gravity is built explicitly against that failure mode:

• No fabricated brands or products. If a name or claim can't be verified, the system uses category and role language instead of guessing at specifics.
• No invented numbers posing as gates. Gravity won't tell you to "test with 50 buyers" or "hit 4 units/door/week" unless that number is sourced from your own intake or a disclosed framework benchmark. The condition is the gate — not a figure that merely sounds authoritative.
• Category Benchmark discipline whenever live competitive data isn't available, rather than fabricated direct-competitor detail — with an explicit pointer to the Reviewed Gravity Brief for verified mapping.
• Patent and proprietary claims are never asserted as fact — they're flagged for verification.

Every brief lands on one of five clear verdicts — build, refine, reposition, validate first, or pause — because a founder making a real capital decision deserves a clear read, not a hedge.

And before any of this reached a live founder, the system was calibrated against known launch archetypes: a genuine in-market winner, a genuine me-too concept, a genuine out-of-coverage category, a strong-but-unproven concept. The point was to confirm the scoring actually separates a real opportunity from a crowded one, instead of flattering everything that comes through it.

Readings, not decisions

None of this exists to hand a founder a verdict from a black box. Gravity gives readings, recommendations, and evidence-grounded guidance — the founder makes the actual call, the same way I made every real call building Amaz. The discipline above isn't there to make Gravity sound authoritative. It's there so that when the system does say something, it has earned the right to.

The part the system can't do: judgment

Here's the honest limit of any diagnostic, this one included. It can read signals with rigor. It can't sit across the table from a buyer, feel a room go quiet at Expo West, or know from experience which "promising" concept has the founder behind it to actually execute. That's not a data problem. It's judgment — and judgment comes from having done it.

Which is why the Reviewed Gravity Brief isn't the Gravity Scan with a nicer cover page. I review and sign every one personally — bringing the same operating experience that took Amaz into Erewhon, Whole Foods, and Sprouts across 400+ doors to bear on your specific findings — and then we spend an hour together going through it, live, so the sharpest interpretation and the most useful next step actually land.

A founder-reviewed read isn't a nice-to-have bolted onto an AI report. It's the difference between a system that's right and a system whose "right" has been tested in the field by someone with real capital and real shelf space on the line.

Start where it makes sense for you

If you're still shaping the idea, start with a free question on Ask Gravity or run the $500 Gravity Scan. If you're heading into a product, launch, retail, or investor decision — or running independent diligence on a concept — the Reviewed Gravity Brief is built for exactly that moment.

Spend follows validation. Everything above exists to make that one line true in practice, not just as a slogan.