INSIGHT I — Presence is not preference  ↗

Optimizing for culture optimizes for AI.

[THE METRIC]Cultural Velocity® v3.1
[EVIDENCE TO DATE]1,182,515 data points · 903 brands
[OUTPUT]A comprehensive diagnostic
[OVERVIEW]Cultural Velocity reads the signals that shape whether your brand is recommended — by people, and by the AI systems trained on them.
[THE ADVANTAGE]

In culture, it's not enough to be seen — you have to be relevant. Cultural Velocity® shows you how.

AI is not an algorithm. It synthesises human decision-making. It infers meaning, relevance and recommendation from the same cultural conditions that shape human preference. Cultural Velocity measures those conditions, outlines where a brand over or underperforms, and sets out a blueprint for optimisation.

[THE DISTINCTION]

Cultural Velocity is the difference between being seen, and being recommended.

[VISIBILITY TOOLS]

Don't measure what got cited yesterday.

GEO monitors track visibility and citation — telling you sources used and results given from AI tools. Yet research proves that  coverage-volume and the citation-count metrics are not accurate predictors of recommendation.

[CULTURAL VELOCITY]

Measure why you'll be recommended tomorrow.

Cultural Velocity® measures the recommendation engine itself — the cultural levers that are the DNA of decision-making. More than visibility, we dive into the 55 cultural signals proven to steer decision-making in both humans and AI.

Cultural Velocity® is the first tool to measure a brand's cultural impact — and how that influences recommendability.

[THE ENGINE]

IMMA. Four pillars. Fifty-five signals.

01ImprintCan the world identify and define you consistently?
02MagnetismIs your meaning distinctive, cohesive and preferred?
03MomentumAre you being reinforced recently, in concentrated, aligned ways?
04AccelerationAre you expanding into new contexts without losing coherence?
[THE PRINCIPLE]

A clear, corroborated identity across every surface that talks about you. Brands with strong Imprint aren't just present — they're understood the same way by everyone who encounters them.

[THE CULTURAL LENS]

People recommend brands they can explain. Clarity precedes preference. When a brand means something different on every shelf, it bleeds credibility.

[THE AI LENS]

LLMs build identity from thousands of sources and penalise contradiction. Entity definition and knowledge-graph legibility determine whether AI resolves you into one coherent entity — or loses confidence.

Every pillar is computed from publicly observable signals across four channels — Earned/PR, Social/UGC, Owned/Listings, Paid/Advertising. 55 verified signals, calibrated every quarter against real consumer prompts across 15 categories in four markets.

[THE INTELLIGENCE] Calibrated (weights v3, cycle 1)
directional · correlational

Six insights Cultural Velocity tells us:

In the July 2026 study, 179,970 recommendation results were correlated with Cultural Velocity scores and read against fifty years of behavioural science, to better understand the DNA of recommendation. The results were interesting…

Close-up of two hands, fingertips almost touching, against a blurred violet background
INSIGHT I

Presence is not preference.

The visibility theory predicts the brands with the most coverage get recommended most. They don't. Across the study's 232 brands, the sheer volume of a brand's coverage shows almost no relationship with AI recommendation. What matters instead is timing — how fresh the news is, and whether the brand is entering rooms it wasn't in before. Those are the variables that move the needle. AI, like us, isn't looking for volume — it's looking for a vibe check.

THE SCIENCE — The availability heuristic at corpus scale (Tversky & Kahneman, 1973). In Ahrefs' citation analysis, web mentions were roughly 3× more predictive of AI visibility than backlinks; 84% of 25 million AI-cited links in Muck Rack's study were earned media.

A person in a dimly lit bar carries a silver boombox on one shoulder
INSIGHT II

Culture moves in moments, not drumbeats.

This is a headline that inverts a decade of SEO best practice. An even drumbeat of coverage correlates negatively with AI recommendation — −0.32 pooled across four markets. Fewer, denser, braver moments that cut through culture are where memorability is made — the silence between them just amplifies their clarity.

THE SCIENCE — Distinct, arousing events are encoded and shared onward; background hum is filtered (Berger & Milkman, 2012). Visible collective attention compounds into cumulative advantage (Salganik, Dodds & Watts, 2006).

Seen from above, a woman in a leather jacket and cap works on a laptop by a city railing as people pass
INSIGHT III

Comparison is always the context.

What a brand says it stands for on its own channels carried almost no influence in recommendation. But when framed as contrast evidence — winning comparisons, substantiated sustainability records, crisply defined differentiation — that's where these statements came alive. In the US, comparison leads all recommendation drivers. In AI recommendation, as with humans, ‘the same but better’ loses out to ‘different’ every time.

THE SCIENCE — Human evaluation is comparative, not absolute — attributes inert in isolation become decisive in joint evaluation (Hsee, 1996). Mental availability belongs to the brand most strongly linked to the category's cues (Romaniuk, 2018; Sharp, 2010).

A smiling woman holds two coupe glasses of champagne over her eyes against a red curtain
INSIGHT IV

Now and Then.

AI has two memories, the current and the remembered. The canon — slow, deep, laid down over years — and the culture — fast, current. These memories hold different sway in different categories; for beer, heritage is a win, but in skincare there's an obsession with what's next. But in every category, you need both to be culturally influential and recommended.

THE SCIENCE — Mere exposure breeds preference (Zajonc, 1968). Repetition breeds believed truth (Hasher, Goldstein & Toppino, 1977). Fluency is read as merit (Alter & Oppenheimer, 2009).

A young man in a camel coat stands in front of a market stall stacked with fruit
INSIGHT V

Like us, machines have a local accent.

While the climate of culturally-inspired recommendation is universal, the weather is local. Every market we covered in our initial research (Manifest Group's home markets of the UK, USA, Australia and Sweden) had a different DNA of cultural relevance. And AI recognises that, behaving very differently in each market, although some categories differ more than others.

THE SCIENCE — Cultural signals explain a similar share of within-category recommendation variance in every market (R² 0.27–0.35) — a substantial, replicated effect for an outcome this noisy.

Two friends laughing together in a white corridor
INSIGHT VI

Technology that makes brands more human.

Step back and the theories converge on one: the machines have inherited our instincts. Freshness dominates because availability does. Bursts win because salience does. Meaning works as contrast because judgement is relative. These machines are not shaped by an algorithm, but by human decision-making, and the meaning of that is that creativity wins.

THE SCIENCE — Canonical cognitive-psychology experiments run on language models find human-like biases in the machinery (Binz & Schulz, 2023, PNAS; Lampinen et al., 2024). This programme extends that literature from lab vignettes to brand choice at scale.

[METHODOLOGY]

The signal score weighting is calibrated by category against real LLM recommendations every quarter, keeping pace with culture as it shifts.

1 — COLLECT

Recommendation behaviour collected via the Anthropic, OpenAI and Gemini APIs; raw signals conditioned by recency decay, independence and source quality.

2 — NORMALISE

Each signal mapped 0–100 against the benchmark cohort for category, geo and window.

3 — AGGREGATE

Pillars aggregate their signals using calibrated weights.

4 — DIAGNOSE

CV score = weighted synthesis of four pillars, adjusted by a risk/noise penalty.

5 — OPTIMISE

Channel influence and prioritised action areas — where the score is built, where it leaks, and the remedies that move it.

1,182,515data points collected to 3 October 2026, across two quarterly calibrations and 86 report datasets. Made up of:
898,723AI recommendation results: one brand, checked in one AI answer
101,132signal measurements across the 55 IMMA signals and three under test
96,297articles and pages logged as evidence
86,363AI answers logged from ChatGPT, Claude and Gemini

Counts include only collected records, not scores derived from them, and exclude failed, scrapped and test runs. In the July 2026 study, IMMA scores explained roughly a quarter to a third of within-category recommendation variance per market. Claims stay correlational, never causal. Calibration is proprietary.

[THE REPORT]

Cultural Velocity is a diagnostic tool that not only measures a brand's cultural impact, but reveals how to maximise it.

[01] IMMA profile — the shape of the score.
[02] Cohort league table — illustrative brands.
[03] Pillar deep-dive — where the score is built.
[04] Optimise — prioritised action areas.

A Cultural Velocity® Blueprint provides a complete benchmark diagnostic, and Tracking adds a live dashboard with custom reporting.

Screenshots are deliberately simplified examples of areas of analysis.
[FORMATS]

Four ways to measure.

[FREE]

Preview

A cohort-led read of your category. Where you rank, and the one thing the diagnostic sees first.

↗ GET YOURS NOW
[BENCHMARK REPORT]

Blueprint

A brand-led diagnostic. Your CV score, pillar shape, channel influence and prioritised action areas.

↗ ENQUIRE
[FULL DIAGNOSTIC]

Playbook

The full read — signal-level diagnostics, prompt analysis, competitive ponds, and a strategy to move the score.

↗ ENQUIRE
[BRAND TRACKING]

Tracking

Ongoing brand tracking with Cultural Velocity: a live dashboard against your cohort, and custom reporting.

↗ ENQUIRE
[WHITE PAPER] The Recommendation Machine What 179,970 recommendation results taught us about AI recommendation
[QUESTIONS]

What people ask about Cultural Velocity.

What is Cultural Velocity?

Cultural Velocity® is a brand diagnostic from Manifest. It measures the cultural conditions that predict whether a brand gets recommended, by people and by the AI systems trained on them. Each brand is scored 0 to 100 against its category cohort, across four pillars (Imprint, Magnetism, Momentum and Acceleration) built from 55 verified signals and calibrated per category against real recommendations from ChatGPT, Claude and Gemini.

How do AI assistants decide which brands to recommend?

Much the way people do. In Manifest's July 2026 study of 179,970 recommendation results, recommendation moved with fresh coverage, concentrated bursts of attention, first appearances in new outlets and winning comparisons. Sheer coverage volume barely registered. Language models are trained on the written record of human judgement, so they inherit its heuristics: availability, salience, comparison and social proof.

How is Cultural Velocity different from GEO and AI visibility tools?

GEO and AI visibility tools track citations and coverage: which sources AI tools use and which answers they give. Cultural Velocity measures what predicts recommendation, and where a brand should strengthen it. In Manifest's research, coverage volume and citation counts were not accurate predictors of recommendation, which is why Cultural Velocity is recommendation-first rather than visibility-first.

Does more coverage make AI recommend your brand?

Not on its own. In the UK record, Mercedes-Benz had roughly twice Toyota's coverage, yet it was recommended in 7% of UK vehicle trials against Toyota's 67%. Across the panel, coverage volume showed almost no relationship with AI recommendation, while freshness, bursts of attention and entry into new outlets showed strong ones (The Recommendation Machine, Manifest, July 2026).

Do schema markup and llms.txt get a brand recommended by AI?

They make a brand legible, and legibility matters, but on their own they don't make it recommended. In Manifest's research, owned-page optimisation alone did not move the pillars that convert to recommendation. AI assistants mostly draw on what others say about a brand: 84% of the 25 million links they cited were earned media and 0.3% paid (Muck Rack, May 2026).

What are the four IMMA pillars?

Imprint asks whether the world can identify and define you consistently. Magnetism asks whether your meaning is distinctive, cohesive and preferred. Momentum asks whether you are being reinforced recently, in concentrated, aligned ways. Acceleration asks whether you are expanding into new contexts without losing coherence. Together they make up a brand's Cultural Velocity score.

How is Cultural Velocity calibrated?

Every quarter, Manifest puts real consumer questions to ChatGPT, Claude and Gemini through their APIs, never naming a brand, and records which brands they recommend. The October 2026 calibration covered 15 categories and 395 brands in the UK, the US, Australia and Sweden. Including the report datasets, Cultural Velocity has collected 1,182,515 data points so far. Weights are fitted per category and kept proprietary.

How much research is Cultural Velocity built on?

1,182,515 data points to 3 October 2026: 898,723 AI recommendation results (one brand checked in one AI answer), 101,132 signal measurements, 96,297 articles and pages logged, and 86,363 AI answers from ChatGPT, Claude and Gemini. They cover 903 brands across two quarterly calibrations and 86 report datasets.

Does Cultural Velocity prove that culture causes AI recommendation?

No, and it doesn't claim to. The findings are correlational. Cultural signals explain roughly a quarter to a third of within-category recommendation variance in each market in the July 2026 study (R² 0.27 to 0.35), a substantial and replicated effect for an outcome this noisy. Every score is relative to the brand's own category cohort and is never compared across categories.

Which markets and categories does Cultural Velocity cover?

The October 2026 calibration covers the UK, the US, Australia and Sweden across 15 categories: CRM software, activewear, soft drinks, skincare, luxury, passenger vehicles, consumer technology, fast food and QSR, household products, beer and lager, high street fashion, spirits, energy providers, mobile networks and broadband providers. Brand audits have also covered New Zealand, Canada, Norway, Finland and Denmark.

How do I get a Cultural Velocity score for my brand?

Request a free Preview report: a cohort-led read of your category showing where you rank and the one thing the diagnostic sees first. Three formats go further: the Blueprint, a full brand diagnostic; the Playbook, which adds signal-level analysis and a strategy to move your score; and Tracking, ongoing tracking with a live dashboard and custom reporting. To commission one, email culturalvelocity@manifest.group.

Who makes Cultural Velocity?

Manifest, the global brand communications agency, with studios in London, Manchester, Stockholm, New York, Los Angeles and Melbourne. Cultural Velocity launched on 30 July 2026, alongside the white paper The Recommendation Machine.

See your brand's Cultural Velocity®.

No commitment necessary. Complete the form and receive your Cultural Velocity score. ↗ Or contact Manifest to commission a full report
CULTURAL VELOCITY®
BY MANIFEST
©2026 ↗ ANOTHER INNOVATION FROM MANIFEST.GROUP Cultural Velocity