You built the brand. AI tells its own version.

We measure how AI assistants describe, rank, and recommend your brand, then test which message moves it.

Request a brand audit See a real finding

01 / What we found

AI called Insta360 a category leader. Their own materials never claim it.

We ran this study on our own initiative, to develop and pressure-test the method. Thirteen of twenty-five answers described Insta360 as a leader or innovator in its category. The official public materials we scoped do not substantiate a market-share or leadership claim, so a brand team repeating it back would be building on something AI invented.

Measured11 Jul 2026n=25 responsesDeepSeek · ungrounded

What AI associates with the brand, and whether the brand can back it

Five fixed description prompts, five repetitions each. Every association below was read and accepted by a human reviewer against a literal span in a stored answer.

Insta360 association recurrence and gap classification
What AI associated	Responses	Finding
Capture everything, frame later	20/25	Reinforced
Smooth, action-ready capture	20/25	Reinforced
360-category leader and innovator	13/25	Unsupported
Workflow and technical trade-offs	8/25	Misframed
Direct-to-share flat video	0/25	Missing

The last row is the one a brand team can act on. Insta360’s official materials now lead on direct-to-share output that preserves the 360 footage. Across all twenty-five answers, the coded evidence contained no distinct association for it at all.

Limits, stated plainly. One engine, one mode, twenty-five stored answers, coded by a single analyst. No confidence interval, inter-rater reliability, or stability claim is made or implied. Insta360 was not a client, did not commission this, and did not endorse it. Read the full study.

02 / Why now

The click no longer tells the whole story.

Paid and organic analytics explain what happens after an impression or visit. AI can compare, describe, and recommend brands before a buyer reaches a site, sometimes without sending a click at all.

Traditional attribution

Where the old scorecard starts too late.

These metrics remain useful. They begin after AI may have already framed, compared, or omitted the brand.

03 / What you get

The old A/B test finds a winner. The new one finds the next message.

An audit measures how AI describes your brand today. Message Lift takes that measured answer as the baseline, then tests one Current message against one New message.

The evidence layer

Measured

Message Lift

Simulated

Current message

What you say today.

Shared contexts.
Only the message changes.

New message

The candidate you are considering.

Two tests run against the pair. Buyer response scores simulated reactions on a 1 to 5 construct. AI recommendation supplies each message as context in brand-neutral situations and reports shortlist rate. Both are comparative simulations, never a prediction of sales or AI ranking.

Want to see the whole thing? Read the full Insta360 study, including every prompt we used and every limit that applies.

04 / Method and limitations

Evidence you can inspect.

Every result shows when it was produced, which model route was tested, how many answers were reviewed, and where the conclusion stops.

Repeated sampling. Audit prompts run multiple times because model answers vary. Rates describe that distribution.
Measured versus simulated. Visibility evidence comes from stored model answers. Both Message Lift test types are comparative simulations. The two evidence classes never mix.
Current/New prompt parity. Both messages use the same prompts and settings. Only the message changes.
Provider and model disclosure. Every figure names the API/model route and mode that produced it.
Sample size and uncertainty. Rates show n and confidence intervals where the metric supports one.
Results that say no. A study that only ever confirms the brief is not measuring anything. Where the evidence does not support a claim, the report says so, and that is the finding.
Snapshot limitations. Conclusions apply only to the tested prompts, routes, modes, and date.

FAQ

Is Resonance a GEO or AEO tool?

GEO and AEO describe work intended to improve how brands appear in AI answers. Resonance measures AI visibility, citations, competitive ranking, and brand perception, then tests messages. It does not publish content or promise ranking improvements.

How does Message Lift work?

Message Lift begins with a stored AI answer from the same audit. One Current message and one New message then go through shared contexts, so only the message changes. Buyer-response and AI-recommendation results are both labeled simulated.

Why run the same audit prompt more than once?

LLM answers are probabilistic. One answer is an anecdote. Repeated answers support a rate, with a confidence interval where the math supports one.

Do results match a consumer chat interface exactly?

No. Resonance uses disclosed API/model routes and settings. Those can differ from a personalized consumer chat session. The site makes no consumer-interface equivalence claim.