Use case Accuracy

Is AI describing your
business accurately?

An answer can name your brand and still describe the company you used to be, or deny a feature you shipped last year. Moose grades every answer that names you against what you wrote in your Brand Truth Profile, and shows you the sentence that does not match.

Watch Moose grade an answer The Brand Truth Profile

Chat Workflows

Klaviyohttps://klaviyo.com

New chat

Home Inbox Context Visibility Workflows Audio Chats Artifact Library Connections

Recents

Where are we described wrong?Chat · 2m Visibility scan · 11 promptsVisibility scan · 20h Brand Truth Profile updatedContext · 1d Capture ChatGPT fan-out qu…Chat · 2d

SamPremium

Observe

Visibility

Klaviyo | paid managed visibility | latest run Aug 1, 10:02 PM

Prompt Manager Run now

Share of mentions

74%

Share of citations

61%

Sentiment

58% positive

Narrative drift

82% aligned

Feature parity

71% accurate

Topic: All AI Engine: All Sentiment: All Narrative: Drifted or partial Parity: Inaccuracy found

Answers that describe you wrong 9 of 51 observations

What is Klaviyo and who is it for? Gemini (local fetch) · “built for small Shopify stores” Drifted Inaccuracy

Best CRM for a consumer brand doing $40M ChatGPT (local fetch) · “no native service or helpdesk layer” Partial Inaccuracy

Klaviyo vs Mailchimp for a growing brand Perplexity (local fetch) · “both are email marketing tools” Partial Needs review

Which platforms do enterprise retailers use? Grok (local fetch) · “more of an SMB tool than an enterprise one” Drifted Good

Gemini (local fetch) response

What is Klaviyo and who is it for?

2026-08-01·Brand & category·Mentioned

Investigate in chat Close

Position #3 Sentiment Neutral Narrative Drifted Parity Inaccuracy found

Why this score

Sentiment: Neutral. The answer states what the product does, with no praise or criticism.

Narrative: "built for small Shopify stores" contradicts the profile, which covers small brands through enterprise retailers.

Parity: The answer denies service and helpdesk tooling. The profile lists both as shipped.

Response, as returned

Klaviyo is a marketing automation tool best known for email, and it is built for small Shopify stores that want automated welcome and abandoned cart flows.

It does not include customer service or helpdesk tooling, so larger retailers usually pair it with a separate support platform.

Sources cited a-review-blog.com/klaviyo-review (2023)

Hi, Moose 0.3.281 Beta Need help?

Step one

Start with the Brand Truth Profile

Write down what is true about the company once in the Brand Truth Profile, or import your brand documents and Moose fills it in from them. Every verdict is measured against what the profile says.

Ground · Brand Truth Profile

How should AI describe your company today?

A B2C CRM for consumer brands: email, SMS, push and reviews on one unified customer profile, with service and analytics in the same platform. Used by small brands through enterprise retailers.

Write the version you would want to see in ChatGPT, Perplexity, Gemini or Google AI answers.

What does AI often get wrong about your company?

Serves the wrong customer type Too small-business focused Too enterprise focused Wrong product category Old products or positioning Misses new products or features Says we lack features we offer Compares us to the wrong competitors Outdated company information Overstates what we do Understates what we do Confuses us with another company

Canonical description and offeringsOne paragraph on what the company is, plus each product or category with its own description. This is the sentence you are trying to see quoted back.

Ideal customer, in a list and in your wordsSegments from a fixed list, plus free text for what a checkbox cannot hold. Drift verdicts mostly come from this field.

Key capabilities and featuresWhat AI should know you can do. Parity is judged against this, so a shipped capability listed here becomes a factual claim an answer can get wrong.

Outdated narratives and vocabularyThe story you have moved on from, the words you prefer, and the words you do not want used about you.

Competitors, with domains and notesWho you compete with, plus the comparisons you want watched closely. This is what catches a competitor feature being credited to you.

Sensitive claims to qualify or avoidRegulated language, unproven numbers, and anything your legal team would not want an engine repeating.

Authoritative and outdated sourcesURLs that tell the current story, and URLs that still tell the old one. When an answer cites the second list, you know why it drifted.

Most damaging mistakes to watch forPick the ones that would cost you a deal: wrong category, wrong customer, incorrect pricing, brand confusion. Those get weighted heaviest.

Without a profile there are no verdicts. Drift and parity are measured against your fields, so Moose leaves them blank.

The two verdicts

Narrative drift and feature parity

Every answer that mentions you gets both verdicts. Each verdict comes with one sentence saying which claim it is about.

Narrative drift

Does the answer describe the company as it is today? It checks the customer type, the category and the positioning against your profile.

Yes The answer contradicts the profile: wrong customer type, wrong category, or positioning you no longer use.

Partial Some claims match the profile and others do not, for example the right category but the wrong customer.

No The description matches what you wrote. Nothing to do.

Feature parity

Does the answer get your capabilities wrong? Leaving a feature out is not counted as an error.

Good Nothing in the answer about your capabilities is wrong, even if it did not list them all.

Needs review A claim is unclear enough that a person should look at it before it is called an error.

Inaccuracy found A claim about what you do or do not do is wrong. The reason quotes the sentence.

What parity flags

Saying you do not have a capability you do have.

Claiming a capability, integration or limit you do not have.

Attributing a competitor's feature to you.

What it does not flag

Leaving features out. An answer that is correct but lists only some of your features is still graded good.

Per engine

Which model is getting you wrong

Every observation is stamped with the engine that produced it, so drift, parity, sentiment, mentions and citations all break down per engine. Every answer where an AI got a feature wrong, and which engine said it, is one filter.

ChatGPT

78%aligned

Two answers still place you in the old category.

Local fetch

Claude

96%aligned

Cites your own docs in four of five answers.

Managed / key

Gemini

61%aligned

Worst drift. Leans on a 2023 review blog.

Local fetch

Grok

74%aligned

Reads you as SMB only on enterprise prompts.

Local fetch

Perplexity

89%aligned

Accurate, but compares you to the wrong two.

Local fetch

Google AI Mode

84%aligned

Category right, customer segment vague.

Local fetch

AI Overviews

80%aligned

Short answers, so one wrong clause carries weight.

Local fetch

Bing Copilot

92%aligned

Mirrors whatever your site says most recently.

Managed / key

DeepSeek

70%aligned

Names the old category on two prompts.

Managed / key

Meta AI

83%aligned

Accurate on features, vague on customer size.

Managed / key

Over time

What changed between runs

Runs go on your schedule, from several times a day to manual, and a scheduled run can email you the summary. What you look at afterwards is the difference between runs.

Trend by day, week or month

Mention rate, citation rate and positive sentiment charted by day, week or month.

Compare two ranges

Compare last month to the month before, or the two weeks either side of a launch, side by side.

Run over run

Narrative alignment, feature accuracy, negative sentiment count, mention rate and citation rate, each with the change since the previous run.

Change highlights per prompt

Lost mention, gained mention, new citation, negative sentiment. Each one names the prompt and engine, with the quote attached.

Filter the whole report

Filter everything by narrative drift or feature parity, so every answer where an AI got a feature wrong is one click.

Diagnostic prompts

Prompts written to catch it

Moose writes prompts from your brand name and the fields you filled in: capabilities, competitor comparisons, customer segments, and the mistakes you already see.

Suggested for review 6 of 11 · you approve each one

Is Klaviyo only for small Shopify stores?Anchored on: misrepresentation: too small-business focused

Does Klaviyo include customer service or helpdesk tools?Anchored on: key capability: service and helpdesk

What is the best B2C CRM for a consumer brand?Anchored on: canonical description and category

Klaviyo vs Attentive for SMS and email togetherAnchored on: comparison to monitor closely

Which platform unifies email, SMS, push and reviews?Anchored on: key capabilities

Do enterprise retailers use Klaviyo?Anchored on: ideal customer segments

Anchored on your brand

Prompts are generated from your brand name and your Brand Truth fields, so the answers they return are about you.

Your questions, word for word

Diagnostic questions you write yourself are passed through unchanged.

Generic queries excluded

Unbranded queries are left out. They do not return answers about you.

You approve every one

Review, edit and approve each suggestion before anything gets tracked. They land in their own Diagnostic Prompts category, separate from your core set.

Where it came from

Where the wrong story came from

When an engine has you wrong, Moose also shows the page it learned it from, which is usually something you can fix.

01 The whole answer is keptFull response text, the snippet that mentioned you, and the sources the engine cited, including whether any of them was one of your own domains.

02 Matched against your outdated sourcesYou listed the pages still telling the old story. When a drifted answer cites one, you can see where it came from.

03 Per-URL citation historyWhether a page is cited now, used to be, or has been replaced by a competitor. That history feeds the AEO audits, so an audit of that page starts from its citation history.

Questions people ask

Do I have to fill in the whole Brand Truth Profile?

Every field is optional, but the profile is what answers are measured against, so an empty one means no drift or parity verdicts at all. The canonical description, ideal customer and key capabilities do most of the work. Start with those three and add the rest as you see what gets flagged.

Does this flag an answer for leaving features out?

No. Feature parity grades misinformation, and leaving a feature out is not a fault. An answer that stays factually correct while mentioning only two of your six products is still good. It only flags a claim that is wrong: a capability denied, a capability invented, or a competitor's feature credited to you.

How do I know a verdict is not a hallucination?

Each verdict carries a one-sentence reason that has to name the specific claim it is about, and sentiment verdicts carry an exact quote copied from the answer. If the scoring call fails, the fields stay empty rather than being filled with a guess.

Which engines can it read?

ChatGPT, Gemini, Grok, Perplexity, Google AI Mode, Google AI Overviews, Claude, DeepSeek, Meta AI and Bing Copilot. The first six are read from the consumer product in a browser session on your computer, with an API fallback if that fetch fails. Every observation records which route produced it.

Can a single bad answer swing my numbers?

Not much, because the same prompt and engine can be sampled up to five times per run and the results carry a Wilson 95% confidence margin. Repeated samples also collapse to one verdict in lists, emails and alerts, so a heavily cited prompt does not flood the report.