Services

/

SEO & GEO

AI now writes your buyer’s shortlist.

A Google search used to hand your buyer the wine list. Now ChatGPT, Gemini and AI Overviews act as the sommelier: they suggest the bottle, and they tell the buyer which one to skip. We measure how often AI answers recommend you, find out why they don’t, and coordinate the site, content and off-site work that changes the verdict.

Illustration of a row of green bottles on a shelf with one blue bottle wearing an orange tag

50%

US consumers intentionally using AI-powered search to guide buying decisions

57.5%

Bought a product because a chatbot recommended it

We find where you’re losing.

By topic, funnel stage, customer segment and AI platform, across prompts written the way your buyers actually ask, including the ones where AI argues against you.

We check whether you should win.

Every gap gets tested against product fit, so the roadmap skips prompts that belong to a specialist.

We trace why you’re losing.

Who gets recommended instead, which sources shaped the answer, and what the model says when asked why.

We report the verdict.

Recommendation rate against your top competitors, next to mentions, citations, organic traffic and conversions.

Every chatbot already has an opinion about your brand. We give it better reasons.

AI visibility is a team sport. Most brands either hand it all to the SEO team, which then owns a number shaped by review sites, Reddit and the press, or let content, PR and social each run their own GEO project. Nobody runs paid search, paid social and programmatic that way. Someone keeps the message consistent, puts budget where it works hardest and times each channel to the others. We bring the same coordination to organic: technical SEO, site content, the blog, digital PR, organic social and creator work run as one program, prioritized by which prompts you’re losing and why. What we learn shapes your paid search copy and ChatGPT Ads too.

We go after the recommendations you should be winning, and skip the ones you shouldn’t.

One answer, many owners

Google’s AI and the chatbots draw on the same mix of sources, in different proportions by platform and prompt. We read the citations before we set priorities.

Swipe to view full table →

What shapes the answer

Usually owned by

What we coordinate

Technical SEO

SEO team or developers

Crawling, indexation, speed, rendering and structured data every engine can read

Site content

Web or product marketing

Product, category and comparison pages that say who you’re right for

Blog

Content team

Coverage of a topic from every angle buyers ask about

Product feeds

E-commerce team

Attributes shopping ads and AI answers both draw from → Feed page

Reviews and roundups

Product marketing

Prioritization of the review sites and listicles cited for your prompts

Reddit, YouTube, creators

Social team

Forum engagement plans and content briefs for your own social channels

Press

PR agency

Digital PR outreach playbook for the publications models cite in your category

Technical SEO

Usually owned by

SEO team or developers

What we coordinate

Crawling, indexation, speed, rendering and structured data every engine can read

Site content

Usually owned by

Web or product marketing

What we coordinate

Product, category and comparison pages that say who you’re right for

Blog

Usually owned by

Content team

What we coordinate

Coverage of a topic from every angle buyers ask about

Product feeds

Usually owned by

E-commerce team

What we coordinate

Attributes shopping ads and AI answers both draw from → Feed page

Reviews and roundups

Usually owned by

Product marketing

What we coordinate

Prioritization of the review sites and listicles cited for your prompts

Reddit, YouTube, creators

Usually owned by

Social team

What we coordinate

Forum engagement plans and content briefs for your own social channels

Press

Usually owned by

PR agency

What we coordinate

Digital PR outreach playbook for the publications models cite in your category

We report recommendation rate.

A mention counts the same whether you’re the first choice, one name in ten, or the brand AI tells buyers to avoid. Recommendations are what move buyers: in Scrunch AI’s clickstream research, they lifted brand search and site visits two to three times as much as neutral mentions. Mentions and citations stay in the report as evidence of reach and of what the model read. Sentiment is tone. A recommendation is a verdict.

Recommendations come in degrees. From best to worst, here is where an answer can put you, and what each metric actually counts.

The recommendation ladder

Where the answer puts you

Mentions

Rec. rate

Absolute

1 · Absolute recommendation

You’re the clear first choice, or the only option offered.

✓
✓
✓

2 · Recommendation

One of several suggested options: “here are the three best trail running shoes to consider.”

✓
✓
—

3 · Neutral mention

The answer names you without saying whether you’re a good or bad choice.

✓
—
—

4 · Not mentioned

You’re absent from the answer.

—
—
—

5 · Recommended against

The answer names you and steers the buyer away: too hard to implement, too expensive to be worth it.

✓still counts

—
—

1 · Absolute recommendation

You’re the clear first choice, or the only option offered.

Mentions

✓

Rec. rate

✓

Absolute

✓

2 · Recommendation

One of several suggested options: “here are the three best trail running shoes to consider.”

Mentions

✓

Rec. rate

✓

Absolute

—

3 · Neutral mention

The answer names you without saying whether you’re a good or bad choice.

Mentions

✓

Rec. rate

—

Absolute

—

4 · Not mentioned

You’re absent from the answer.

Mentions

—

Rec. rate

—

Absolute

—

5 · Recommended against

The answer names you and steers the buyer away: too hard to implement, too expensive to be worth it.

Mentions

✓still counts

Rec. rate

—

Absolute

—

The bottom rung is the one most dashboards miss. A brand that gets recommended against still logs a mention, and it’s worse off than a brand that never came up.

Recommendations come in degrees. From best to worst, here is where an answer can put you, and what each metric actually counts.

The recommendation ladder

Where the answer puts you

Mentions

Rec. rate

Absolute

1 · Absolute recommendation

You’re the clear first choice, or the only option offered.

✓
✓
✓

2 · Recommendation

One of several suggested options: “here are the three best trail running shoes to consider.”

✓
✓
—

3 · Neutral mention

The answer names you without saying whether you’re a good or bad choice.

✓
—
—

4 · Not mentioned

You’re absent from the answer.

—
—
—

5 · Recommended against

The answer names you and steers the buyer away: too hard to implement, too expensive to be worth it.

✓still counts

—
—

1 · Absolute recommendation

You’re the clear first choice, or the only option offered.

Mentions

✓

Rec. rate

✓

Absolute

✓

2 · Recommendation

One of several suggested options: “here are the three best trail running shoes to consider.”

Mentions

✓

Rec. rate

✓

Absolute

—

3 · Neutral mention

The answer names you without saying whether you’re a good or bad choice.

Mentions

✓

Rec. rate

—

Absolute

—

4 · Not mentioned

You’re absent from the answer.

Mentions

—

Rec. rate

—

Absolute

—

5 · Recommended against

The answer names you and steers the buyer away: too hard to implement, too expensive to be worth it.

Mentions

✓still counts

Rec. rate

—

Absolute

—

The bottom rung is the one most dashboards miss. A brand that gets recommended against still logs a mention, and it’s worse off than a brand that never came up.

Recommendations come in degrees. From best to worst, here is where an answer can put you, and what each metric actually counts.

The recommendation ladder

Where the answer puts you

Mentions

Rec. rate

Absolute

1 · Absolute recommendation

You’re the clear first choice, or the only option offered.

✓
✓
✓

2 · Recommendation

One of several suggested options: “here are the three best trail running shoes to consider.”

✓
✓
—

3 · Neutral mention

The answer names you without saying whether you’re a good or bad choice.

✓
—
—

4 · Not mentioned

You’re absent from the answer.

—
—
—

5 · Recommended against

The answer names you and steers the buyer away: too hard to implement, too expensive to be worth it.

✓still counts

—
—

1 · Absolute recommendation

You’re the clear first choice, or the only option offered.

Mentions

✓

Rec. rate

✓

Absolute

✓

2 · Recommendation

One of several suggested options: “here are the three best trail running shoes to consider.”

Mentions

✓

Rec. rate

✓

Absolute

—

3 · Neutral mention

The answer names you without saying whether you’re a good or bad choice.

Mentions

✓

Rec. rate

—

Absolute

—

4 · Not mentioned

You’re absent from the answer.

Mentions

—

Rec. rate

—

Absolute

—

5 · Recommended against

The answer names you and steers the buyer away: too hard to implement, too expensive to be worth it.

Mentions

✓still counts

Rec. rate

—

Absolute

—

The bottom rung is the one most dashboards miss. A brand that gets recommended against still logs a mention, and it’s worse off than a brand that never came up.

Recommendations come in degrees. From best to worst, here is where an answer can put you, and what each metric actually counts.

The recommendation ladder

Where the answer puts you

Mentions

Rec. rate

Absolute

1 · Absolute recommendation

You’re the clear first choice, or the only option offered.

✓
✓
✓

2 · Recommendation

One of several suggested options: “here are the three best trail running shoes to consider.”

✓
✓
—

3 · Neutral mention

The answer names you without saying whether you’re a good or bad choice.

✓
—
—

4 · Not mentioned

You’re absent from the answer.

—
—
—

5 · Recommended against

The answer names you and steers the buyer away: too hard to implement, too expensive to be worth it.

✓still counts

—
—

1 · Absolute recommendation

You’re the clear first choice, or the only option offered.

Mentions

✓

Rec. rate

✓

Absolute

✓

2 · Recommendation

One of several suggested options: “here are the three best trail running shoes to consider.”

Mentions

✓

Rec. rate

✓

Absolute

—

3 · Neutral mention

The answer names you without saying whether you’re a good or bad choice.

Mentions

✓

Rec. rate

—

Absolute

—

4 · Not mentioned

You’re absent from the answer.

Mentions

—

Rec. rate

—

Absolute

—

5 · Recommended against

The answer names you and steers the buyer away: too hard to implement, too expensive to be worth it.

Mentions

✓still counts

Rec. rate

—

Absolute

—

The bottom rung is the one most dashboards miss. A brand that gets recommended against still logs a mention, and it’s worse off than a brand that never came up.

Work with us

Outgrow your category.

Tell us where growth has stalled. We’ll benchmark you against your competitors, show you where you’re losing ground, and bring the first moves we’d make. The audition is on us.

Work with us

Outgrow your category.

Tell us where growth has stalled. We’ll benchmark you against your competitors, show you where you’re losing ground, and bring the first moves we’d make. The audition is on us.

Work with us

Outgrow your category.

Tell us where growth has stalled. We’ll benchmark you against your competitors, show you where you’re losing ground, and bring the first moves we’d make. The audition is on us.

Work with us

Outgrow your category.

Tell us where growth has stalled. We’ll benchmark you against your competitors, show you where you’re losing ground, and bring the first moves we’d make. The audition is on us.