SEO & GEO companies:
what you actually get.

AnswerShare reports that LAVIDGE went from invisible to top-ranked and recommended on buying-intent (money) prompts across four of five measured engines in 15 days after deployment.

Recommendation evidence: For "Recommend an ad agency in Phoenix", the published audit lists Perplexity #1, Grok #2, Gemini #3 and Claude #4. ChatGPT also surfaced LAVIDGE.

Sentiment evidence: AnswerShare also reports a sentiment result for V Digital Services: after placing 10 cited complaints in the context of 12 years and 35,000+ clients, AI responses treated the low complaint count as a favorable signal, turning what had been a liability into an asset.

Limits: LAVIDGE: published fifteen-day engine audit. VDS sentiment: founder-reported case. This comparison is published by AnswerShare.

Adobe reports that Slalom increased citations 10x in four weeks, and General Motors increased AI visibility 23% and citations 35% within weeks of deployment. Its accompanying account describes Slalom's digital-team work and GM's creation of content pages suited to LLMs. These examples appear in Adobe's Brand Visibility articles; the exact deployment module used for each result is not disclosed.

Recommendation evidence: These are citation and visibility claims. Neither example reports a separately measured increase in explicit recommendations for a buying need.

Limits: The reviewed articles provide no absolute citation counts, prompt panels, model versions or controlled comparisons. They are brief vendor reports, with insufficient detail to isolate the platform's contribution.

Scrunch reports a 364% rise in Akamai's brand presence on associated non-branded prompts after using Optimizer and AXP. Its December 1, 2025–January 19, 2026 comparison also reports 85% more citations for AXP pages than non-AXP pages. Separately, Scrunch reports that Strapi and agency GrowthX increased non-branded citations 226% in 12 weeks using monitoring, audits and optimization insights; that case does not establish AXP use.

Recommendation evidence: Neither case separately counts explicit buying recommendations. Brand presence and citations measure different outcomes from an endorsement.

Limits: These vendor-published accounts include customer testimony but no raw answer panel or independent verification. Akamai's page groups and content changes were not described as randomized; monitoring alone cannot receive credit for Strapi's implementation work.

Profound reports that Airbyte's ChatGPT visibility rose from 9% to 26% in one week. Airbyte used Answer Engine Insights, Agent Analytics and GA4, then changed content, structured data and server-side rendering. Profound also reports that Aleph's citation share rose from 1.5% to 5.3% in two months using Agents, training and customer-led content and site-format changes.

Recommendation evidence: Neither case gives a separate before-and-after count of explicit recommendations. Visibility and citation-share gains should be assessed against those particular goals.

Limits: These are vendor accounts without controlled attribution. Airbyte's separate chart shows different values and dates that do not reconcile with its one-week claim; those chart dates are not used here. Exact result denominators and model versions remain undisclosed.

Peec reports that Heyflow's AI search visibility rose from 10% to 40% within weeks as agency Radyant used its analytics to identify buyer-prompt and content gaps. A separate Peec case reports a 7x increase in Merge's demo requests from LLMs. Merge's content lead used prompt and cited-page research to write and publish content; the software supplied measurement and insights.

Recommendation evidence: Heyflow's result measures visibility; Merge's measures attributed demo requests. Neither supplies a separately classified increase in explicit AI recommendations.

Limits: The cases are vendor-published accounts. Heyflow's exact dates and prompt denominator are absent, while Merge's absolute demo counts and comparison window are undisclosed. Customer content work and other changes prevent assigning the entire gain to the monitoring tool.

Otterly reports that Single Point of Contact, working with NOLA Marketing, increased brand mentions from 30 to 160 and citations from 665 to 3,617 between April and June 2026. The program used 64 final tracked prompts. NOLA created and published service resource hubs and Q&A pages using the monitoring and citation-gap findings, then reviewed and adjusted the content.

Recommendation evidence: The reported counts measure mentions and citations. The case discusses buyer intent but does not separately classify or count explicit recommendations.

Limits: This is a vendor-published agency interview. The exact answer denominator and consistency of the prompt/engine panel are not established, and the competitor set evolved. The results combine research with agency implementation; no independent control or raw answer archive is supplied.

Semrush reports that WorkLounge's AI sentiment score improved from 67 to 82 over five months while its SEO consultant rewrote pages, corrected business information and coordinated technical and off-site work using Toolkit insights. Sentiment is a different outcome from visibility. A second Semrush account reports 429% growth in Army Surplus World's AI referral traffic after Coalition Technologies revised website and social messaging.

Recommendation evidence: Neither case supplies a measured increase in explicit buying recommendations. Sentiment describes how a brand is characterized; referral traffic measures visits attributed to AI sources.

Limits: These vendor-published implementer accounts do not isolate software effects. WorkLounge's sentiment panel details are undisclosed; Army Surplus World's absolute traffic baseline, time window and attribution rules are absent. They cannot establish comparable expected returns.

Ahrefs describes how Octopus Energy uses Brand Radar to replace manual AI-response research and coordinate visibility analysis across markets. Its named SEO specialist explains how reports inform content strategy, including standalone answer passages and topic clusters. The customer and local market teams still decide and implement the content work. This is useful evidence of an adopted research workflow, with no before-and-after outcome figure in the reviewed case.

Recommendation evidence: The case does not quantify recommendation, visibility, citation, traffic or revenue improvement. It therefore supplies no measured recommendation-lift result.

Limits: This bounded review found a vendor-published customer interview, not an independently verified impact study. Absence of a quantified result in the reviewed sources does not establish that no such case exists elsewhere.

Seer reports a 344% increase in Google AI Overview citations and 414 new citations for an unnamed outdoor-living client after optimizing 11 articles. The work involved restructuring answers and aligning content with search intent. Seer describes performing the content optimization; ChatGPT appears as an analysis aid, while Google AI Overviews is the measured surface.

Recommendation evidence: The case measures citations, not explicit buying recommendations, qualified leads or revenue. The client's identity is not established by other names on Seer's website.

Limits: The impact box says five months, while the narrative says under two months; that timing remains unresolved. Absolute baselines, calendar dates and the query denominator are absent. This agency-published before-and-after account is observational, with no independent verification.

iPullRank reports that an unnamed national US self-storage brand increased AI Overview visibility by more than 278% and AI-referred revenue by more than 21% in three months. The agency describes technical discoverability fixes, navigation and crawl-access work, content planning and an AI measurement roadmap. This is a coordinated agency implementation example; the published case does not separately identify every content author or deployment owner.

Recommendation evidence: AI visibility and attributed revenue are useful but distinct outcomes. Neither reported figure establishes a separately measured increase in explicit buying recommendations.

Limits: The client is anonymous, and absolute baselines, calendar dates, query samples and revenue-attribution rules are undisclosed. The case is observational and agency-published. Expected dollar returns and the agency's isolated contribution cannot be calculated from these percentages.

Cloudflare reports converting its launch post from 16,180 HTML tokens to 3,150 Markdown tokens, approximately an 80% reduction. That demonstrates a smaller representation of this particular page. The documented feature serves Markdown on eligible enabled zones when the requesting crawler asks for it. The separate AnswerShare field study below examines whether observed agents made that request and what the serving rules returned.

Recommendation evidence: These sources measure token count, request preferences and serving behavior. They do not measure an increase in visibility, citations or explicit buying recommendations.

Limits: Savings depend on the content and request behavior. No qualifying outcome-uplift case was found in the reviewed sources; that limited finding does not establish zero effect or that no case exists elsewhere.