-
Online revenue
300× growthPeriod: Sep–Nov 2021
-
Monthly active users (MAU)
5× growthPeriod: Sep–Nov 2021
-
Brand online buzz
50th+ #1Buzz ranking within the category
-
Category word-of-mouth
10th+ #1Word-of-mouth ranking within the category
PARTNERS
Company structure and track record.
GROUP BRANDS
Group brand matrix
The parent company, the technology licensor, and five outward-facing solutions. Every cell links to a corresponding public resource.
TRACK RECORD
The founding team's past campaigns
All from projects the team was involved in before CiteTek was established; clients have been anonymized.
BUSINESS TYPES
Business types we take on
AI recommends the brands it can parse, verify and safely quote. Every business type carries its own retrieval gap — what AI reads today, the three weakest of our twelve diagnostic dimensions, and how we measure the correction. Engine coverage scales with the plan, and every card states how far it reaches.
-
01
Live for 8 weeks
Owner reports a clear lift in footfall
B2B software startup
Applies to SaaS, enterprise software, subscription tools
Retrieval scenario
"Which SaaS vendors cover this category?" "How does A differ from B?" — the comparison queries buyers run during vendor selection.
Diagnostic findings
Positioning, company-size fit and pricing sit on separate pages and contradict each other. The model cannot resolve a single product entity, so the brand never enters the shortlist.
Engines monitored
- ChatGPT
- Claude
- Gemini
- Perplexity
- Grok
- Mistral
- Google AI Overview
Primary gap dimensions
-
Semantic Structure
conflicting positioning single resolved entity
Site, pricing page and FAQ describe the same product the same way.
-
Content Citability
no extractable passages selection criteria citable
Fit, limits and competitive comparison each stand alone as a quotable passage.
-
Structured Data
product markup missing plans and pricing machine-readable
Modules, tiers and supported platforms emitted as structured fields.
Measurement
We re-run the same selection queries across six engines weekly, logging mention rate, the product category assigned, and any drift from your official wording.
-
02
Live for 8 weeks
Owner reports a clear lift in footfall
Technical apparel brand
Applies to E-commerce, consumer goods, direct-to-consumer
Retrieval scenario
"How do I choose a sun-protection jacket?" "Which brands make this kind of gear?" — pre-purchase comparison and filtering.
Diagnostic findings
Product pages carry names and prices only. Fabric, fit, sizing and care conditions were never written out, so marketplaces and price-comparison sites own the answer.
Engines monitored
- ChatGPT
- Gemini
- Perplexity
- Grok
- Google AI Overview
Primary gap dimensions
-
Structured Data
product fields incomplete full item specifications
Materials, specs, availability and price emitted as Product fields.
-
Content Citability
sources sit with resellers own passages citable
Selection criteria and material comparisons rewritten as standalone passages.
-
Snippet Control
summaries pulled from third parties summaries pulled from your site
Extractable ranges defined so reseller copy stops being the default answer.
Measurement
We track the source-domain distribution behind the same buying queries, log your site's share, and check every material and spec claim against the product page.
-
03
Live for 7 weeks
Owner reports a clear lift in footfall
Specialist clinic
Applies to Medical practices, group clinics, regional care
Retrieval scenario
"Where do I go locally for this specialty?" "What does the treatment pathway involve?" — pre-visit evaluation.
Diagnostic findings
Service scope, physician expertise and location data are split across pages with no verifiable credential signal, so the model files the clinic as a generic provider.
Engines monitored
- ChatGPT
- Claude
- Gemini
- Perplexity
- Google AI Overview
Primary gap dimensions
-
E-E-A-T Authority
weak credential signal verifiable practice record
Licences, specialty scope and service area corroborate each other.
-
FAQ Readiness
long-form patient education discrete answer units
Symptoms, procedure, risk and aftercare each map to an answerable question.
-
Semantic Structure
specialty boundaries blurred primary service explicit
Heading hierarchy and internal links separate core specialty from secondary services.
Measurement
We re-ask the questions patients actually type and verify specialty scope and physician expertise line by line. Overstated efficacy or misattribution enters the current correction cycle.
-
04
Live for 7 weeks
Owner reports a clear lift in footfall
Vocational training provider
Applies to Online courses, education, corporate training
Retrieval scenario
"How do I get started in data analysis?" "What order should I learn this in?" — learning-path and course-selection planning.
Diagnostic findings
Course pages list a syllabus but no prerequisites, outcomes or target learner. The model cannot rank difficulty or sequence courses relative to each other.
Engines monitored
- ChatGPT
- Gemini
- Perplexity
- Grok
- Google AI Overview
Primary gap dimensions
-
Semantic Structure
courses stand isolated path relationships modelled
Prerequisite, advanced and follow-on courses are structurally linked.
-
FAQ Readiness
selection questions unanswered selection Q&A complete
Which path for which background, how long it takes, what you produce — each answered.
-
Content Citability
outcomes stated vaguely concrete learning outcomes
Assessable skills and portfolio output written as quotable statements.
Measurement
We run one round each for beginner, career-change and upskilling phrasing, logging whether courses land in the correct difficulty tier and whether the suggested order respects prerequisites.
-
05
Live for 6 weeks
Owner reports a clear lift in footfall
Investment advisory firm
Applies to Financial services, insurance broking, investor education
Retrieval scenario
"Where do I start with retirement planning?" "How do the fees compare?" — knowledge and fee verification before a decision.
Diagnostic findings
Only scattered articles exist. Service scope, risk disclosure and audience fit never form a coherent record, so the model cites third-party media over your own pages.
Engines monitored
- ChatGPT
- Claude
- Gemini
- Perplexity
- Grok
- Mistral
- Google AI Overview
Primary gap dimensions
-
E-E-A-T Authority
practice signals scattered credentials and disclosure in place
Advisor licensing, regulatory status and risk statements are centralised and checkable.
-
Content Citability
fee data not extractable fees and conditions citable
Fee structure, audience fit and constraints written as standalone units.
-
Structured Data
services unmarked services and Q&A machine-readable
Service lines and common concerns emitted as structured fields.
Measurement
We check the model's paraphrase against your compliance wording line by line and log whether risk notes surface alongside return claims. Misleading simplification sends us back to the source page.
-
06
Live for 6 weeks
Owner reports a clear lift in footfall
Commercial office brokerage
Applies to Property advisory, local agencies, leasing management
Retrieval scenario
"What are office rents like in this district?" "What should a company check when leasing?" — market and process queries before a decision.
Diagnostic findings
The site is a listing feed with no district research, leasing process or industry-specific service narrative, so portals and news own the answer.
Engines monitored
- ChatGPT
- Claude
- Gemini
- Perplexity
- Grok
- Google AI Overview
Primary gap dimensions
-
Content Citability
listings only market insight citable
District comparisons, rent bands and process notes become extractable knowledge units.
-
Structured Data
service scope unmarked coverage machine-readable
Cities, districts and service lines emitted as structured fields.
-
E-E-A-T Authority
thin professional evidence local expertise verifiable
Licensing, transaction types and published research corroborate each other.
Measurement
We verify every district and rent figure in the answer against your monthly report, and track how the citation mix shifts from portals back to your site.
-
07
Live for 8 weeks
Owner reports a clear lift in footfall
Long-established parts maker
Applies to Traditional manufacturing, industrial parts, contract production
Retrieval scenario
"Who supplies this specification?" "What's a cross-reference for this part number?" — engineering and procurement lookups.
Diagnostic findings
Specifications are sealed in catalogue PDFs. Target industries and material limits were never written as text, so crawlers retrieve no comparable spec values.
Engines monitored
- ChatGPT
- Claude
- Gemini
- Perplexity
- DeepSeek
- 通義千問
- Mistral
- Google AI Overview
Primary gap dimensions
-
AI Crawler Access
specs locked in PDF spec pages retrievable
Catalogues split into indexable pages so crawlers reach the spec fields.
-
Semantic Structure
part numbers without hierarchy product taxonomy resolved
Series, part number and spec conditions parse as a hierarchy.
-
Content Citability
applications undocumented selection criteria citable
Fit conditions, cross-references and industry applications become standalone passages.
Measurement
We test by part number, by specification and by application, then compare every retrieved value against the catalogue for exact agreement.
-
08
Live for 5 weeks
Owner reports a clear lift in footfall
Independent consultant
Applies to Personal brands, experts, speakers, founders
Retrieval scenario
"Who can I ask about this field?" "Who is this person and what do they do?" — discovery and identity verification.
Diagnostic findings
The record is stitched from platform bios, interviews and social posts. Titles and history disagree, so the person entity never resolves to one individual.
Engines monitored
- ChatGPT
- Claude
- Gemini
- Perplexity
Primary gap dimensions
-
E-E-A-T Authority
conflicting biographies cross-verifiable record
Title, publications, press and speaking history corroborate each other.
-
Structured Data
person unmarked identity fields complete
Name, role, affiliation and external profiles linked as structured fields.
-
Language Naturalness
bio reads as keyword stuffing expert prose
Expertise stated the way a practitioner would state it, not as boilerplate.
Measurement
We fact-check across "who they are", "what they do" and "what you'd hire them for", logging each wrong statement and correcting it at its source page.
-
09
Live for 4 weeks
Owner reports a clear lift in footfall
Tourism factory
Applies to Attractions, local venues, destinations
Retrieval scenario
"How should I plan a day here?" "What's the rainy-day option?" — itinerary planning by place and time slot.
Diagnostic findings
Operating details contradict each other across your site, social and third-party platforms. The model prefers travel blogs, and prices and hours run years out of date.
Engines monitored
- ChatGPT
- Claude
- Gemini
- Perplexity
- Google AI Overview
Primary gap dimensions
-
Structured Data
operating fields missing venue record complete
Hours, admission, transport and parking emitted as structured fields.
-
FAQ Readiness
itinerary contexts absent scenario Q&A complete
Family, rainy-day, group and half-day visits each map to an answerable question.
-
Snippet Control
stale data extracted current data prioritised
Extraction scope limited so superseded pricing stops being quoted.
Measurement
We run one round per season and per weather scenario, focusing on hours and admission. Any stale citation is corrected and retested within the week.
-
10
Live for 4 weeks
Owner reports a clear lift in footfall
Restaurant chain
Applies to Restaurants, cafes, local food
Retrieval scenario
"Where should we book for a group dinner?" "Somewhere good for a date?" — situational filtering before booking.
Diagnostic findings
The site carries no citable menu or signature-dish description, so the model leans on review platforms and blogs, and branch details fail to match opening hours.
Engines monitored
- ChatGPT
- Gemini
- Perplexity
- Google AI Overview
Primary gap dimensions
-
Structured Data
menu unmarked menu and branches machine-readable
Dishes, price band, hours and booking method emitted as structured fields.
-
FAQ Readiness
booking conditions unclear dining conditions complete
Private rooms, minimum spend, parking and allergens each answered.
-
Content Citability
signature dishes undescribed menu prose citable
Cooking style and signature items written as standalone passages.
Measurement
We spot-check group dinner, date night and family phrasing, verifying dishes, price band and hours against your site.
-
11
Live for 11 weeks
Owner reports footfall and revenue doubled
Long-established pawnshop
Applies to Local businesses, storefronts, lifestyle brands
Retrieval scenario
"Where can I borrow against something nearby?" "How is the interest calculated?" — local queries made at the moment cash is needed.
Diagnostic findings
Information exists only in social posts and map reviews. Service lines, fee basis, licensing and store details have no indexable official version.
Engines monitored
- ChatGPT
- Claude
- Gemini
- Perplexity
- Grok
- Google AI Overview
Primary gap dimensions
-
Structured Data
storefront unmarked business and services machine-readable
Hours, service lines, service area and contact details emitted as structured fields.
-
E-E-A-T Authority
legitimacy unclear licensing verifiable
Issuing authority and licence number presented together as checkable signals.
-
AEO Readiness
concrete questions unanswered lending questions answerable
Rates, process and required documents each map to a direct answer.
Measurement
We test by pledged item, by rate and by process, logging whether the store enters local recommendations and whether fee and licensing statements match your site.
-
12
Live for 7 weeks
Owner reports a clear lift in footfall
Multi-partner law firm
Applies to Professional firms, accounting, design, consulting
Retrieval scenario
"Who handles this kind of case?" "How should an SME pick an accountant?" — credential and specialisation queries before engagement.
Diagnostic findings
Information is scattered across thin service pages and third-party directories, with no verifiable official record of licensing, case types or service area.
Engines monitored
- ChatGPT
- Claude
- Gemini
- Perplexity
- Grok
- Google AI Overview
Primary gap dimensions
-
E-E-A-T Authority
credential signals scattered licensing verifiable
Licence numbers, firm registration and professional writing corroborate each other.
-
Content Citability
services described generically case types citable
Client fit, process and fee assessment written as extractable passages.
-
Semantic Structure
specialisation boundaries blurred practice areas distinct
Practice area and service territory parse as a clear hierarchy.
Measurement
We test with "who should handle this" phrasing and verify specialisation and service area statement by statement, correcting the service page and retesting on any error.
-
13
Live for 6 weeks
Owner reports a clear lift in footfall
Dental practice
Applies to Aesthetics, dentistry, health screening
Retrieval scenario
"How do I choose between these treatments?" "How do screening packages compare?" — treatment and package evaluation.
Diagnostic findings
The model cites forums, ad pages and outdated articles. Treatment specs, candidate criteria and contraindications are conflated, and risk information gets dropped.
Engines monitored
- ChatGPT
- Claude
- Gemini
- Perplexity
- Grok
- Google AI Overview
Primary gap dimensions
-
E-E-A-T Authority
low-credibility sources clinical evidence in place
Clinician licensing, equipment provenance and clinical references form a checkable chain.
-
FAQ Readiness
contraindications not framed as questions constraints answerable
Candidate criteria, contraindications and aftercare each map to a question.
-
Snippet Control
only benefits extracted risk extracted alongside
Extraction scope set so outcomes and limits surface as a pair.
Measurement
We spot-check by treatment name and comparison phrasing, focusing on whether contraindications and risk notes appear alongside efficacy claims. Omissions send us back to the source page.
-
14
Live for 3 weeks
Owner reports a clear lift in footfall
Boutique guesthouse
Applies to Hotels, inns, resorts
Retrieval scenario
"Family-friendly places to stay?" "Somewhere that takes pets?" — condition filtering before booking.
Diagnostic findings
Room types, facilities and surcharge rules have no official version. The model depends on OTAs and review platforms, and stay conditions often contradict your site.
Engines monitored
- ChatGPT
- Gemini
- Perplexity
- Google AI Overview
Primary gap dimensions
-
Structured Data
room types unmarked rooms and rates machine-readable
Room type, occupancy, facilities and surcharge rules emitted as structured fields.
-
FAQ Readiness
stay conditions unclear house rules complete
Pets, extra beds, breakfast and parking each answered.
-
Content Citability
descriptions sourced from OTAs own description citable
Guest fit and nearby itineraries written as standalone passages.
Measurement
We test family and pet scenarios and verify room type, surcharge rules and stay conditions against your site.
-
15
Live for 3 weeks
Owner reports a clear lift in footfall
Personal training studio
Applies to Gyms, yoga, boxing, sports coaching
Retrieval scenario
"Which gym suits a beginner?" "What class works postnatally?" — class and level queries before a trial.
Diagnostic findings
Information exists only in social posts and map reviews. Class level, trainer specialisation and target group have no indexable structure.
Engines monitored
- ChatGPT
- Gemini
- Perplexity
- Google AI Overview
Primary gap dimensions
-
Semantic Structure
classes ungraded level hierarchy modelled
Entry, advanced and specialist classes parse as related tiers.
-
FAQ Readiness
trial process unclear trial Q&A complete
Price, duration, equipment and prerequisites each answered.
-
AEO Readiness
cannot answer who it suits audience fit explicit
Beginner, postnatal, senior and rehab each map to a direct answer.
Measurement
We test beginner, postnatal and strength phrasing, logging whether classes land in the correct level tier and whether trainer specialisation is stated correctly.
-
16
Live for 2 weeks
Owner reports a clear lift in footfall
Industry conference
Applies to Exhibitions, talks, concerts, local festivals
Retrieval scenario
"What's on this week?" "What's the agenda for this conference?" — timing and content queries before registering.
Diagnostic findings
Expired event pages coexist with current ones. The model cannot tell whether registration is open, and often retrieves last year's speakers and agenda.
Engines monitored
- ChatGPT
- Gemini
- Perplexity
- Google AI Overview
Primary gap dimensions
-
Structured Data
sessions unmarked schedule and ticketing machine-readable
Date, venue, ticket type and registration status emitted as Event fields.
-
Snippet Control
prior-year data mixed in current edition prioritised
Extraction scope limited so expired sessions stop reading as current.
-
AI Crawler Access
agenda pages unreachable agenda indexable
Agenda and speaker pages rebuilt as indexable standalone pages.
Measurement
We spot-check before and after the event, logging whether expired sessions are misread as open for registration. Any misread triggers a status-markup fix and a retest.
-
17
Live for 3 weeks
Owner reports a clear lift in footfall
AI tool
Applies to Apps, Chrome extensions, developer products
Retrieval scenario
"Any good tools for this?" "What's an alternative to X?" — comparison and integration queries before adoption.
Diagnostic findings
The only legible content is the homepage tagline. Feature modules, use cases, pricing tiers and integration limits were never written as retrievable text.
Engines monitored
- ChatGPT
- Claude
- Gemini
- Perplexity
- Grok
- Mistral
- DeepSeek
- Google AI Overview
Primary gap dimensions
-
Content Citability
tagline only use cases citable
Use cases, templates and integration flows each become extractable passages.
-
AEO Readiness
limits unanswerable capability boundaries explicit
Supported scope, data handling and known limits each map to a direct answer.
-
AI Crawler Access
agents cannot reach it retrieval paths opened
AI bot rules in robots and documentation paths opened to major agents.
Measurement
We test recommendation, alternative-to and integration phrasing, logging the accuracy of capability and limit statements and whether the product enters comparison lists.
CONTACT
Want to be among the first publicly named partners?
We'll publish research cases for our first cohort of partner brands on this page. If you're interested, let's talk.