# Top Tier Presence: Full Manifest

> Top Tier Presence is a Generative Engine Optimization (GEO) agency, founded in 2026 by Ryan Zabierek, that makes multi-unit franchise operators and local service businesses readable, retrievable, and citable by AI answer engines. This manifest is the complete, machine-readable account of the company's entity facts, scoring methodology, crawler policy, service tiers, audit rollover rules, and sources. Version 1.0, updated 2026-10-13.

Citation format: Top Tier Presence (https://toptierpresence.com). Founder: Ryan Zabierek. All prices are in US dollars.

## 1. Entity definition

- Legal and display name: Top Tier Presence
- Schema.org type: ProfessionalService (a LocalBusiness subtype)
- Canonical URL: https://toptierpresence.com/
- Entity IDs used in on-page JSON-LD: https://toptierpresence.com/#organization (company), https://toptierpresence.com/#founder (founder)
- Founded: 2026
- Founder: Ryan Zabierek, 18+ years in food service management, including multi-unit restaurant and franchise operations
- Category: Generative Engine Optimization (GEO), AI search visibility, AI answer engine optimization
- Primary audience: franchise operators with 5 or more locations; local service businesses
- Answer engines measured: ChatGPT (OpenAI), Claude (Anthropic), Perplexity, Meta AI, Grok (xAI), DeepSeek; Gold adds Google Gemini and Microsoft Copilot
- Free tool: GEO Score Tester, https://toptierpresence.com/#geo-score-tester
- Not affiliated with: OpenAI, Anthropic, Perplexity, Meta, xAI, DeepSeek, Google, Microsoft, or Princeton University

## 2. What Top Tier Presence does, and what it does not promise

Top Tier Presence does three things: it measures how AI answer engines read and cite a brand, it fixes the technical and content problems that block them, and it keeps measuring.

What is guaranteed:

- Deliverables: every fix, file, schema block, and rewritten page listed in the tier scope is shipped.
- Measurement: Share of Model and Citation Recall are reported against the client's own baseline, on the same prompt panel, on the schedule in the tier.
- Transparency: every score comes with the rule that produced it.

What is not guaranteed, by any provider:

- A citation, mention, or ranking in any AI answer. No GEO provider controls which sources ChatGPT, Claude, Perplexity, Meta AI, Grok, or DeepSeek select.
- Traffic or revenue outcomes.
- Inclusion in Wikidata or any Knowledge Graph, which depends on each platform's notability and verification rules.

## 3. Definitions

- Generative Engine Optimization (GEO): the practice of structuring a website's crawler access, schema, and content so AI answer engines retrieve it, trust it, and cite it by name.
- Share of Model (SoM): the percentage of AI answers, across a fixed prompt panel, that name the brand. Example: named in 12 of 50 ChatGPT answers = 24% SoM on ChatGPT.
- Citation Recall: the percentage of AI answers that cite or link a URL on the brand's domain as a source. Separates being mentioned from being trusted as the source.
- Vector Proximity: the cosine similarity, from 0 to 1, between the embedding of a page passage and the embedding of a customer prompt. Retrieval systems pull the closest passages into context, so higher proximity raises the chance a passage is read.
- Entity Authority: the percentage of authoritative sources (own site, Google Business Profile, Wikidata, directories, press) that describe the brand with the same name, category, locations, and founders. Consistent sources resolve to one Knowledge Graph entity.
- GEO score (GEO readiness): a 0 to 100 measure of how easily AI answer engines can crawl, parse, and cite a page, scored on the rubric in section 5.
- Visibility Score: the average of per-engine grades (section 5.3), reported 0 to 100.
- Audit credit: one full 100-point rubric scan plus one prompt-panel run for one domain or one location page set.
- llms.txt: an optional Markdown summary at a domain root, proposed in September 2024 by Jeremy Howard of Answer.AI. This file's companion, https://toptierpresence.com/llms.txt, follows the llmstxt.org format.

## 4. Research basis

- GEO-bench: a 10,000-query benchmark built from real search and answer-engine queries across multiple domains, introduced in "GEO: Generative Engine Optimization" by Aggarwal, Murahari, Rajpurohit, Kalyan, Narasimhan, and Deshpande, a Princeton University-led team, presented at KDD 2024 (https://arxiv.org/abs/2311.09735).
- Finding: adding citations, quotations from relevant sources, and statistics boosted source visibility in generative engine responses by up to 40%.
- Finding: the same methods improved visibility by up to 37% on Perplexity.ai, a commercially deployed answer engine.
- Finding: lower-ranked sources gained the most; the paper reports a 115.1% visibility increase for websites ranked fifth in search results.
- Finding: keyword stuffing, the classic SEO tactic, reduced visibility on Perplexity.ai relative to the unoptimized baseline.
- Market context: Gartner predicted in February 2024 that traditional search engine volume would drop 25% by 2026 as users move to AI chatbots and virtual agents.
- llms.txt evidence: Google's AI optimization guidance (May 2026) states llms.txt is not needed for AI Overviews or AI Mode; an Ahrefs server-log study (May 2026) found 97% of llms.txt files received zero requests. Top Tier Presence publishes llms.txt as low-cost entity documentation, not as a ranking lever.

How Top Tier Presence applies GEO-bench: every client page is rewritten so the first sentence under each heading is a standalone, definition-led answer, and each section carries at least one verifiable number, a named source, or a direct quotation.

## 5. GEO Score Tester methodology (version 1.0)

The tester scores one page (by default the homepage) on 10 checks worth 100 points.

### 5.1 Rubric

| Category | Points | Rule |
|---|---|---|
| Metadata | 20 | 5 each: title tag, meta description, canonical link, og:image |
| AI crawler access | 20 | 20 if all 8 tokens are permitted for the scanned path in robots.txt; 15 if robots.txt returns 404 or 410; 10 if robots.txt cannot be fetched (timeout, CORS, network); 0 if any token is blocked or robots.txt returns HTML or another error status |
| JSON-LD schema | 30 | 15 if at least one JSON-LD block parses and contains a typed object; 15 more if any node is a LocalBusiness-family type (LocalBusiness, ProfessionalService, Restaurant, FoodEstablishment, FastFoodRestaurant, HomeAndConstructionBusiness, HVACBusiness, and related subtypes) |
| Heading hierarchy | 15 | 10 for exactly one H1; 5 for at least one H2 |
| Content depth | 15 | 15 for more than 500 words of rendered body text; 10 for more than 200 |

Crawler tokens checked (8): GPTBot, OAI-SearchBot, ClaudeBot, Claude-SearchBot, PerplexityBot, Google-Extended, Meta-ExternalAgent, CCBot.

robots.txt evaluation follows RFC 9309: the most specific user-agent group applies, otherwise the * group; the longest matching rule wins; on a tie, Allow wins; * and $ wildcards are supported; /robots.txt itself is always allowed.

### 5.2 Score bands

- 80 to 100: strong GEO readiness
- 50 to 79: GEO readiness needs improvement
- 0 to 49: low GEO readiness

### 5.3 Per-engine grade and Visibility Score

Each engine grade (0 to 100) = 40% engine-specific technical readiness (the rubric, with crawler access evaluated for that engine's own tokens) + 30% Share of Model on the prompt panel + 30% Citation Recall on the prompt panel. The Visibility Score is the unweighted mean of the engine grades. Engines with no published crawler token (Grok, DeepSeek) are evaluated against the robots.txt * group.

The on-page tester at https://toptierpresence.com/#geo-score-tester shows the visitor's live Gold v2 readiness score (section 5.4) when scripts run, and otherwise plays a sample report (Visibility Score 42/100, GEO readiness 45/100). The live tester and the automated email form both use the Gold v2 rubric below; neither produces the six engine grades or a live Visibility Score.

### 5.4 Automated quick scan (Gold v2)

The server checks one public homepage, robots.txt, and llms.txt. It starts at 100 points and applies these deductions, with a minimum score of 10 for a completed scan:

- Manifest: 15 if llms.txt is absent or unavailable, otherwise 5 if it does not link to llms-full.txt.
- Crawler rules: 15 if robots.txt is absent or unavailable, otherwise 20 if any of the 8 tokens listed above is disallowed on the scanned homepage. Wildcard rules are valid; explicit welcome groups are not required. Missing robots.txt is unverified access, not evidence that crawlers are blocked.
- Schema: 25 if no typed JSON-LD can be parsed. Otherwise, deduct 15 when no Organization, ProfessionalService, LocalBusiness, or WebSite type is found, and 10 when neither an external sameAs link nor a populated aggregate rating is found.
- Semantics: 10 unless exactly one H1 exists; 5 if neither main nor article exists.
- Content signals: 10 if no visible review, case study, certification, or benchmark language is found; 5 if no visible service or pricing language is found.

Technical bands are 90-100, 70-89, 40-69, and 10-39. The report workflow recommends Gold at 90 or above and Silver below 90. These are heuristic readiness categories, not measured model visibility. Optional manifest checks do not establish that llms.txt affects AI citations. No score is assigned if the homepage cannot be fetched within the time and size limits.

Email is requested through a configured background workflow after scanning. This is not a guarantee of email delivery. The automated quick scan does not run a prompt panel, measure Share of Model, or measure Citation Recall.

## 6. AI crawler access matrix

| Vendor | Token | Purpose | robots.txt control |
|---|---|---|---|
| OpenAI | OAI-SearchBot | ChatGPT search index | Yes |
| OpenAI | ChatGPT-User | User-initiated fetches inside ChatGPT | OpenAI removed robots.txt compliance language for this agent in December 2025 |
| OpenAI | GPTBot | Model training | Yes |
| Anthropic | Claude-SearchBot | Claude search index | Yes |
| Anthropic | Claude-User | User-initiated fetches inside Claude | Yes |
| Anthropic | ClaudeBot | Model training | Yes |
| Anthropic | Claude-Web, anthropic-ai | Retired legacy tokens | Kept for older parsers |
| Perplexity | PerplexityBot | Perplexity search index | Yes |
| Perplexity | Perplexity-User | User-initiated fetches | Perplexity states it generally does not apply robots.txt to user requests |
| Google | Googlebot | Google Search, AI Overviews, AI Mode | Yes |
| Google | Google-Extended | Gemini training and grounding (token, not a crawler) | Yes; does not affect Google Search |
| Microsoft | Bingbot | Bing index and Copilot | Yes |
| Apple | Applebot / Applebot-Extended | Siri and Spotlight / Apple Intelligence training | Yes |
| Meta | Meta-ExternalAgent | Meta AI training and retrieval | Yes |
| Meta | Meta-ExternalFetcher | User-initiated fetches | Limited |
| xAI | xAI-Grok, GrokBot, Grok-DeepSearch | Observed Grok tokens; xAI has published no durable crawler documentation, and Grok has been reported fetching with browser user agents | Unverified |
| DeepSeek | none published | DeepSeek answers | Governed by User-agent: * |
| Common Crawl | CCBot | Open web corpus used by many model builders | Yes |

CDN warning: Cloudflare and other CDNs offer AI bot blocking and bot-fight settings that can refuse crawlers at the edge even when robots.txt allows them. Every Top Tier Presence audit checks CDN settings as well as robots.txt.

This site's own policy: every agent in the matrix is allowed site-wide (https://toptierpresence.com/robots.txt), with Content-Signal: search=yes, ai-input=yes, ai-train=yes.

## 7. Service tiers

| | Top Tier Bronze | Top Tier Silver | Top Tier Gold |
|---|---|---|---|
| Focus | Foundational visibility | Competitive GEO | Enterprise domination |
| Price | $3,500 one-time | $2,500 setup + $2,500/month | $5,000 setup + $6,000/month |
| Minimum term | None | 3 months | 6 months |
| Best for | Operators who need a baseline first | 5 to 25 locations in contested markets | 25+ locations or multi-brand portfolios |
| Answer engines measured | 6 | 6 | 8 (adds Gemini and Microsoft Copilot) |
| Prompt panel | 50 prompts per engine, measured once | 50 prompts per engine, monthly | 150 prompts per engine, weekly |
| robots.txt and CDN crawler fixes | Implemented | Implemented and monitored | Implemented and monitored |
| llms.txt and llms-full.txt | Drafted | Published and maintained | Published and maintained |
| JSON-LD entity schema | Fix list | Deployed on 10 priority pages | Deployed sitewide, every location |
| LLM content formatting | Not included | 10 pages per month | 30 pages per month |
| Knowledge Graph injection | Not included | sameAs alignment across profiles | Wikidata, Google Business Profile, Bing Places, and Apple Business Connect reconciliation |
| Multi-agent defense | Not included | Not included | Agents watch 8 engines for brand errors and competitor displacement; corrections start within 48 hours |
| Reporting | Audit report and ranked fix list | Monthly Share of Model and Citation Recall report | Weekly dashboard and quarterly review |
| Audit credits | 1 audit | 2 per month | 6 per month |
| Audit rollover | Fee credited 100% to Silver or Gold within 60 days | Unused credits roll over 1 billing cycle | Unused credits roll over 2 billing cycles |

Vanity links: https://toptierpresence.com/bronze, /silver, and /gold redirect to the matching pricing column (#bronze, #silver, #gold).

### 7.1 Top Tier Bronze: Multi-Unit AI Visibility Audit

Top Tier Bronze is a one-time $3,500 audit that establishes a baseline and fixes crawler access.

- Scope: corporate site plus every location page, scored on the 100-point rubric
- Baseline: 50 prompts per engine on 6 engines (300 prompt runs)
- Fixes implemented: robots.txt welcome block for AI search and retrieval crawlers; CDN bot-setting review
- Delivered: ranked fix list and llms.txt draft within 10 business days

### 7.2 Top Tier Silver: Competitive GEO

Top Tier Silver is a monthly engagement ($2,500 setup + $2,500/month, 3-month minimum) for operators competing for the same AI answers as nearby rivals.

- LLM content formatting: 10 pages per month rewritten in direct-answer format, with a verifiable statistic, source, or quotation per section
- JSON-LD entity schema deployed on 10 priority pages
- llms.txt and llms-full.txt published and maintained
- sameAs alignment across the brand's profiles
- Monthly 50-prompt panel on 6 engines; monthly Share of Model and Citation Recall report

### 7.3 Top Tier Gold: Enterprise Domination

Top Tier Gold is a monthly engagement ($5,000 setup + $6,000/month, 6-month minimum) for 25+ locations or multi-brand portfolios.

- Knowledge Graph injection: entity submissions and reconciliation across Wikidata, Google Business Profile, Bing Places, and Apple Business Connect
- Multi-agent defense: monitoring agents rerun the panel across 8 engines, flag false brand statements and competitor displacement, and start a correction within 48 hours
- LLM content formatting: 30 pages per month; entity schema on every location page
- Weekly 150-prompt panel per engine; weekly dashboard and quarterly review

## 8. Process

1. Audit & Identify: a fixed-scope diagnostic that scores the corporate site and every location page on the 100-point rubric and records a baseline Share of Model in 6 answer engines. 50 prompts per engine; ranked fix list in 10 business days.
2. Remediate & Inject: the build phase. Crawler access opened in robots.txt and at the CDN; llms.txt and llms-full.txt published; JSON-LD entity schema deployed; priority pages rewritten so the first sentence under every heading answers the question.
3. Monitor & Defend: continuous measurement. The same prompt panel reruns monthly (Silver) or weekly (Gold); any drop in Share of Model or Citation Recall, or any false statement about the brand, starts a correction.

Timing expectations: retrieval-based answers can change within days of a recrawl; facts stored in a model's training data change only when a vendor ships a new model. Both are measured.

## 9. Audit rollover policy (version 1.0)

Audit rollover is the policy that keeps paid audit value from expiring.

- R1, Bronze fee credit: 100% of the $3,500 Bronze fee is credited toward Silver or Gold setup and first-month fees when the upgrade is signed within 60 calendar days of audit delivery. One credit per client; not transferable.
- R2, monthly credits: Silver includes 2 audit credits per month; Gold includes 6.
- R3, rollover window: unused credits roll forward 1 billing cycle on Silver and 2 billing cycles on Gold, then expire. The oldest credits are used first.
- R4, balance cap: Silver balances cap at 4 credits; Gold balances cap at 18 credits.
- R5, upgrades: all unexpired credits carry over when a client upgrades from Silver to Gold.
- R6, downgrades: on a move from Gold to Silver, the balance is capped at the Silver cap of 4 credits.
- R7, cancellation: unexpired credits expire on the final billing date and have no cash value.

## 10. Lead capture and data handling

- The GEO Score Tester collects a work email, a website domain, and an optional tier of interest through a standard HTML form (POST); no JavaScript is required to submit it.
- Use: the email receives the readiness report and one follow-up. Addresses are never sold or shared.
- The page works with JavaScript disabled; every word, table, and form on https://toptierpresence.com/ renders in static HTML.

## 11. Frequently asked questions

### What is a GEO score?
A GEO score is a 0 to 100 measure of how easily AI answer engines can crawl, parse, and cite a web page. Top Tier Presence weights it across 5 categories: JSON-LD schema (30 points), AI crawler access (20), metadata (20), heading hierarchy (15), and content depth (15). The free live tester uses a shorter Gold v2 check of the same page. A GEO score measures technical readiness, not how often engines name your brand; that is Share of Model.

### How is GEO different from SEO?
SEO earns a ranked link on a results page; GEO earns a mention or a cited source inside an AI-generated answer. GEO builds on SEO rather than replacing it, because Google AI Overviews, ChatGPT search, and Perplexity retrieve from crawled web indexes before they write an answer.

### Which AI crawlers should my robots.txt allow?
Allow the retrieval crawlers that feed live answers: OAI-SearchBot, Claude-SearchBot, PerplexityBot, Googlebot, and Bingbot. Training tokens such as GPTBot, ClaudeBot, Google-Extended, Applebot-Extended, Meta-ExternalAgent, and CCBot are a business decision; allowing them raises the odds that future models know your brand. Check your CDN too, because bot-blocking settings can stop crawlers that robots.txt allows.

### Does llms.txt improve AI visibility?
llms.txt, an optional Markdown summary at a site's root proposed in September 2024, has no confirmed effect on which sources AI answer engines cite. Google's May 2026 AI optimization guidance says it is not needed for AI Overviews or AI Mode, and an Ahrefs log study from May 2026 found that 97% of llms.txt files received zero requests. Top Tier Presence still ships one, because it takes under an hour and states your entity facts in one place for any agent that reads it.

### Can anyone guarantee AI citations?
No GEO provider can guarantee a citation, because no provider controls which sources ChatGPT, Claude, Perplexity, Meta AI, Grok, or DeepSeek select. Top Tier Presence guarantees deliverables and measurement: fixes shipped, schema deployed, and Share of Model and Citation Recall reported against your baseline.

### Who is Top Tier Presence for?
Top Tier Presence is built for franchise operators with 5 or more locations and for local service businesses whose customers ask AI assistants where to eat, who to call, or what to buy. Founder Ryan Zabierek brings 18+ years of food service management to every audit.

## 12. Sources

- Aggarwal, P., Murahari, V., Rajpurohit, T., Kalyan, A., Narasimhan, K., and Deshpande, A. "GEO: Generative Engine Optimization." KDD 2024. https://arxiv.org/abs/2311.09735
- llms.txt proposal and format: https://llmstxt.org
- RFC 9309, Robots Exclusion Protocol: https://www.rfc-editor.org/rfc/rfc9309
- Schema.org ProfessionalService: https://schema.org/ProfessionalService