Get named when buyers ask ChatGPT or Perplexity who to pick
Why recommendations go to competitors today — and the exact crawler rules, copy patterns, schema graph, and 10-minute monthly check that put your brand on the shortlist.
Enter your work email (@company.com) to unlock all playbooks and PDFs.
Free · Work email unlocks all 10 guides · Privacy policy
A quiet shift in B2B discovery means buyers shortlist 2–3 trusted options inside AI chat before opening a single browser tab.
Locked
02
How bots read raw HTML & Cloudflare firewall rules
Separate live citation indexers (OAI-SearchBot, PerplexityBot, Claude-SearchBot) from offline training scrapers cleanly without WAF blocks.
Locked
03
Publishing /llms.txt cleanly without menu bloat
Give coding agents and retrieval bots a crisp < 2 KB Markdown reading guide under the llmstxt.org specification.
Locked
04
Answer-first pages, HTML tables & unified @graph schema
Replace poetic slogans with 25–40 word answer capsules, native HTML <table> pricing specs & connected company JSON-LD.
Locked
05
The 10-minute monthly check inside Free AEO Checker
Spot caching regressions, accidental 403 blocks, or disconnected schema nodes monthly before rankings slip.
Locked
06
Your 4-week implementation roadmap & impact benchmarks
Sequence crawler access, /llms.txt, answer-first rewrites & prompt testing across four crisp weekly passes.
Locked
INSTANT WORK-EMAIL ACCESS
Read all 6 chapters online & download the 15 pages PDF
Enter your work email once to open this playbook and every other guide in the library.
PB-01 · CHATGPT & AI VISIBILITY
Get named when buyers ask ChatGPT or Perplexity who to pick
By Cr8v Stacks Engineering · Updated Nov 2026 · 6 chapters · 15 pages
CHAPTER 01
Why buyers ask ChatGPT before they click Google
A quiet shift in B2B discovery means buyers shortlist 2–3 trusted options inside AI chat before opening a single browser tab.
Five years ago, a buyer looking for a custom website or software vendor skimmed ten links on Google. Today that initial shortlist happens inside ChatGPT Search, Perplexity, or Claude in under thirty seconds.
3 in 5 B2B buyers
Compare vendors inside AI assistants before shortlisting agencies to contact directly.
< 1.2 sec budget
Live AI fetchers give your server response before dropping slow scripts and moving to a clearer site.
30-point audit
Every factor determining whether your brand is cited or skipped fits cleanly into six measurable pillars.
Across the sites we audit inside Free AEO Checker, the gap is rarely company quality—it is basic readability:
92 / 100
cr8vstacks.com verified live grade
8 / 8
AI search bots returning HTTP 200
42%
Sites silently blocking PerplexityBot
0.9s
Server-rendered HTML delivery
What changed between classic SEO and AI recommendations
Google rewarded keyword density, backlink volume, and ten links per page. Answer engines compress twenty candidate tabs into two named recommendations and expect structured proof right inside initial HTML.
Real takeaway: A pretty homepage built only for human scrolling leaves ChatGPT Search with nothing concrete to verify or quote when your next buyer asks who to hire.
Ten Blue Links Era (2015–2023)
Buyers clicked 5–8 tabs, scanned vague hero copy, opened pricing pages, and compared agencies manually.
AI Shortlist Era (2025+)
Models fetch raw HTML once, extract exact capabilities, price ranges & case proof, and name 2–3 trusted options inline.
KEY TAKEAWAY
Being readable only to human browsers is no longer enough: if AI retrieval agents cannot extract your capabilities inside raw HTML, you never enter the shortlist.
CHAPTER 02
How bots read raw HTML & Cloudflare firewall rules
Separate live citation indexers (OAI-SearchBot, PerplexityBot, Claude-SearchBot) from offline training scrapers cleanly without WAF blocks.
Unlike classic two-wave Googlebot indexing, live AI fetchers parse your initial HTTP 200 payload directly in milliseconds—and frequently get dropped silently by Cloudflare Bot Fight Mode or JS CAPTCHAs.
Blocking GPTBot in /robots.txt protects content from offline training without hurting ChatGPT Search citations—as long as OAI-SearchBot and ChatGPT-User stay explicitly allowed.
Bot User-Agent
Vendor & Role
Trains AI?
Recommended Rule
OAI-SearchBot
OpenAI · ChatGPT Search index
No
PASS
ChatGPT-User
OpenAI · Live buyer chat fetch
No
PASS
PerplexityBot
Perplexity · Live citation index
No
PASS
Claude-SearchBot
Anthropic · Claude web search
No
PASS
GPTBot
OpenAI · Foundation model training
Yes
Policy choice
ClaudeBot
Anthropic · Model training corpus
Yes
Policy choice
The hidden Cloudflare Bot Fight Mode trap
Roughly two out of five sites we audit already allow search bots inside /robots.txt—only for Cloudflare WAF or security plugins to silently serve an HTTP 403 JavaScript challenge before a single word loads.
KEY TAKEAWAY
Confirm HTTP 200 SSR text via curl -A "OAI-SearchBot", explicitly allow search indexers inside /robots.txt, and allowlist verified crawler IPs in your WAF.
CHAPTER 03
Publishing /llms.txt cleanly without menu bloat
Give coding agents and retrieval bots a crisp < 2 KB Markdown reading guide under the llmstxt.org specification.
Proposed at llmstxt.org, /llms.txt hands AI assistants your canonical company summary, core services, case studies, and tools directly at domain root without navigation menus or cookie modals.
Myth: "`/llms.txt` ranks me #1 on ChatGPT"
False. Adding /llms.txt does not boost authority on its own or override blocked crawlers and weak service copy.
Reality: "A clean 2-KB reading map for agents"
True. Under llmstxt.org, it hands coding tools and retrieval agents your canonical URLs without nav clutter or menus.
Standard `llmstxt.org` 4-Block Order
1
2
3
4
1
# Official Brand Title (H1)Exact registered business name matching your Organization schema.
2
> Two-Sentence BlockquotePlain-English explanation of what you build, who you serve, and offices.
3
Core Services Links (`##`)Annotated [Title](https://...) links to pricing, flagship builds, and guides.
4
Proof & Tools Section (`##`)Direct links to case studies, free calculators, and verified contact/legal pages.
A clean /llms.txt file takes fifteen minutes to ship
Keep yours strictly under 80 lines, plain UTF-8 Markdown, served at HTTP 200 from your root domain. Link only evergreen canonical URLs.
Skip marketing fluff inside descriptions
Write clear 8–12 word descriptions after each colon (: Fixed-scope WordPress builds from $3,500) instead of adjectives.
Ensure your caching plugin does not wrap /llms.txt inside a themed HTML 404 or redirect chain.
cr8vstacks.com/llms.txt — Minimal Valid Reference
# Cr8v Stacks
> Custom WordPress, WooCommerce, and AI-ready web engineering studio.
## Core Services
- [Custom Websites](https://cr8vstacks.com/services/): Fixed-scope web builds.
- [Free AEO Checker](https://cr8vstacks.com/toolkits/free-aeo-checker/): Live 30-point audit.
- [Field Guides](https://cr8vstacks.com/guides/): Playbooks for founders and leads.
KEY TAKEAWAY
Keep /llms.txt focused under 80 clean lines, served as UTF-8 Markdown, linking only high-signal pages.
CHAPTER 04
Answer-first pages, HTML tables & unified @graph schema
Replace poetic slogans with 25–40 word answer capsules, native HTML <table> pricing specs & connected company JSON-LD.
Retrieval models lift passage chunks anchored around clear headings and validate claims against connected Organization, Service, and off-site sameAs citations.
4 -> 1 JSON-LD
Replacing fragmented plugin snippets with one server-rendered @graph cuts ambiguity immediately.
5+ SameAs Links
Matching LinkedIn, GitHub, Clutch, Google Business Profile & domain records proves off-site authenticity.
One `@graph`, zero plugin clashes
Many WordPress sites emit 3 separate Organization islands from SEO, review, and local plugins. Merging under @graph creates one undeniable source of truth.
Deterministic `@id` anchors
Link Service -> provider: {"@id": "/#organization"} and WebPage -> publisher so models resolve your entity graph in a single pass.
Off-site provenance: models double-check claims outside your domain
Clean schema on your site only finishes half the job. Retrieval models cross-verify your claims against directories, legal policies, and public social profiles.
Consistent entity facts everywhere
Keep company name, core services, named leadership, and physical office addresses (for example Lagos & Ibadan) word-consistent across your footer, Privacy Policy, LinkedIn, and Google Business Profile.
Include verified review & case-study URLs inside `sameAs`
Explicitly connecting verified Clutch, GitHub, and YouTube handles helps answer engines tie third-party praise back to your domain.
Every top service H2 opens with a 25–40 word verbatim answer capsule
Organization, Service, and FAQPage linked inside one clean @graph with zero schema warnings
/llms.txt updated with new flagship work, case studies, and interactive utilities
KEY TAKEAWAY
Grade-A AI citations compound when technical crawlability and answer clarity stay verified every 30 days.
CHAPTER 06
Your 4-week implementation roadmap & impact benchmarks
Sequence crawler access, /llms.txt, answer-first rewrites & prompt testing across four crisp weekly passes.
You do not need a website rebuild to win AI visibility. Most teams ship the full stack cleanly in four weeks—starting with access checks in Week 1 and finishing with live prompt tests in Week 4.
Where four weeks of effort delivers the highest return
You get 80% of the citation lift from fixing crawl access, surfacing clear pricing/service tables in HTML, and unifying company identity.
Do not start with content volume alone. Fix crawler access and server readability first—then every clear page you publish enters the index immediately.
Impact on AI Shortlist Inclusion vs. Team Effort
Based on Cr8v Stacks live diagnostics & before/after client re-crawls.
Fixing robots.txt + Cloudflare WAF
96% liftDefault: Day 1 fix
Moving Key Copy Into SSR HTML
92% liftDefault: JS bundle
25-Word Leads + HTML Spec Tables
88% liftDefault: Vague copy
Unified @graph JSON-LD + /llms.txt
85% liftDefault: Fragmented
KEY TAKEAWAY
Always fix technical access (robots.txt, WAF, SSR HTML) before editing headlines so every improvement actually reaches crawlers.