What a corporation receives: analysis, plan and proposal with measurable targets.
Illustrative case: an industrial manufacturer based in Mexico, selling in the United States and Canada, with a three-language website. Every figure on this page is a sample and is labeled as such; in your analysis they come from your data.
buying questions for which ChatGPT, Gemini or Perplexity cited the manufacturer in the United States market. Three competitors were cited instead, one of them in 9 of the 12 answers.
Sample figure. In a real analysis it comes from the AI answers collected per question, market and language.
Baseline, six-month target and how each one is measured.
The analysis opens with this table and the monthly report closes with it. Nothing goes in the plan without a line here.
| Indicator | Baseline | Target at 6 months | How it is measured |
|---|---|---|---|
| AI citation share, United States (12 buying questions) | 0% | 25% | Monthly sample of answers in ChatGPT, Gemini, Perplexity and Google AI Overviews |
| Non-brand organic sessions, United States | 1,200 per month | 2,400 per month | GA4, organic channel, brand terms excluded |
| Quote requests (RFQRFQ, request for quotation: a buyer asking for price, lead time and terms on a specific product or service, by form, email or phone. In a corporation it is the conversion that counts; traffic without RFQs is cost.) from the website | 9 per month | 25 per month | Form with channel attribution plus the CRM |
| Cost per paid lead | 180 USD | 110 USD | Ad platforms against RFQRFQ, request for quotation: a buyer asking for price, lead time and terms on a specific product or service, by form, email or phone. In a corporation it is the conversion that counts; traffic without RFQs is cost.s attributed to paid |
| Pages readable by AI crawlers | 38% | 100% | Crawl with GPTBot, ClaudeBot and PerplexityBot user agents |
| Average position, 10 core terms, United States | 18 | 6 | Rank tracking, weekly |
Six findings that explain the baseline.
- 01
AI crawlers blocked
robots.txt disallows GPTBot and ClaudeBot on the whole site. The assistants cannot read what they are asked to recommend.
- 02
Country versions without hreflang
A visitor in the United States receives the Spanish version. Google treats the three versions as duplicates.
- 03
Catalog locked in PDF
240 technical sheets live in PDF files with no HTML page. Search and AI see 240 blanks.
- 04
The competitor is cited for its guides
The competitor cited in 9 of 12 answers publishes 42 buying guides. The manufacturer publishes none.
- 05
No structured data
No Organization, Product or FAQ markup. The site tells machines nothing about what it sells.
- 06
RFQRFQ, request for quotation: a buyer asking for price, lead time and terms on a specific product or service, by form, email or phone. In a corporation it is the conversion that counts; traffic without RFQs is cost. without attribution
The quote form records no channel. Nobody knows which market or channel produces the 9 requests.
Plan: five moves in 90 days, each with its own metric.
- 01
Weeks 1 to 2
Unblock the AI crawlers, publish llms.txt, hreflang on the three versions, Organization schema.
Metric: pages readable by AI crawlers, 38% to 100%.
- 02
Weeks 3 to 6
The 40 technical sheets with the most demand become HTML pages with Product schema; the PDF stays as download.
Metric: non-brand impressions on those 40 products.
- 03
Weeks 4 to 8
Answer center: 20 buying questions per market, written from the sales team’s real objections.
Metric: AI citations per question and market.
- 04
Weeks 6 to 10
Attribution on the quote form, the CRM and the phone line, with the AI assistant option.
Metric: RFQRFQ, request for quotation: a buyer asking for price, lead time and terms on a specific product or service, by form, email or phone. In a corporation it is the conversion that counts; traffic without RFQs is cost.s per channel and market.
- 05
Months 4 to 6
Second round by market: Canada in French, press and profiles that the assistants use as trust signals.
Metric: AI citation share per market.
Proposal: what is included, and what is quoted apart.
The proposal prices the analysis and the plan as one product, with the reading session and one adjustment to your operation. It names the owner of every move: your team, your provider or me.
What the plan asks to fix or build on the site is quoted separately, case by case, after you read the analysis. Supervising another provider is never on the table. No mandatory retainer and no ranking promises.
Conversions by channel: where the quote requests come from.
This is the table the client reads every month. The point of the whole plan is the RFQRFQ, request for quotation: a buyer asking for price, lead time and terms on a specific product or service, by form, email or phone. In a corporation it is the conversion that counts; traffic without RFQs is cost. column, not the traffic.
| Channel | RFQRFQ, request for quotation: a buyer asking for price, lead time and terms on a specific product or service, by form, email or phone. In a corporation it is the conversion that counts; traffic without RFQs is cost. per month, baseline | RFQRFQ, request for quotation: a buyer asking for price, lead time and terms on a specific product or service, by form, email or phone. In a corporation it is the conversion that counts; traffic without RFQs is cost. per month, target at 6 months |
|---|---|---|
| Organic search, non-brand | 4 | 12 |
| AI assistants (referred from ChatGPT, Perplexity, Gemini) | 0 | 5 |
| Direct and brand | 3 | 5 |
| Paid | 2 | 3 |
| Total | 9 | 25 |
This is the shape of the work. The figures of your case come from your data.
Write to me with the site, the brands and the markets. I confirm scope, fee and delivery date first, and nothing is paid until you approve them.