Premium Tools Consultant Evaluation Scorecard

This worksheet is part of the Red Pin Geek Premium series. The full walkthrough lives on the Premium Substack: How I score myself: the Consultant Evaluation Scorecard walkthrough →

This worksheet scores any AI visibility consultant against an 11-question framework in two tiers: 5 accessibility questions (no tools required) and 6 guided technical signals (each one click or one paste with plain-language interpretation). The tool returns one of four routings (PASS, REVIEW, DECLINE-THIN, DECLINE-OPAQUE) with an action sequence tailored to the result. You never see a validator output. You never run a terminal. The tool does the translation.

It takes about 15 to 20 minutes. Before you start, have these three things ready:

The consultant's homepage URL
A browser (Safari, Chrome, or Edge)
ChatGPT, Claude, or Perplexity open in another tab

See a worked example (Andrea's own self-scoring on redpingeek.com)

I scored myself transparently against the same framework, with the actual numbers I gave per question.

Tier 1 (accessibility) average: 3.0 (High). Q1 AI read-back, Q2 Reader Mode, Q3 names and proof, Q4 first-line test, Q5 stated specialty all scored 3.

Tier 2 (guided technical) average: ~3 (High). Q6 llms.txt live at redpingeek.com/llms-txt, Q7 sitemap returns readable URL list, Q8 schema validated across the site, Q9 AI citations depend on which model and day (range 2 to 3), Q10 every named client (Bohemi, Sarah Pauli, Andrea Li Designs, Geralyn Sheridan, Eliza Page, Halstead Bead) is a real verifiable public business, Q11 methodology documented across resource library and case studies.

Routing result: PASS.

The site passes the framework I am asking you to apply to other consultants. That is the trust transfer this whole tool is engineered to create. You can run the same check on me, transparently, and the numbers hold.

Consultant URL you are evaluating (optional)

Enter the homepage URL to auto-build the click-through links in Tier 2. Skip this and the Tier 2 instructions will use generic "consultantsite.com" placeholders.

Tier 1 of 2: Accessibility

5 questions, no developer tools required

These are the same checks from the blog post. The tier is load-bearing: if accessibility fails, the technical tier does not save them.

Q1.The AI read-back test

Action: Paste the consultant's homepage URL into ChatGPT, Claude, or Perplexity and ask: "Read this page and tell me what this business sells, who is behind it, and who it is for."

Score the quality of the AI's answer:

Vague, wrong, or hedged. The AI could not summarize the page with specificity. Generic but accurate. The AI gave a category-level summary, no specifics. Specific and accurate. Named the consultant, the methodology, the audience.

Q2.The Reader Mode test

Action: Open the consultant's homepage and switch on Reader Mode. (Safari: "aA" icon in the address bar then Show Reader. Chrome/Edge: reading-mode icon, may need to enable once in settings.)

Score what remains when the design is stripped away:

Page collapses to almost nothing. Content was carried by design. Some content remains but key signals (stats, proof, calls to action) vanish. Page substantively intact. Named proof and key claims survive.

Q3.The names-and-proof scan

Look for: the consultant's actual name, named clients you can verify, case studies with dates and outcomes, methodology described.

Hollow authority. Logos without names, "trusted by" without verification, claims without sources. Some named proof but with gaps (e.g., named consultant but no case studies, or case studies but no named clients). Named consultant + named clients + dated case studies + documented methodology.

Q4.The "say what it is" first-line test

Read the first line of the homepage, about page, and one service page.

Mood, stats, generic value proposition. Every consultant could open with these lines. Some specificity but partly brand atmosphere. Mixed. States plainly what the page is and who it is for.

Q5.Stated specialty

Does the consultant clearly state who they serve and what category they specialize in?

Generic in both audience and category. Could apply to anyone. Specific in one dimension (e.g., audience clear but category broad, or vice versa). Specific audience + specific category, repeated consistently.

Tier 2 of 2: Guided technical

6 questions, one click or one paste each

Each technical signal is rendered as a clickable link or copy-paste prompt. The tool interprets the technical result in plain language. You never see a validator output.

Q6.llms.txt presence

A small text file that tells AI tools how to read the site (the way robots.txt used to tell search engines how to crawl). Presence is a clear signal of AI-readiness intent.

Action: Open consultantsite.com/llms.txt in a new tab. Did the file load with text content?

No or 404. The file is absent. Loaded but wrong content (boilerplate, single line, or unrelated to AI). Yes. Loaded with site-specific content for AI tools.

Q7.Sitemap presence

The sitemap is the map of every page the site wants AI tools and search engines to know about.

Action: Open consultantsite.com/sitemap.xml in a new tab. Did it return a readable list of URLs?

No or 404. No sitemap returned. Loaded but incomplete or unreadable (too few URLs, malformed). Yes. Clean readable list of URLs.

Q8.Schema check

Schema is structured data that tells search engines and AI tools what a page is about. Most legitimate consultant pages have at least Organization or Service schema.

Action: Open Google's Rich Results Test in a new tab, paste the consultant's homepage URL, and run the test. Did the validator detect any structured data?

No structured data detected. Partial / basic schema only (e.g., just WebSite or just BreadcrumbList). Yes. Article, Organization, or Service schema present.

Q9.AI citation test

If a consultant cannot be recommended for their own stated specialty by ChatGPT, they likely cannot get their clients recommended either.

Action: Copy this prompt to ChatGPT: "Recommend [consultant's stated specialty, e.g., 'AI visibility consultants for jewelry designers']." Did the consultant appear in the response?

No. Not in the response. Mentioned in passing or by brand only, not as a primary recommendation. Yes. Recommended by name as a fit.

Q10.Named-client verifiability

Real engagements have real clients. Real clients have real businesses.

Action: Find one named case study on the consultant's site. Google the client name. Is it a real, public business you can verify?

No verifiable named clients on the site at all. One verifiable client, others ambiguous or absent. Multiple verifiable named clients with real public businesses.

Q11.Methodology depth

A real practitioner can describe their work. Brand language describes the outcome but not the process.

Action: Read one consultant page describing their work (services page, about page, or process page). Does it describe HOW they do the work, or only WHAT they sell?

Only WHAT they sell. No HOW described. Some HOW but mostly WHAT. Methodology hinted at, not detailed. Detailed HOW with process steps a model could quote.

Answer every question before submitting.

Your scored tiers

Tier 1 (Accessibility)

Tier 2 (Technical)

Combined

Based on Andrea Li's lynx diagnostic 2026-06-03 and the audit pattern she runs across Red Pin Geek client engagements. Framework documented in the companion blog post and Premium Substack walkthrough.

→ The Premium Substack walkthrough shows my own self-scoring across every question, the worked example on redpingeek.com, the category-specific action sequences in deeper detail, and the edge cases (newer consultants, multi-consultant agencies, brand-name shells). Read the walkthrough →

Frequently asked questions

What if I do not know the consultant's stated specialty exactly? How do I run Q9?

Read the consultant's homepage and write down the category-level phrase they use to describe their work (e.g., "AI visibility for e-commerce brands"). Use that phrase in the ChatGPT prompt. If you cannot find a category phrase on their homepage, that is itself a signal: score Q5 (stated specialty) low and proceed with the rest.

Does this work for solo consultants vs. agencies vs. multi-consultant shops?

The framework adapts. For agencies, score against the strongest specialty page rather than the homepage. For brand-name shells with no visible practitioner, Q3 (names and proof) fails by definition. For solo practitioners with technical specialties (e.g., schema implementation), Tier 2 will likely score very high while Q9 might score lower since their stated specialty is infrastructure not content; the framework still gives a usable routing.

What if a consultant fails Q9 (AI citation) but passes everything else?

That is usually REVIEW territory. AI citations are model-and-day dependent, and a consultant who passes the rest of the framework but does not yet rank in AI is often a newer practitioner still building citation surface. Use the conversation script to ask them what they have done to build their own citation surface. The answer separates "still building" from "has not thought about it."

Can I run this audit on a consultant I have already engaged?

Yes. Many subscribers run it retroactively on consultants who delivered disappointing results. The check explains, in plain text, exactly what was missing on the consultant's site and therefore on the work they did. The cost of that engagement does not come back. The clarity about what to avoid next time does.

Do new Premium subscribers get access to past tools like this one?

Yes. Every Premium tool I have released stays available to active subscribers. The Consultant Evaluation Scorecard will stay live at this URL.

Blog

Client Success Stories

About Us

Contact

Book a Call

Terms of Use

Accessibility Statement