D:\Hermes\Kavenacc (44 HTML pages, 31 blog articles, llms.txt, robots.txt, sitemap.xml), cross-referenced against the empirical findings in the awesome-aeo-seo skill (citation drivers, retrieval mechanics, measurement). Live SERP/AEO visibility was not measured (no web tooling). Recommendations are ranked by the reference's evidence hierarchy: prompt-content alignment > earned media > structure; schema/Core Web Vitals show no independent citation signal per the cited studies.
0. Headline verdict
llms.txt, an AI-crawler-specific robots.txt, FAQ schema, BlogPosting schema, and 31 topical articles. The single biggest gap is earned media: per the Discovered Labs 2M-point study, third-party authoritative mentions dominate citations far more than owned content or schema. The site is strong on the factors it can control and weak on the one that matters most for AI visibility.1. What already works (keep doing this)
| Signal | Evidence found | Why it matters (reference) |
|---|---|---|
| llms.txt | Present, detailed: disambiguation of old vs current positioning, packages + prices, founder facts, verified profiles. | Direct machine-readable context for LLMs — a leading AEO practice; surfaces accurate facts and kills stale "web design agency" descriptions. |
| AI-crawler robots.txt | Named groups for GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot, Applebot + users, all Allow:/ (only /admin/, /api/ blocked). | Signals to answer engines they are explicitly welcome — removes the #1 accidental exclusion risk. |
| Question-style content | Blog titles & H2s are phrased as queries: "WhatsApp AI Agent… Pricing?", "How much does a WhatsApp AI agent cost in Kenya?", "Why You Must Own Your Code". | Prompt-content alignment is the dominant citation signal (β=+0.37, ~3× next). Writing the way a model frames a query is exactly what the research rewards. |
| FAQ / Question schema | index.html wraps 4 Q&A pairs in FAQPage; 30 BlogPosting, 12 Question/Answer nodes across site. | The reference flags FAQ/TLDR blocks as positive citation signals; schema helps engines parse the Q&A structure. |
| Topical coverage breadth | 31 articles spanning salons, clinics, real estate, logistics, ecommerce, schools, consultants — vertical-specific intent. | Retrieval consistency across methods + topic relevance dominate citations; broad relevant coverage raises the odds of matching a query's natural framing. |
| Canonical + OG tags | canonical present on index & services; full OpenGraph set. | Prevents duplicate-content dilution; clean social/AEO snippet rendering. |
2. Gaps & fixes (ranked by impact)
🔴 Gap 1 — Earned media is near-zero (highest-impact)
Evidence: All 31 blog links point internally (index, sibling articles). No outbound citations to authoritative third parties; no evidence of being cited by others.
Reference: Discovered Labs 2M-point model — prompt-content alignment is the top signal, but earned/third-party mentions far outperform owned content across ChatGPT/Perplexity/Gemini. The "Generative Engine Optimization at Scale" study (100K+ responses) shows best-of listicles ≈21% of citations and YouTube/Reddit dominate non-corporate sources.
Fix: (a) Get listed/cited on Kenyan tech press, Clutch/GoodFirms-style directories, and AI-roundup newsletters. (b) Publish the llms.txt claims as a press release / guest post so third parties repeat them. (c) Build a "As featured in / cited by" strip — but only after real earned mentions exist (no fake badges; your own audit flags inflated proof).
🟠 Gap 2 — FAQ schema on blog articles is inconsistent
Evidence: index.html has proper FAQPage wrapper; grep shows 0 FAQPage in blog/*.html even though many contain Question/Answer nodes. Likely the blog Q&A is not wrapped in a FAQPage item (or uses a different container).
Fix: Wrap each article's Q&A block in a single FAQPage → ItemList of Question/Answer. This is the single highest-ROI structural fix (FAQ blocks are a positive signal per the reference).
🟠 Gap 3 — Reciprocal internal linking is thin / no hub structure
Evidence: internal links exist but are flat (articles link to index + siblings). No clear pillar-page → cluster hub that concentrates authority on high-value terms like "AI agent Kenya", "custom AI software", "WhatsApp AI agent pricing".
Fix: Designate 3 pillar pages (Services, AI-Agent-Launchpad, OpsCommand) and make every relevant blog article link UP to its pillar with descriptive anchor text. This also improves the "SEO is the floor" foundation the Discovery Gap study says predicts LLM visibility.
🟡 Gap 4 — No visible author/entity authority signals on-page
Evidence: Person + Organization schema exist (30 each) — good — but the blog articles don't visibly showcase founder credentials (26 yrs IT, 50+ apps, 2 books, Amazon author) in a way that reads as an authority marker to both users and models.
Reference: "Authority Signals in AI Cited Health Sources" generalizes: author credentials + institutional affiliation act as trust proxies for LLMs. For a services business, founder authority is the proxy.
Fix: Add a consistent author byline block (with Person schema already present) citing the books/Amazon and years-of-experience on every article.
🟡 Gap 5 — Measurement is absent
Evidence: No AI-visibility tracking (no brand-mention monitoring across ChatGPT/Perplexity/Gemini).
Reference: "Quantifying Uncertainty in AI Visibility" — citation distributions follow power laws; single-run checks are statistically unreliable. Need bootstrap CIs + minimum sample sizes.
Fix: Run a 30–50 query monthly brand-visibility prompt across the 3 engines; track with confidence intervals. (Pairs with your competitor-news-monitor / researchmatt skills.)
🟡 Gap 6 — Minor: image alt + Core Web Vitals
Evidence: Only 7 alt= attributes on index.html; ImageObject schema present (60) but alt text thin. (Per reference, Core Web Vitals show no independent citation signal — so this is a classic-SEO win, not an AEO one.)
Fix: Add descriptive alt text to hero/images for accessibility + classic SEO; don't over-invest expecting AEO lift.
3. Priority action plan
| # | Action | Effort | AEO impact | Use this skill |
|---|---|---|---|---|
| 1 | Earn 5–10 real third-party mentions (press, directories, guest posts) | High | HIGHEST | client-lead-research, researchmatt |
| 2 | Wrap all blog Q&A in FAQPage schema | Low | HIGH | ai-seo, onpage-seo |
| 3 | Build pillar→cluster internal linking | Med | HIGH | programmatic-seo, seo |
| 4 | Add founder-authority byline to every article | Low | MED | ai-seo |
| 5 | Stand up monthly AI-visibility measurement (3 engines, CIs) | Med | MED | researchmatt, competitor-news-monitor |
| 6 | Alt-text pass on images | Low | LOW (SEO) | onpage-seo |
llms.txt disambiguation — it directly fights the stale "web design agency" positioning that would otherwise get cited by models.