Executive Summary

Faddom’s AI Agent Readiness score is 52/100 (UX 64, Discovery 34), which indicates moderate human usability but weak machine discoverability and agent interoperability. The most immediate issues are in Elementor-generated interactive markup (non-semantic clickable elements, missing ARIA/tabindex, unlabeled form controls) and missing AI discovery standards (llms-full.txt, markdown negotiation, .well-known manifests).

Business risk is twofold: (1) AI assistants and crawlers will miss or misinterpret key conversion actions (demo/contact/report CTAs), reducing qualified AI-sourced traffic and lead capture; and (2) absence of machine-readable capability metadata means Faddom will be underrepresented in agent-mediated workflows compared with competitors that publish these standards.

Against current enterprise baselines, a score in the low-50s is below “agent-ready” expectations (typically 70+), especially for B2B SaaS brands that rely on high-intent discovery and self-serve evaluation.

---

Audit Results

1) Semantic Interactive Elements (Warn, score 74)

Metrics: 181 non-semantic clickables; 163 without ARIA/tabindex.
Examples: span.elementor-button-content-wrapper, span.elementor-button-text, div.elementor-element.elementor-element-b2dc2f1.

Business impact: AI agents and assistive tech may not reliably identify CTAs as actionable, reducing successful completion of “request demo,” “access report,” and navigation flows.
Severity: P1
Affected AI user flows: CTA click-through, menu interactions, modal close controls, form initiation.
WordPress/Elementor notes: This pattern is common when Elementor widgets or custom JS bind clicks to wrapper <div>/<span> nodes. Fixes should be done in Elementor widget settings first, then child-theme/JS fallbacks for legacy templates.

---

2) ARIA Roles on Non-Semantic Elements (Fail, score 10)

Metrics: 181 clickable non-semantic elements; only 18 have role+tabindex; 163 missing ARIA behavior.

Business impact: Agents parsing DOM semantics cannot infer intent of controls, causing navigation failure and lower confidence in autonomous task execution.
Severity: P0
Affected AI user flows: Header nav expansion, dismiss controls (close-bar), in-page toggles.
WordPress/Elementor notes: Elementor Nav Menu and Icon widgets can emit proper <button> in many cases, but custom icon wrappers often remain <div>. Resolve via widget replacement and controlled DOM post-processing in a child theme.

---

3) Label–Input Associations (Warn, score 63)

Metrics: 8 inputs total; 3 unlabeled (#form-field-first_name, #form-field-email, #form-field-description).

Business impact: Form completion reliability drops for both accessibility tooling and AI form-fill agents; placeholders are not stable accessible names and may disappear on input.
Severity: P1
Affected AI user flows: Contact/demo form auto-fill and submission.
WordPress/Elementor notes: Elementor Form widget supports visible labels and custom labels per field. This should be fixed in the Form widget panel, not by ad hoc frontend scripts.

---

4) Tap Target Size (Warn, score 73)

Metrics: 21/78 visible interactives below 24×24; examples include button#e-n-menu-dropdown-icon-1631 and -1632 at 13×26, and tiny skip-link rendering.

Business impact: Small targets reduce successful activation by touch users and pixel-coordinate agents, increasing misclicks and abandonment in mobile nav flows.
Severity: P2
Affected AI user flows: Mobile menu expansion, skip-link activation, secondary utility controls.
WordPress/Elementor notes: Address in Elementor global button/nav spacing and theme CSS overrides; avoid per-page inline tweaks.

---

5) robots.txt — AI Bot Rules (Warn, score 29)

Metrics: Explicit rules for 2/7 bots; missing GPTBot, ClaudeBot, anthropic-ai, Google-Extended, CCBot.

Business impact: Policy ambiguity leads to inconsistent crawling behavior across AI platforms and weak governance over content usage.
Severity: P1
Affected AI user flows: Content ingestion and citation by AI assistants.
WordPress/Kinsta/Cloudflare notes: On Kinsta + Cloudflare, set robots centrally and avoid plugin conflicts (Yoast/RankMath virtual robots vs physical robots.txt).

---

6) llms.txt Structure (Warn, score 60)

Metrics: file exists, but missing blockquote summary, sections, and markdown links.

Business impact: Agents find the endpoint but cannot extract prioritized canonical resources; discoverability remains low despite file presence.
Severity: P1
Affected AI user flows: Knowledge grounding and source selection.
WordPress notes: Generate from WP content model (key pages, docs, legal, product) via scheduled task/plugin route.

---

7) llms-full.txt Missing (Fail, score 0)

Metrics: 404 at /llms-full.txt.

Business impact: Agents must crawl many pages instead of one corpus file, lowering ingestion completeness and freshness.
Severity: P2
Affected AI user flows: Bulk knowledge import, long-context retrieval.
WordPress notes: Best implemented as generated static artifact (WP-CLI cron job) cached via Cloudflare.

---

8) Markdown Content Negotiation Missing (Fail, score 0)

Metrics: Accept: text/markdown returns HTML.

Business impact: Prevents efficient machine consumption of page content and increases parsing errors.
Severity: P2
Affected AI user flows: Direct fetch-and-summarize of canonical pages.
Kinsta/Cloudflare/WordPress notes: Implement rewrite/endpoint strategy (e.g., ?format=md or .md routes) instead of true HTTP conneg if hosting stack limitations apply.

---

9) Missing .well-known Agent Capability Files (Fail: MCP card, agent-skills, OAuth discovery, API catalog, WebMCP signals)

Business impact: Agents cannot discover tool/action interfaces programmatically, reducing ability to execute high-intent tasks (query systems, authenticated operations, workflow invocation).
Severity: P2 now, P1 strategic for 2026 readiness.
Affected AI user flows: Tool invocation, authenticated integration, capability discovery.
Stack notes: Add static JSON files via WordPress rewrite/static delivery; if product APIs exist off-domain, publish pointers at root domain.

---

Prioritized Action Plan

1. Fix non-semantic clickables + missing ARIA/tabindex in Elementor templates

   • Priority: P0
   • Effort: M
   • Score impact: +10 to +15
   • Replace clickable wrapper div/span patterns with real <a>/<button> widgets; patch remaining legacy nodes via child-theme JS/PHP hooks.

2. Repair form labeling in Elementor Form widgets

   • Priority: P1
   • Effort: S
   • Score impact: +4 to +6
   • Enable visible labels and explicit field labels for name/email/description; keep placeholders as helper text only.

3. Standardize mobile tap target sizes in nav and utility controls

   • Priority: P1
   • Effort: S
   • Score impact: +3 to +5
   • Enforce min 24×24 (44×44 for primary CTAs) through Elementor global styles + child-theme CSS.

4. Publish explicit AI crawler directives in robots.txt

   • Priority: P1
   • Effort: S
   • Score impact: +4 to +6
   • Add intentional Allow/Disallow for GPTBot, ClaudeBot, anthropic-ai, Google-Extended, CCBot.

5. Rebuild llms.txt to spec + add llms-full.txt

   • Priority: P1/P2
   • Effort: M
   • Score impact: +8 to +12
   • Include H1, summary blockquote, H2 sections, markdown links; generate full corpus export nightly.

6. Introduce markdown endpoint strategy for key pages

   • Priority: P2
   • Effort: M
   • Score impact: +4 to +7
   • Serve markdown at predictable paths (/page.md or ?format=md) with correct Content-Type.

7. Publish .well-known capability manifests (mcp, agent-skills, optional OAuth/API catalog)

   • Priority: P2 (strategic)
   • Effort: M/L
   • Score impact: +6 to +10
   • Start with static JSON metadata; expand as product APIs/tooling mature.

---

Implementation Examples

A) Elementor/WordPress: fix clickable non-semantic elements

// wp-content/themes/your-child-theme/functions.php
add_action('wp_footer', function () { ?>
<script>
document.querySelectorAll('div.close-bar.elementor-widget-icon, span.elementor-button-content-wrapper')
  .forEach(el => {
    if (!el.hasAttribute('role')) el.setAttribute('role','button');
    if (!el.hasAttribute('tabindex')) el.setAttribute('tabindex','0');
    el.addEventListener('keydown', (e) => {
      if (e.key === 'Enter' || e.key === ' ') { e.preventDefault(); el.click(); }
    });
  });
</script>
<?php });

Elementor-side (preferred):

• Edit affected widget → Advanced → HTML Tag: set to button or a where possible.
• For icon close control (.elementor-element-b2dc2f1), replace Icon widget click handler with Button widget styled as icon-only button.

B) Elementor Form labels

In Elementor Form widget:

• Field settings for first_name, email, description:
  • Label: set explicit text
  • Show Label: Yes
  • Keep placeholder optional.

If forced visually hidden labels:

.elementor-form .screen-reader-text{
  position:absolute!important; width:1px; height:1px; overflow:hidden;
  clip:rect(1px,1px,1px,1px); white-space:nowrap;
}

C) Tap target fixes

/* Child theme style.css */
button.e-n-menu-dropdown-icon,
a.ea11y-skip-to-content-link {
  min-width: 24px;
  min-height: 24px;
  padding: 8px;
}
.elementor-button, .elementor-icon {
  min-width: 44px; min-height: 44px;
}

D) robots.txt explicit AI rules

User-agent: GPTBot
Allow: /

User-agent: ClaudeBot
Allow: /

User-agent: anthropic-ai
Allow: /

User-agent: Google-Extended
Allow: /

User-agent: CCBot
Allow: /

User-agent: ChatGPT-User
Allow: /

User-agent: PerplexityBot
Allow: /

(Adjust to policy. Keep one canonical source; avoid duplicate plugin-generated directives.)

E) llms.txt and llms-full.txt

/llms.txt example:

# Faddom AI Access Guide

> Canonical resources for understanding Faddom product capabilities, deployment, pricing, and contact workflows.

## Product
- [Platform Overview](https://faddom.com/)
- [Use Cases](https://faddom.com/use-cases/)
- [Pricing](https://faddom.com/pricing/)

## Learn
- [Blog](https://faddom.com/blog/)
- [Documentation](https://docs.faddom.com/)

Generate /llms-full.txt via WP-CLI export job and serve as cached static text.

F) .well-known/mcp.json starter

{
  "name": "Faddom",
  "description": "Discovery metadata for Faddom agent-invokable capabilities.",
  "url": "https://faddom.com",
  "tools": []
}

Place under web root .well-known/ via deploy pipeline or MU-plugin route.

---

Verification

1. Re-run the same AI Agent Readiness audit after deployment to production (not staging blocked by robots/auth).
2. Confirm expected metric deltas:
   • aria-roles: missingAria from 163 → near 0
   • label-input: unlabeled from 3 → 0
   • tap-target-size: tooSmall from 21 → <5
   • robots-ai-bots: botsWithRules from 2 → 7
   • llms-txt: hasSummary true, sections >=1, linkCount >0
   • llms-full.txt: HTTP 200
   • markdown endpoint: returns text/markdown on designated route
   • .well-known/*: HTTP 200 + valid JSON.
3. Validate with manual checks:
   • Keyboard-only navigation (Tab/Enter/Space) on nav, CTA, modal close.
   • Screen reader name announcement for the three form controls.
   • Mobile viewport tap testing for dropdown icons.
4. Track business outcomes for 30 days:
   • AI referral traffic volume
   • Form completion rate on AI-referred sessions
   • Crawl/log visibility for declared bots.