See the simple version before the technical detail
This is a real PageLens AI audit for https://www.experi.co.uk/. Start with the plain-English verdict, then open the full evidence when your builder needs it.
Export the report as Markdown and hand the fixes to Claude, Cursor, or Codex.
90 / 100
Excellent
No critical fixes
Priority verdict
Your site is in good shape.
90/100 - excellent, but not finished. The score is strong; the value is knowing what to fix first.
The biggest issue is heavy third-party load (506 kb from 5 domain(s)). Start with the priority fixes, then use the evidence and technical details when a developer needs proof.
Excellent
3 priority fixes identified
Score
90 / 100 — Excellent
No major launch blockers found.
Biggest risk
Heavy third-party load (506 KB from 5 domain(s))
speed and loading experience
Fastest win
Heavy third-party load (506 KB from 5 domain(s))
A practical first fix for your builder or AI agent.
Estimated impact
Higher confidence and discoverability
Fix the priority items first, then re-scan to confirm the evidence.
Biggest area to improve: Performance, which mostly means speed and loading experience.
Priority verdict
This site is technically inspectable, but the priority is making the next fix obvious.
Fix the highest-impact issue first, then use structured evidence and developer prompts so search engines, AI answer engines, and real visitors can understand the site more reliably.
What's working
- Accessibility
- AI Search
- Design
What needs attention
- Heavy third-party load (506 KB from 5 domain(s))
- No cookie consent banner detected
- Heavy third-party load (538 KB from 6 domain(s))
What to do first
- 1.Heavy third-party load (506 KB from 5 domain(s))
- 2.No cookie consent banner detected
- 3.Heavy third-party load (538 KB from 6 domain(s))
Read the consultant narrative
Experi maintains an exceptional technical foundation characterized by elite Core Web Vitals and a sophisticated SEO schema implementation.
The website demonstrates high-tier performance engineering, specifically on desktop where the Largest Contentful Paint (LCP) reaches a remarkable 416ms. The design utilizes a clean visual hierarchy and intentional whitespace to guide users toward high-contrast primary calls to action. Furthermore, the site is highly optimized for search engines and AI crawlers through comprehensive JSON-LD structured data and a functional sitemap.
The most critical technical weakness is a significant security vulnerability regarding email authentication. The domain lacks properly configured SPF, DMARC, and DKIM records, leaving Experi highly susceptible to domain spoofing and phishing attacks. Additionally, both desktop and mobile versions suffer from an excessive DOM size of 1,573 nodes and heavy third-party script loads exceeding 530 KB, which threatens long-term scalability and interaction responsiveness.
There is a major opportunity to improve brand authority and search visibility by refining the /llms.txt file to meet full specification standards. Correcting the current structural omissions will ensure AI-driven discovery tools accurately parse and represent the product. Simultaneously, tightening the DOM structure and optimizing third-party script execution will safeguard the current "Good" performance ratings as the site grows in complexity.
To secure the brand and stabilize the technical core, the following must be addressed within the next 30 days:
- Implement and validate SPF, DKIM, and DMARC records to prevent domain spoofing.
- Audit and consolidate third-party scripts to reduce total payload and render-blocking overhead.
- Refactor the DOM structure to reduce node count below the 1,500-node threshold.
Scan details
- Scanned from
- London
- Device
- Desktop + Mobile
- Pages checked
- 3
- Scan date
- 24 Apr 2026
Checks performed: SEO, UX, accessibility, performance, security, AI search.
Fix workflow
Treat this report as a queue: send the ready fixes, accept any intentional risks, then re-scan the production URL after changes land.
11
Ready to send
93
Not started
0
Accepted risk
0
Re-scan queued
Top 3 things to fix
The full report has all the proof. This is the owner-friendly version of what to do first.
Heavy third-party load (506 KB from 5 domain(s))
Plain-English reason
This affects speed and loading experience. PageLens marked it as worth fixing so you know where it belongs in the queue.
What to ask your builder or AI agent to do
Audit every third-party tag: drop ones you're no longer measuring, switch from <script> to async/defer, route analytics through a single tag manager, and self-host fonts as woff2 (most font CDNs add 50-150 KB per family).
Show the audit wording
Third-party scripts (analytics, embeds, ad pixels, font CDNs, chat widgets) are hosted outside your control and often render-block, INP-block, or both. Each adds DNS resolution + TLS overhead and can fail independently of your own infrastructure.
No cookie consent banner detected
Plain-English reason
This affects tracking & analytics. PageLens marked it as worth fixing so you know where it belongs in the queue.
What to ask your builder or AI agent to do
Implement a consent management platform (Cookiebot, OneTrust, Osano, Termly, or the free Klaro). Configure it to block tracking scripts until the visitor opts in, and enable Google Consent Mode v2 for recovery of modelled conversions.
Show the audit wording
No consent management platform (CMP) was found on this page. Under GDPR, UK PECR, and ePrivacy rules, setting non-essential cookies or firing tracking pixels without explicit consent exposes you to regulatory fines of up to 4% of annual revenue. Beyond compliance, missing consent also means you cannot use Google's Consent Mode — which recovers up to 70% of lost conversion signal through privacy-safe modelling.
Heavy third-party load (538 KB from 6 domain(s))
Plain-English reason
This affects speed and loading experience. PageLens marked it as worth fixing so you know where it belongs in the queue.
What to ask your builder or AI agent to do
Audit every third-party tag: drop ones you're no longer measuring, switch from <script> to async/defer, route analytics through a single tag manager, and self-host fonts as woff2 (most font CDNs add 50-150 KB per family).
Show the audit wording
Third-party scripts (analytics, embeds, ad pixels, font CDNs, chat widgets) are hosted outside your control and often render-block, INP-block, or both. Each adds DNS resolution + TLS overhead and can fail independently of your own infrastructure.
Recommended first 30 days
A practical roadmap for turning the audit into progress.
Today
Heavy third-party load (506 KB from 5 domain(s))
Audit every third-party tag: drop ones you're no longer measuring, switch from <script> to async/defer, route analytics through a single tag manager, and self-host fonts as woff2 (most font CDNs add 50-150 KB per family).
This week
No cookie consent banner detected
Implement a consent management platform (Cookiebot, OneTrust, Osano, Termly, or the free Klaro). Configure it to block tracking scripts until the visitor opts in, and enable Google Consent Mode v2 for recovery of modelled conversions.
This month
Heavy third-party load (538 KB from 6 domain(s))
Audit every third-party tag: drop ones you're no longer measuring, switch from <script> to async/defer, route analytics through a single tag manager, and self-host fonts as woff2 (most font CDNs add 50-150 KB per family).
Your action plan
Same data as the full report, grouped by what a non-technical owner should do with it.
Fix first
High-impact fixes that should usually be tackled before anything else.
Nothing in this bucket.
Plan next
Important fixes that may need more development time or a design decision.
Nothing in this bucket.
Can wait
Polish and lower-priority work. Useful, but not where to start.
- Heavy third-party load (506 KB from 5 domain(s))
- No cookie consent banner detected
- Heavy third-party load (538 KB from 6 domain(s))
Need the detail?
The full report still has every finding, evidence, rule ID, filters, screenshots, and technical panels.
This is a real audit — not a mockup. Yours will look like this.
No subscription · Refund if we find nothing actionable
