AI Audit by Metamode: find where your website is losing you deals

brightpathmortgages.com.au

Overview

54/ 100

Clear opportunity

“Your hero says ‘Your future, secured’ and little else. Nothing on the first screen tells a visitor you're a mortgage broker, or who you help.”

brightpathmortgages.com.au scores 54 out of 100 across four lenses, with AI & Search Visibility the weakest at 8 of 20. 12 findings came back; clearing the actionable ones would put the score at about 84.

Screenshot of brightpathmortgages.com.au

By lens

Where the score comes from

Report details

Site audited: brightpathmortgages.com.au
Audit date: 6 July 2026
Lenses scored: 4
Findings: 12
Checks passed: 15
Projected after fixes: 84 / 100

Worst first

The three findings costing you most

Hero sells on its own
Conversioncritical
The hero reads “Your future, secured. Tailored home loan solutions.” It names no borrower, no outcome, and no reason to act. A visitor who reads only this screen can't tell what you do or who you do it for.
AI crawler access
AI & Search Visibilitycritical
robots.txt blocks GPTBot, PerplexityBot. Those engines can't read the site.
Largest Contentful Paint
Page Speedcritical
Largest Contentful Paint: 4.6 s, in the red. On mobile the hero takes over four seconds to appear.

Conversion

Executive summary

Does this page turn visitors into enquiries?

Conversion is where this site is losing the most, and it's losing it in the first screen. “Your future, secured. Tailored home loan solutions.” could sit on any broker's homepage in the country, which means a first-home buyer landing here has nothing to tell them they're in the right place.

The mechanics underneath are sound: the headline is short, the copy is readable, and the buttons say what they do. The problem is what the page is saying, not how it's built.

For a mortgage broker, the homepage is doing the job a first phone call would do: establishing who you help and why you're worth the enquiry. Rate-shoppers compare on specifics, and there aren't any here yet.

How we score Conversion

20/ 35

Projected after fixes: 28 / 35

Report details

Lens: Conversion
Score: 20 / 35
Checks passed: 4 / 7
Findings: 3

Critical gaps

Hero sells on its own
critical
The hero reads “Your future, secured. Tailored home loan solutions.” It names no borrower, no outcome, and no reason to act. A visitor who reads only this screen can't tell what you do or who you do it for.
One primary CTA
major
9 different button labels compete for the click (“Apply now”, “Get a quote”, “Book a call”, “Learn more”…). None can win.
Testimonials present
major
No testimonials, reviews, or named client quotes anywhere on the page. Trust has nothing to attach to.

Current strengths

Headline under 10 words
Your headline runs 6 words: “Your future, secured. Tailored home loan solutions.”
A short headline lands at a glance, so visitors keep reading.
Fifth-grade readability
Body copy reads at about grade 8, which is comfortable for a skim-reader.
Plain copy is easy to skim, so more visitors reach your form.
CTA says what happens next
A button that names the outcome gets more clicks.
OG image present
An og:image tag is set on the page.
Shared links preview properly, so more people click through.

Priority action roadmap

3 fixes

Quick wins

Small, contained changes. Most are an afternoon or less.

Nothing in this tier.

Foundation

The structural work the rest of the fixes rest on.

P1
One primary CTA
Pick one action, make it visually dominant, demote the rest to text links.
2 hours
P1
Testimonials present
Add two named testimonials with a concrete result, ideally with a face.
A day, once you have the quotes

Strategic

Positioning and design decisions worth deciding properly.

P0
Hero sells on its own
Lead with the plain version: what you do, for whom, e.g. “Home loans for first-home buyers in Brisbane, sorted in days not weeks.”
A positioning decision, then a day of copy

Where this lands

The hero is the hinge. Fix what it claims and the proof gap behind it matters less; leave it and every other fix on this lens is decoration on a page that hasn't said what it does.

The opportunity

Naming the borrower and the outcome is a copy change, not a rebuild. The rewrite below is the shape of it, and the two testimonial and CTA fixes are an afternoon each.

7 checks ran

Detailed audit

Hero sells on its owncritical

Found

The hero reads “Your future, secured. Tailored home loan solutions.” It names no borrower, no outcome, and no reason to act. A visitor who reads only this screen can't tell what you do or who you do it for.

Costs you

A visitor who can't tell what you do in five seconds leaves. Most won't scroll to find out.

Fix

Lead with the plain version: what you do, for whom, e.g. “Home loans for first-home buyers in Brisbane, sorted in days not weeks.”

Effort

A positioning decision, then a day of copy

One primary CTAmajor

Found

9 different button labels compete for the click (“Apply now”, “Get a quote”, “Book a call”, “Learn more”…). None can win.

Costs you

When every button is a priority, visitors choose none and leave without enquiring.

Fix

Pick one action, make it visually dominant, demote the rest to text links.

Effort

2 hours

Testimonials presentmajor

Found

No testimonials, reviews, or named client quotes anywhere on the page. Trust has nothing to attach to.

Costs you

With nothing to vouch for you, a cautious borrower picks the broker who shows proof.

Fix

Add two named testimonials with a concrete result, ideally with a face.

Effort

A day, once you have the quotes

Headline under 10 wordsPassed

Found

Your headline runs 6 words: “Your future, secured. Tailored home loan solutions.”

Earns you

A short headline lands at a glance, so visitors keep reading.

Fifth-grade readabilityPassed

Found

Body copy reads at about grade 8, which is comfortable for a skim-reader.

Earns you

Plain copy is easy to skim, so more visitors reach your form.

CTA says what happens nextPassed

Result

This check ran against the page and cleared.

Earns you

A button that names the outcome gets more clicks.

OG image presentPassed

Found

An og:image tag is set on the page.

Earns you

Shared links preview properly, so more people click through.

What a one-page automated read can't see

Beyond this audit

Your actual conversion rate
This audit reads the page, not your analytics. Whether these findings are costing you 5% or 50% depends on traffic, intent, and what happens after the click.
Everything past the homepage
The scan reads one page. Your pricing page, your signup flow, and the form someone actually fills in are all unexamined.
What your visitors are trying to do
We can tell you the offer isn't clear. We can't tell you what your customers came looking for; that takes talking to them.

See every finding

3 more findings on this lens, plus every finding on the others, each with its fix. Enter your email and the whole report opens.

Trust & Credibility

Executive summary

Does this site look credible enough to trust with your money?

Nothing here reads as broken, but a lot of it reads as bought. Inter at its defaults, a purple-leaning palette, and a gradient heading are the three most common signals of a template that was filled in rather than designed.

That matters more for a broker than for most businesses: you're asking someone to hand over their financial details, and the page is the only evidence they have that you take care with things.

Trust is judged in the first few seconds and mostly unconsciously. For a services business at this price point, the site is the credential.

How we score Trust & Credibility

21/ 35

Projected after fixes: 31 / 35

Report details

Lens: Trust & Credibility
Score: 21 / 35
Checks passed: 5 / 8
Findings: 3
Didn't apply: 1

Critical gaps

Overused default font
major
“Inter” is the only typeface on the page, at defaults. It's the single most common template tell.
Gradient text
major
The page uses gradient-clipped text. It's decorative, hurts contrast, and reads as template.
Marketing buzzwords
minor
Buzzwords doing the selling: “seamless”, “tailored solutions”, “empower”. They read as filler, not proof.

Current strengths

Glassmorphism as decoration
Restraint with effects reads as craft, which is what buyers trust.
Em-dash overuse
Copy that reads as human-written builds trust.
Justified text
Even spacing keeps your copy easy to read.
Tiny body text
No body, paragraph, or root rule sets font-size below 14px.
Text people can read on a phone actually gets read.
Mobile viewport
The page fits a phone screen, where most of your visitors are.

Priority action roadmap

3 fixes

Quick wins

Small, contained changes. Most are an afternoon or less.

P1
Gradient text
Use a single solid colour; carry emphasis with weight and size.
30 minutes

Foundation

The structural work the rest of the fixes rest on.

P2
Marketing buzzwords
Say what you do in the words a customer would use.
2 hours

Strategic

Positioning and design decisions worth deciding properly.

P1
Overused default font
Pair a distinctive display face with a clean body face on a contrast axis.
A type decision, then half a day

Where this lands

The typography and palette choices are the two that carry the signal. Neither is a redesign, but both are decisions rather than tickets.

The opportunity

Dropping the gradient heading takes half an hour. A typeface with some character and a palette that isn't the AI default are the two calls worth making properly.

8 checks ran

Detailed audit

Overused default fontmajor

Found

“Inter” is the only typeface on the page, at defaults. It's the single most common template tell.

Costs you

A stock template look reads as generic. With money on the line, generic doesn't earn trust.

Fix

Pair a distinctive display face with a clean body face on a contrast axis.

Effort

A type decision, then half a day

Gradient textmajor

Found

The page uses gradient-clipped text. It's decorative, hurts contrast, and reads as template.

Costs you

Headings that are harder to read, and a generic look, cost you trust at a glance.

Fix

Use a single solid colour; carry emphasis with weight and size.

Effort

30 minutes

Marketing buzzwordsminor

Found

Buzzwords doing the selling: “seamless”, “tailored solutions”, “empower”. They read as filler, not proof.

Costs you

Buzzwords don't tell a borrower why to pick you, so they keep looking.

Fix

Say what you do in the words a customer would use.

Effort

2 hours

Glassmorphism as decorationPassed

Result

This check ran against the page and cleared.

Earns you

Restraint with effects reads as craft, which is what buyers trust.

Em-dash overusePassed

Result

This check ran against the page and cleared.

Earns you

Copy that reads as human-written builds trust.

Justified textPassed

Result

This check ran against the page and cleared.

Earns you

Even spacing keeps your copy easy to read.

Tiny body textPassed

Found

No body, paragraph, or root rule sets font-size below 14px.

Earns you

Text people can read on a phone actually gets read.

Mobile viewportPassed

Result

This check ran against the page and cleared.

Earns you

The page fits a phone screen, where most of your visitors are.

1 left out of the score

Checks that didn't apply

These ran but had nothing to measure on this page, so they were renormalised out rather than counted against you.

Image alt text

Skipped

The page carries fewer than three images, too few to judge whether alt text is handled systematically.

What a one-page automated read can't see

Beyond this audit

Measured contrast and spacing
These checks read the markup and the stylesheet. A full pass renders the page and measures real contrast ratios, tap-target sizes, and how many greys you're actually shipping.
How it holds up on a real device
Layout is judged from the source, not from a rendered screenshot at every breakpoint. Genuine responsive breakage needs the page opened, not parsed.
Brand consistency across the product
One page can look composed while the app behind the login looks like a different company built it.

See every finding

3 more findings on this lens, plus every finding on the others, each with its fix. Enter your email and the whole report opens.

Page Speed

Executive summary

Does this page load fast enough to keep visitors?

Google's mobile Lighthouse run scores this 48 out of 100. The hero takes 4.6 seconds to paint, which is past the point where most mobile visitors have already gone back to the search results.

Layout stability and first paint are both fine, so this isn't a broken build; it's a heavy hero and too much blocking script.

Speed is the only lens Google measures on your behalf and feeds into ranking directly. For a broker competing on local search, it compounds.

How we score Page Speed

5/ 10

Report details

Lens: Page Speed
Score: 5 / 10
Checks passed: 3 / 6
Findings: 2

Critical gaps

Largest Contentful Paint
critical
Largest Contentful Paint: 4.6 s, in the red. On mobile the hero takes over four seconds to appear.
Total Blocking Time
minor
Total Blocking Time: 410 ms, needs work. The main thread is busy long enough to make taps feel laggy.

Current strengths

Cumulative Layout Shift
0.04, in the green.
The page holds still as it loads, so nothing gets mis-tapped.
First Contentful Paint
1.6 s, in the green.
Something appears quickly, so the page never looks broken.
Speed Index
3.1 s, in the green.
The page looks usable sooner, so visitors keep reading.

Priority action roadmap

2 fixes

Quick wins

Small, contained changes. Most are an afternoon or less.

Nothing in this tier.

Foundation

The structural work the rest of the fixes rest on.

P0
Largest Contentful Paint
Get the hero painting under 2.5s: preload the LCP asset, serve it in a modern format at the right size, and cut render-blocking CSS/JS.
A day
P2
Total Blocking Time
Ship less JavaScript: code-split, defer non-critical scripts, and drop unused third-party tags so the main thread frees up faster.
A day

Strategic

Positioning and design decisions worth deciding properly.

Nothing in this tier.

Where this lands

Largest Contentful Paint is the one that matters here. Total Blocking Time is the supporting problem, and both point at the same hero.

The opportunity

Preloading and resizing the hero image, then deferring the non-critical scripts, is the whole job. We won't put a projected score on it: Page Speed comes from a live Lighthouse run, so it has to be measured rather than predicted.

5 checks ran

Detailed audit

Largest Contentful Paintcritical

Found

Largest Contentful Paint: 4.6 s, in the red. On mobile the hero takes over four seconds to appear.

Costs you

Every extra second of load loses visitors before the page appears, and most enquiries come from mobile.

Fix

Get the hero painting under 2.5s: preload the LCP asset, serve it in a modern format at the right size, and cut render-blocking CSS/JS.

Effort

A day

Total Blocking TimePartial · half credit

Found

Total Blocking Time: 410 ms, needs work. The main thread is busy long enough to make taps feel laggy.

Costs you

A page that feels sluggish on the first tap makes visitors doubt the business behind it.

Fix

Ship less JavaScript: code-split, defer non-critical scripts, and drop unused third-party tags so the main thread frees up faster.

Effort

A day

Cumulative Layout ShiftPassed

Found

0.04, in the green.

Earns you

The page holds still as it loads, so nothing gets mis-tapped.

First Contentful PaintPassed

Found

1.6 s, in the green.

Earns you

Something appears quickly, so the page never looks broken.

Speed IndexPassed

Found

3.1 s, in the green.

Earns you

The page looks usable sooner, so visitors keep reading.

What a one-page automated read can't see

Beyond this audit

Speed for your actual visitors
Lighthouse runs one synthetic mobile test from Google's infrastructure. Real-user data (the Core Web Vitals field report from people on your site) tells a fuller story.
Every other page
Homepages are usually the fastest page a site has. Templates, listings, and anything behind a search box are often much worse.
Why it's slow
The metrics say what's slow. Finding the specific script, image, or query responsible takes a trace.

See every finding

2 more findings on this lens, plus every finding on the others, each with its fix. Enter your email and the whole report opens.

AI & Search Visibility

Executive summary

Can AI assistants and search engines find, understand, and recommend you?

This is the weakest lens on the audit, and the reason is a single line in a text file: robots.txt blocks GPTBot and PerplexityBot. When someone asks an assistant for a mortgage broker, this site isn't eligible to be the answer.

The structure underneath is better than the score suggests. Headings are ordered, there's FAQ markup, and the sitemap is in place. The site is readable, it just isn't allowed to be read.

These are presence checks: whether an engine can reach, parse and attribute the site. They don't test whether an assistant actually names you; that needs live prompting against real queries.

How we score AI & Search Visibility

8/ 20

Projected after fixes: 20 / 20

Report details

Lens: AI & Search Visibility
Score: 8 / 20
Checks passed: 3 / 7
Findings: 4

Readiness by engine

Who can actually read you

Same checks, gated per crawler

ChatGPT

9 / 20

GPTBot

Blocked in robots.txt

Claude

13 / 20

ClaudeBot

No rule in robots.txt

Perplexity

9 / 20

PerplexityBot

Blocked in robots.txt

Google AI

13 / 20

Google-Extended

Allowed in robots.txt

Derived from the presence checks below, gated by each crawler's robots.txt access. No engine was queried; whether they actually cite you is a different question, and one worth asking.

Critical gaps

AI crawler access
critical
robots.txt blocks GPTBot, PerplexityBot. Those engines can't read the site.
Structured data (JSON-LD)
major
No JSON-LD on the page. Engines have nothing structured to confirm who you are.
Meta & Open Graph
minor
Missing og:image.
llms.txt
minor
No /llms.txt, so AI crawlers get no curated guide to the site.

Current strengths

Semantic HTML & headings
One H1, ordered headings, landmark elements present.
A clean structure lets an AI assistant quote your page accurately.
Question-shaped content
Question headings with FAQPage markup.
Answering real questions is what answer engines look for.
XML sitemap
/sitemap.xml found.
Crawlers can find every page instead of guessing your page list.

Priority action roadmap

4 fixes

Quick wins

Small, contained changes. Most are an afternoon or less.

P0
AI crawler access
Remove the Disallow rules for GPTBot, ClaudeBot, PerplexityBot, and Google-Extended in robots.txt.
15 minutes
P2
Meta & Open Graph
Ship a 50–60 char title, a 140–155 char meta description, and og:title + og:image.
30 minutes
P2
llms.txt
Publish /llms.txt listing your key pages with a one-line description of each.
30 minutes

Foundation

The structural work the rest of the fixes rest on.

P1
Structured data (JSON-LD)
Add JSON-LD with name, logo, and URL: Organization plus the type that matches the page.
Half a day

Strategic

Positioning and design decisions worth deciding properly.

Nothing in this tier.

Where this lands

Crawler access is the finding that unlocks the rest. Structured data is the next one, and it's what lets an engine confirm you're a real business rather than inferring it from prose.

The opportunity

Unblocking the crawlers is a fifteen-minute change to one file. Adding llms.txt is another half hour. Those two alone are most of the gap.

7 checks ran

Detailed audit

AI crawler accesscritical

Found

robots.txt blocks GPTBot, PerplexityBot. Those engines can't read the site.

Costs you

When someone asks ChatGPT or Perplexity for a broker, you can't be the answer. A competitor is.

Fix

Remove the Disallow rules for GPTBot, ClaudeBot, PerplexityBot, and Google-Extended in robots.txt.

Effort

15 minutes

Structured data (JSON-LD)major

Found

No JSON-LD on the page. Engines have nothing structured to confirm who you are.

Costs you

Search and AI can't confirm your business details, so you're less likely to be shown or recommended.

Fix

Add JSON-LD with name, logo, and URL: Organization plus the type that matches the page.

Effort

Half a day

Meta & Open GraphPartial · half credit

Found

Missing og:image.

Costs you

Search and AI have no clear summary to show, so your listing gets passed over for one that does.

Fix

Ship a 50–60 char title, a 140–155 char meta description, and og:title + og:image.

Effort

30 minutes

llms.txtminor

Found

No /llms.txt, so AI crawlers get no curated guide to the site.

Costs you

AI assistants have to guess what you offer, so they're less likely to recommend you accurately.

Fix

Publish /llms.txt listing your key pages with a one-line description of each.

Effort

30 minutes

Semantic HTML & headingsPassed

Found

One H1, ordered headings, landmark elements present.

Earns you

A clean structure lets an AI assistant quote your page accurately.

Question-shaped contentPassed

Found

Question headings with FAQPage markup.

Earns you

Answering real questions is what answer engines look for.

XML sitemapPassed

Found

/sitemap.xml found.

Earns you

Crawlers can find every page instead of guessing your page list.

What a one-page automated read can't see

Beyond this audit

Whether engines actually cite you
Every check here is a presence check: can an engine read and parse your site. Whether ChatGPT or Perplexity names you when a customer asks for a recommendation (or names a competitor instead) takes live testing against real prompts.
What you rank for today
This audit reads the signals. It doesn't pull your rankings, your impressions, or the queries you're already winning.
Your competitors' visibility
Being readable matters less than being more readable than the alternative. That comparison needs their sites scanned too.

See every finding

4 more findings on this lens, plus every finding on the others, each with its fix. Enter your email and the whole report opens.

Action plan

What to do, and in what order

12 fixes across every lens, sorted by how much work each one is and then by what it costs you. Quick wins first; most are an afternoon.

Working through the quick wins and foundation columns takes this site from 54 to about 84 out of 100. The strategic column isn't in that number; those are decisions, not tasks.

Everything worth fixing, in order

12 fixes

Quick wins

Small, contained changes. Most are an afternoon or less.

P0
AI crawler access
AI & Search Visibility
Remove the Disallow rules for GPTBot, ClaudeBot, PerplexityBot, and Google-Extended in robots.txt.
15 minutes
P1
Gradient text
Trust & Credibility
Use a single solid colour; carry emphasis with weight and size.
30 minutes
P2
Meta & Open Graph
AI & Search Visibility
Ship a 50–60 char title, a 140–155 char meta description, and og:title + og:image.
30 minutes
P2
llms.txt
AI & Search Visibility
Publish /llms.txt listing your key pages with a one-line description of each.
30 minutes

Foundation

The structural work the rest of the fixes rest on.

P0
Largest Contentful Paint
Page Speed
Get the hero painting under 2.5s: preload the LCP asset, serve it in a modern format at the right size, and cut render-blocking CSS/JS.
A day
P1
One primary CTA
Conversion
Pick one action, make it visually dominant, demote the rest to text links.
2 hours
P1
Structured data (JSON-LD)
AI & Search Visibility
Add JSON-LD with name, logo, and URL: Organization plus the type that matches the page.
Half a day
P2
Total Blocking Time
Page Speed
Ship less JavaScript: code-split, defer non-critical scripts, and drop unused third-party tags so the main thread frees up faster.
A day
P1
Testimonials present
Conversion
Add two named testimonials with a concrete result, ideally with a face.
A day, once you have the quotes
P2
Marketing buzzwords
Trust & Credibility
Say what you do in the words a customer would use.
2 hours

Strategic

Positioning and design decisions worth deciding properly.

P0
Hero sells on its own
Conversion
Lead with the plain version: what you do, for whom, e.g. “Home loans for first-home buyers in Brisbane, sorted in days not weeks.”
A positioning decision, then a day of copy
P1
Overused default font
Trust & Credibility
Pair a distinctive display face with a clean body face on a contrast axis.
A type decision, then half a day

See every finding

12 more findings on this lens, plus every finding on the others, each with its fix. Enter your email and the whole report opens.

We rewrote your weakest line

Hero headline

Yours

Your future, secured. Tailored home loan solutions.

Rewritten

Home loans for first-home buyers, approved in days not weeks

Why

Names the borrower (first-home buyers), the outcome (approved), and a time promise. A broker selling to someone else can't paste it unchanged.

What's next

This is the free snapshot. The fix is a site built to pass it.

We design and build new websites for Australian small businesses, then look after them on a care plan that keeps your search and AI visibility working: SEO, AEO and GEO. Scope and price are agreed on the call, before anything is invoiced.

Book a 15-minute call

Website audit · brightpathmortgages.com.au

Overview

Hero sells on its own

AI crawler access

Largest Contentful Paint

Conversion

Does this page turn visitors into enquiries?

Critical gaps

Hero sells on its own

One primary CTA

Testimonials present

Current strengths

Headline under 10 words

Fifth-grade readability

CTA says what happens next

OG image present

See every finding

Trust & Credibility

Does this site look credible enough to trust with your money?

Critical gaps

Overused default font

Gradient text

Marketing buzzwords

Current strengths

Glassmorphism as decoration

Em-dash overuse

Justified text

Tiny body text

Mobile viewport

See every finding

Page Speed

Does this page load fast enough to keep visitors?

Critical gaps

Largest Contentful Paint

Total Blocking Time

Current strengths

Cumulative Layout Shift

First Contentful Paint

Speed Index

See every finding

AI & Search Visibility

Can AI assistants and search engines find, understand, and recommend you?

ChatGPT

Claude

Perplexity

Google AI

Critical gaps

AI crawler access

Structured data (JSON-LD)

Meta & Open Graph

llms.txt

Current strengths

Semantic HTML & headings

Question-shaped content

XML sitemap

See every finding

Action plan

See every finding

This is the free snapshot. The fix is a site built to pass it.