Understand the problem
See what's working, what's getting in the way and what still needs checking.
SEO and AI visibility, checked by a real person
Ask ChatGPT who to call. Does it name you, or your competitor?
More customers now ask AI to recommend a local business. I check what ChatGPT, Gemini, Perplexity and Claude say about you, and who they name instead. Then I check your website and Google data, because AI answers draw on the same basics. You get a plan that shows what to fix first.
A$1,490 + GST. Report within 5 business days. If I can't find at least 3 evidence-backed fixes worth doing, you get a full refund. Nobody can promise an AI will name you, and I won't.
Run by me, Wade Cockfield. Getting businesses found on Google since 2007.
Get a clear next step
Your website should be bringing in more work. Where do you start?
You may have paid for SEO without knowing what changed, or have a long list of recommendations and no idea which matter. RightView gives you a clear explanation and an order to work through.
Know what to do first
Focus on the important repairs before committing to more pages or ongoing work.
Give your team a usable plan
Each priority has a reason, an owner and a way to check that the work is done.
See what a useful audit looks like
What should an SEO audit include?
Learn what evidence, priorities and handover details to expect from an SEO audit.
Read more →Two service pages, two different SEO decisions
See why one RightView workshop review recommended combining a service page and keeping a location page.
Read more →How much does an SEO audit cost in Australia?
Compare published Australian SEO audit prices by scope, deliverables, GST and implementation.
Read more →Under the hood
Not one tool and a score out of 100. A full diagnostic system.
Most SEO audits come from a single crawler and an automated score. RightView combines five sources of evidence and a classifier model of how Google evaluates businesses, and checks the results against what Google has published about how search works. Then I verify every finding by hand before it reaches you.
205Ranking and trust classifiers in the model I run
411Checks in my audit workbook, updated to Google's current guidance
4AI assistants asked your customers' questions
2007The year I started getting businesses found on Google
The evidence I gather
Five sources, captured and dated.
- Your Google data. Search Console and analytics, so findings rest on what actually happened on your site.
- A full crawl of your site. Important faults are re-checked by hand, not taken on a tool's word.
- Live Google results. The search results and keyword data for your services, captured on the day of your audit.
- Real AI answers. ChatGPT, Gemini, Perplexity and Claude, asked the questions your customers ask.
- Your wider footprint. Your Business Profile, directory listings, public records and the sites that link to you.
The classifier diagnostics I run
Six ways of looking at how Google sees you.
- Entity comparison. How you compare with the businesses that get named for your main search, and why you were left out.
- Drift check. Where your online presence has drifted from what you actually sell, and where you sell it.
- Page integrity. Whether your important pages add something new, show real experience, and avoid anything that holds them back.
- Business Profile audit. Serious red flags that can put a profile at risk, kept separate from smaller warnings.
- Credential check. Whether your licences, memberships and qualifications can actually be verified online.
- Link value. The relevance and risk of the sites linking to you, not a raw link count.
The model flags. I decide.
The classifier model is fast and thorough, but it produces leads to follow up, not verdicts. Every flag is checked against your real data before it goes in the report. If the evidence doesn't hold up, it stays out, and I tell you what was ruled out.
The research behind it
I reverse-engineer what Google has published. Then I test it on your site.
Google has put a lot on the public record about how search works: patents and patent applications, research papers, documentation and the guidelines its human quality raters follow. I study it, work out what it means for a business like yours, and turn it into checks. Here is some of the research that shapes a RightView report.
2012 · Knowledge Graph
Things, not strings.
Google announced it would understand real things, such as people, places and businesses, rather than only match the words typed in. So I check whether Google and the AI assistants recognise your business as one clear, consistent thing, or mix it up with someone else.
Shapes: entity authority.
2018 · Patent application US20200050707
Judging a whole website against its field.
This Google application describes building a combined profile of the websites in one subject area, for example sites written by experts and sites that aren't, then classifying a new site by how close it sits to each. So I look at your whole site against the sites already trusted in your category, not just page by page.
An application, not a granted patent: it was abandoned in the US in 2021. It shows how Google's engineers framed the problem, not what runs today.
2019 · BERT in Search
Meaning, not keyword counts.
Google applied its BERT language model to Search to understand what a whole search means, not just its individual words. So I check whether your pages cover the whole topic your customers ask about, in the words they use, instead of repeating a keyword.
Shapes: perception alignment and information gain.
2021 · Passage ranking
One good paragraph can be the answer.
Google can rank a single passage from a longer page. AI answers work in a similar way, lifting a clear paragraph to quote. So I check whether each important customer question has a clear, self-contained answer on your site.
Shapes: information gain and AI visibility.
September 2025 · Quality Rater Guidelines
What Google asks people to look for.
Google's raters don't set rankings, but their guidelines show what Google wants its systems to reward: real experience, expertise, trust and a clear person or business behind the website. My checks are aligned to the September 2025 edition.
Shapes: trust, credentials and page integrity.
Now · AI answers
How AI picks who to name.
AI assistants search, read several sources, write an answer, then point to pages that back it up. Nobody outside those companies knows the exact process, so I don't guess. I ask the assistants your customers' questions and record who they name and which websites they lean on.
Shapes: AI visibility, across all six areas.
A patent is a clue, not a guarantee.
Patents describe methods a company has filed, not what it runs today. Research papers describe experiments. So every idea from this research is treated as something to test against your real data. You won't read "Google's algorithm does this" in my report unless Google has said so publicly.
A real example, simplified
Sometimes the next step is to fix what you already have.
Adapted from a delivered diagnostic. Client details and figures have been removed.
From evidence to a decisionA recommendation, not a results claim
What I found
Several pages were appearing for the same service searches. For a different group of searches, one useful article was already being found.
Why it matters
Those need different decisions. Adding more pages or repeating keywords would not address which existing page should do the job.
What I recommend
For the overlapping pages, decide which one best answers the customer and bring the others into line. For the useful article, improve the answer on that page.
See the evidence behind that recommendation
I filtered the search terms in Search Console, then checked which pages appeared for the same searches. Several pages appearing is a reason to investigate, not automatic proof that they harm each other.
Where pages served the same need, the recommendation was to agree the best destination, preserve useful content and align links and any necessary redirects. Where one relevant article already received the visibility, the recommendation was a clearer answer and useful decision detail.
Completion means checking the agreed page and link changes. Search performance and enquiries are measured separately afterwards. This example does not claim an implemented change or a measured gain.
The handover
A decision you can understand. Detail your team can use.
You get a private online report, a PDF and a walkthrough with me. The supporting working files turn the recommendations into specific actions. These previews show the structure used in real report packs, with client data removed.
Output 01 · Finding card
The reasoning stays attached to the action.
- Finding
- What is happening, and which pages it affects.
- Evidence
- The retained source, capture date and page references.
- Commercial impact
- Why it could matter to discovery or enquiries, with uncertainty stated.
- Action
- What to change, why it takes priority and how to check it.
Output 02 · Page decision inventory
A reason for every page decision.
URLCurrent titleSearch clicksSearch impressionsAverage positionDecisionWhy
These are the columns from an actual page inventory. It records whether to keep, improve, combine or redirect a page and why. A low traffic number alone is not a reason to delete it.
Your working file contains your real URLs and the measured period. No client rows are reproduced here.
Output 03 · Implementation register
Enough detail to do the job and verify it.
The link-repair register records priority, source URL, current anchor and destination, link position, problem, recommended destination, required action and acceptance test. The roadmap adds the owner and dependency.
Action
Update links to a page selected for retirement so they point directly to the agreed destination.
Owner
Your chosen SEO and web team.
Dependency
Confirm the page decision and preserve useful content before the redirect release.
Completion test
A fresh crawl finds no internal links to the retired page. Its redirect reaches the agreed destination in one step.
Adapted from a real implementation handover. Descriptive labels replace private page addresses; this is a format preview, not a client screenshot.
The detail includes what must be confirmed before publishing.
Where the assessment calls for it, the pack includes recommended business-identity structured data and instructions for publishing and validating it. Unverified addresses, credentials or claims stay out until the evidence is confirmed.
The six-part method
Six areas, checked in every report.
Every RightView report covers the same six areas, so nothing important is skipped. You don't need to learn the terms: the report explains what each finding means for your business.
01 · Entity authority
Does the web recognise the right business?
I compare the website, business profile, public records and search results. A correct business identity with thin supporting evidence needs a different response from a business being confused with someone else.
Decision: which identity details need confirming or bringing into line.
02 · Perception alignment
Does the site describe what you actually do?
I compare your real services and operating area with page copy, titles, internal links and structured data. I check whether the right page is appearing for the right customer need.
Decision: what each important page should cover, and which claims need evidence.
03 · Technical foundation
Can people and crawlers reach the useful pages?
I check live responses, indexing instructions, preferred URLs, redirects, internal links and the mobile enquiry path. I verify important crawl findings on the site rather than treating every warning as a fault.
Decision: the specific repairs, dependencies and tests for your web team.
04 · Backlink authority
Who independently vouches for you?
I look at the relevance and credibility of referring sites, the pages they point to and opportunities to recover useful links. A large link count on its own tells me very little.
Decision: what to preserve or recover, and where relevant corroboration is missing.
05 · Information gain
What can a customer learn only from you?
I check for repeated content and repeated purpose separately. Then I look for first-hand knowledge: real projects, constraints, decisions, photographs and expert answers that help someone choose.
Decision: what to combine, improve or create, and the evidence needed to write it.
06 · Performance trajectory
What changed, and what might explain it?
I compare useful periods of search performance, query groups and the pages receiving visibility. I separate brand demand, changing page selection and potential site problems before drawing conclusions.
Decision: the baseline and measures that will show whether the chosen work helped.
AI visibility runs across all six.
I check sampled AI answers, business mentions and citations alongside identity, content and crawler access. A crawler being allowed in robots.txt is different from it receiving a successful page response. Neither proves that an AI service will cite you. I record what was tested and where evidence is missing.
I also report what the evidence rules out.
In a separate completed diagnostic, content comparison did not support a suspicion of widespread copied text. Some pages still served the same purpose. The recommendation distinguished overlapping intent from duplicate wording, rather than claiming a test had found something it hadn't.
Anonymised method example, not an implementation or results claim.
How I set priorities
Repair what you can verify. Test what you can't yet know.
I separate known faults from opportunities that need further evidence. That keeps a practical repair plan from turning into a promise about future rankings.
First · Stabilise
Verified, recoverable faults
Broken routes, conflicting instructions and links to retired pages have concrete completion tests. Clear the important faults and capture the baseline.
Next · Consolidate and prove
Decisions and dependencies
Agree page purposes, preserve useful evidence and confirm business facts. Give each action an owner, a dependency and a test before release.
Then · Grow
Opportunities to test
New topics, relevant mentions and stronger project evidence need a business case. Track search, sampled AI visibility and enquiries without treating a forecast as a result.
Working together
You bring the business context. I bring the diagnosis.
We start with your priorities, profitable services, genuine service area and what a worthwhile enquiry looks like. I agree the scope and the evidence needed, including search and analytics access where available.
I gather and check the evidence, develop the recommendations, then walk you through the findings and handover. Missing access or unresolved facts are called out as limits, not filled with invented scores.
The report stands on its own.
Take it to your team, developer or another agency. It explains what to do, why, in what order and how to check completion. There is no obligation to buy ongoing work from me.
Want help putting the plan into action?
If you'd like me to carry out the work, we can agree priorities and scope after the assessment.
Explore SEO & AI VisibilityKnow what you're buying
What RightView is, and what it isn't.
RightView is
- A diagnosis of why you show up where you do, and why you don't
- Built on Google's own published research, tested on your real data
- Evidence you can check, with the source and date for every finding
- A fix list in priority order, each with a way to prove it's done
- Run, checked and explained by me
RightView isn't
- An automated score out of 100
- A rank tracker or a keyword report
- A promise that Google or an AI will name you
- Guesses dressed up as facts
- A sales pitch for ongoing work
Find out who AI recommends, and what it will take to get on that list.
I ask the questions your customers ask, the way they ask them. I note who gets named and which websites the answers lean on. Then I go through your site and search data and tell you what to fix first.
For A$1,490 + GST, you get:
- A private online report and a PDF, in plain English
- A walkthrough with me
- A fix list in priority order, with the evidence and a way to check each fix is done
- Checks drawn from Google's patents, research and rater guidelines, verified on your own data
- Working files your web team can use straight away
Your report arrives within 5 business days. If I can't find at least 3 evidence-backed fixes worth doing, I'll refund you in full. I do it myself, start to finish. Not a junior, and not a free tool spitting out a score.