The Complete AI SEO Guide with OptimAI

This AI SEO guide walks through how modern search has shifted - and how OptimAI measures it. AI engines now answer questions your customers used to type into Google. OptimAI is the audit tool built for that shift: we check whether AI engines can extract your content, whether they're actually citing your domain right now, and what they're answering for your target keyword - then hand you a prioritised to-do list of what to fix, why it matters, and how it's prioritised by impact.

Where traditional SEO tools hand you a spreadsheet of metrics, OptimAI scores both halves of modern search in a single scan:

• AI Search Readiness - 40% of your total score, across six dimensions legacy SEO tools don't measure at all: structured data, Q&A patterns, direct answer density, citation worthiness, entity richness, content freshness.
• AI Search Visibility - direct citation checking on Perplexity, OpenAI (ChatGPT) and Google. We ask each engine your keyword and tell you whether, where and how your domain gets cited.
• AI Search Opportunity - reverse-engineers what AI engines are actually answering for your keyword and flags the topics, questions and formats your page is missing.
• Traditional SEO fundamentals - Content, On-Page, Technical and Performance (60% of the total). Everything you'd expect from a serious audit tool, alongside everything the incumbents leave out.

Every suggestion is personalised to the specific page and keyword you scanned. No generic advice, no filler, no "industry best practices" copy-pasted from a template - the scorecard reads your page and tells you what to do about it.

This is the definitive walkthrough of OptimAI, its scoring methodology and why this is the best tool to get your site visible in AI and improve your Google ranking. Every panel, every dimension, every tip worth knowing - plus a real-world example scan of Salesforce's UK product page end-to-end, so you can see exactly how the scorecard reads on a real site.

If you just want the UI tour, skim section 2. If you want to understand the scoring, head to section 3. If you're evaluating Pro tier, section 4 covers everything. For a working rhythm once you're up and running, see section 5.

1. Getting Started

OptimAI is an AI SEO audit tool. Paste a URL and a target keyword, get a 20-dimension scorecard covering how well the page ranks in traditional Google search and in the new generation of AI answer engines (also called Answer Engine Optimization, or AEO). The free trial gives you enough credits for ten full analyses, no card required.

1.1 Logging in

Sign in with Google or Microsoft. A Standard account is created automatically with 120 free credits - enough for ten full analyses. You'll land on the empty dashboard:

• Scan bar - URL field + keyword field + Analyze button (showing the credit cost for the scan)
• Free credits badge - your starting balance
• Empty state - once you run a scan, every scan you ever run will appear in this table

1.2 Run your first scan

Paste a URL. Enter a target keyword. Click Analyze. The Analyze button shows the credit cost dynamically so you always know what you're spending before you spend it.

1.3 Your first report

A scan takes around 45-60 seconds. The report opens automatically when it's done. You get a single long scrollable view broken into sections:

• The scan bar at the top showing the scanned URL + keyword, with a Refresh button (runs a fresh scan at the reduced refresh cost)
• A section nav bar: AI Search / Content / On-Page / Technical / Performance / Keywords / SERP / History - click to jump straight to that section
• The Report / Action Plan tabs top-right - switch between the full report and a todo-list view of every suggestion across every dimension, prioritised by impact
• The top panel with your overall score, five per-category scores, and a dimension heatmap
• Each category section in turn - top panel, per-dimension panels, expandable metrics drawers
• The Download PDF button bottom-right

Tip: Throughout the system you'll see small (i) icons - click for a plain-English explainer of that item. Every dimension, every metric, every weighting has one.

1.4 Real example explained: Salesforce

If you just want the UI tour, skip to Getting Around. This section walks through a real report end-to-end and is the clearest single view of how to read OptimAI - worth 5 minutes.

The Salesforce UK products page scoring 47 is a fair real-world example - big brands aren't immune, especially on AI search readiness. Let's walk through a real scan (salesforce.com/uk/products against the keyword "salesforce products", download the full PDF here) and unpack what each number is telling us.

The headline: 47/100 - dragged down by AI Search

If we excluded AI Search entirely and just scored the traditional SEO dimensions (Content + On-Page + Technical + Performance, 60% of our weighting), this page would land around ~58 - solidly amber, a "decent page that needs work" rather than "in trouble". The combined score of 47 isn't an error; it's what happens when you honestly measure a page optimised for 20 years of traditional SEO but zero years of AI search.

AI Search Readiness: 31 - "low readiness"

This is the single biggest lever.

Two red flags stand out: Structured Data at 15 (schema markup nearly absent - no FAQPage, no Article, no Organization) and Entity & Data Richness at 4 (almost no concrete data points, no pricing, no customer counts). AI systems favour extractable facts over marketing prose, and this page reads like marketing prose.

Salesforce's page doesn't clear either bar.

Content: 72 - the saving grace

If AI Search is the red flag, Content is why the overall isn't in the red. The page has 1,749 words of substantive copy covering multiple product subtopics. Headings are clean, readability sits in the fair-to-good band, the writing is clear. This is the category where Salesforce's size shows up - they've had editors polishing this copy for years.

On-Page (45), Technical (54), Performance (50) - the "basics overlooked" story

Even at 20 years of SEO investment, the basics have gaps. A few worth flagging:

• Images: 14 - only 25% of the page's 211 images have alt text. 159 images have no alt attribute. That's a 5-minute-per-image fix and a compounding signal.
• Keyword not in H1 - "salesforce products" doesn't appear exactly in the H1. (Caveat: we guessed this keyword from the SERP title; Salesforce's real target could differ.)
• Meta description 61 characters - outside the optimal 50-60 window, so Google will rewrite the SERP snippet itself.
• External links: 47 - excessive, dilutes link equity.

AI Search Visibility: cited by 2 of 3 engines

When we ask the AI engines "salesforce products":

Two of three engines cite the Salesforce domain - not always the exact /uk/products page we scanned, but a page from the domain. That's the domain-authority effect: AI engines know Salesforce is an authoritative source and pull in some page from the domain, even if it's not the best-optimised one.

Why 47 is the right score

Salesforce is the 1,500-pound gorilla of CRM, ranks #2 in Google SERP for "salesforce products", is cited by two of three AI engines, and has 1,749 words of clean copy. And it scores 47. That's not a flaw in the tool; that's the tool doing its job. Other SEO audit tools would give this page an 80+ and a gold star, because they'd be measuring 20-year-old signals. We're measuring what actually drives discovery in 2026: can AI engines extract, understand, and cite your content? For this page, the honest answer is "partially, and only by leaning on domain authority".

The receipts: OptimAI scoring its own landing page

We run the same tool against our own landing page (optimai.live/free-ai-seo-audit, keyword "free ai seo audit") to prove we eat our own dog food:

The difference isn't domain authority or word count - Salesforce wins both on raw inputs. The difference is deliberate structure: FAQPage and Article schema, explicit entity naming, short paragraphs, concrete data points, cited external sources. Every suggestion the tool gives other pages, we've already taken on our own.

P.S. This scan was run pre-launch, before the site was indexed - so you'll see zero AI citations and no ranked keywords on our own report. That's expected for a brand-new domain. The point of the comparison is the structural score (how well the page is built for AI search and on-page fundamentals), not the visibility score - visibility will build naturally as the site gets indexed and linked.

2. Getting Around

A quick tour of the UI surfaces. This is shallow on purpose - the meat is in Section 3.

2.1 Dashboard

After your first scan, the dashboard becomes your home view. From any report, click the OptimAI logo (top-left) or the app back button (next to the Page Report heading) to return to the dashboard at any time.

• Your Pages table - every scan you've ever run (we keep lifetime history). Columns: # / Page / Keyword / Score / Change / Scans / Last Scan / Refresh. Click any row to open the report.
• Filter input (top-right of table) - filters by URL or keyword, client-side, instant.
• Refresh button - re-scans the same URL+keyword at the reduced refresh cost.
• Region banner (top, dismissible) - tells you which country we're scoring against. Change from the settings menu.

2.2 Settings menu

Open from the avatar bottom-left of any page. Header shows your name, email, current credit balance. Menu items: Account (opens the Account modal), Buy Credits, Transactions, Region, Sign Out.

2.3 Account modal

Your full account profile: identity, tier, credit balance, recent activity, upgrade path to Pro, and a Deactivate Account option (your scan history and credit balance are preserved - contact support to reactivate).

2.4 Buy credits

Five credit packs available in GBP and USD. VAT exclusive on GBP transactions (UK 20%), rest of world exempt. Credits never expire. Need more than £1,000? There's a bulk pricing contact link.

2.5 Transactions

Full ledger of every credit movement on your account. Every scan debit, every purchase, every bonus, every refund - dated, categorised, with running balance. Reachable from settings menu or the Account modal's Recent Activity link.

2.6 Region picker

Region matters. If you're optimising for a UK audience but your account is set to United States, search volumes, keyword rankings and competitor data will all reflect US Google rather than the Google your customers actually use. We auto-detect region from your IP, but you can override it any time from the region picker in the settings menu.

Changing the region only affects new scans. Existing scans stay tied to the region they were scanned under - if you want to re-region a historical report, run a fresh scan after switching.

2.7 Same page, different keywords

Each URL+keyword pair is treated as a separate scan and gets its own dashboard row. Scanning the same page against multiple keywords is a legitimate and useful pattern - different keywords surface different strengths and weaknesses.

3. Report Deep Dive

The single-page report is one long scrollable view broken into sections. Every score on the page comes from the same scan - the dashboard, the report, the PDF and the action plan are all reading the same numbers.

3.1 Three navigation surfaces

The report gives you three navigation surfaces, each tuned for a different reading style:

Section Nav Bar

Buttons for every section (AI Search, Content, On-Page, Technical, Performance, plus Keywords, SERP and History when data exists). Active section highlights as you scroll.

Top panel score gauges

Five clickable category score circles in the top panel (AI Search / Content / On-Page / Technical / Performance). Each is a smooth-scroll link to that category's section. Also serves as a compact weight indicator - click the info icon to see what % of total each category contributes.

Dimension Heatmap

Every dimension shown as a coloured tile, sized by its weight in the total score and coloured by its rating (red / amber / green). Click any tile to jump to that dimension's panel. The at-a-glance "where's the damage" view - the biggest red rectangle is your biggest problem.

3.2 AI Search Readiness (40% of total)

The biggest category by weight, and the part most legacy SEO tools don't measure at all. The AI Search top panel shows the category score, a plain-English diagnosis, an explainer, and a horizontal bar chart of the 6 dimensions ranked. Below the top panel, each dimension gets its own panel with score, verdict line, up to 3 top priority suggestions, and an expandable "N metrics" pill that reveals every measured signal.

Tip - Suggestions: Click the "N suggestions" toggle on any dimension panel to expand and see what the most important fixes are for that dimension. These are page-specific and keyword-specific, ordered by impact. Recommendations will update over time as you resolve issues.

Tip - Metrics: Click the "N metrics" pill at the bottom of any dimension panel to reveal the specific signals feeding that dimension's score - what we measured, what value we found, and how heavily each one weighs into the total. Don't forget the (i) info icon on each metric for a plain-English explanation of what it means.

Tip - Back to top: OptimAI packs a lot of information into each report. A small round up-arrow button appears in the bottom-right once you start scrolling - click it any time to jump back to the top of the page.

3.3 AI Search Visibility

Below the 6 dimension panels, the AI Search Visibility panel tells you whether AI engines are actually citing your domain right now for the keyword you scanned.

Layout: status banner (green tick or red X), citation timeline sparkline, three engine panels (Google / OpenAI / Perplexity), and a Competitor Analysis table comparing your page vs the average of the top cited pages.

Each engine panel shows a badge: Cited (green), Mentioned (amber - in text but not formally cited), Not Found (red), or Unavailable (engine API errored). Cited engines expand to show the actual citation snippets where available.

Caveat on Google: the way we query Google differs from how we query Perplexity and OpenAI - which means for Google we can confirm whether your domain was cited but not always show the exact snippet that referenced you. For Perplexity and OpenAI you get the full context around your cited reference.

AI search results are non-deterministic. Re-running a scan can produce different results because the engines themselves don't always cite the same sources. Patterns matter more than single data points - which is why the timeline bubble chart exists.

You'll only see the bubble chart timeline after you have run more than one scan. It's possible to run more than one scan in a day, each of which may get entirely different citation results. The bottom row of the bubble chart shows how many scans were run on a given date - in the screenshot above you can see the highlighted column shows 3 scans run on 27th March. We show the highest count of citations from the scans run that day - here you can see one scan got 3 citations from OpenAI and 1 from Google, which could have come from different scans that day.

If you click one of the green bubbles you'll get a tooltip telling you which scan from that day got you the citations - click the timestamp link where it says "Source" to view that report in full, so you can check the exact details.

When possible (this depends on whether AI engines return any citations at all, which isn't guaranteed) we also analyse the other sites that were cited against the same AI Search Readiness metrics as your page. We distill this down into a comparison of which AI Search Readiness dimensions other cited sites are achieving higher scores in than you - giving you a real-time AI Search competitor analysis.

3.4 AI Search Opportunity

The companion to AI Search Visibility. While AI Search Visibility tells you whether AI engines are citing your domain, AI Search Opportunity reverse-engineers what AI engines are actually answering for your keyword and flags what's missing from your page.

How it works: when you scan, we query Perplexity, OpenAI and Google for your keyword and analyse their responses. We extract the questions AI engines are answering, the topics they cover, and the domains they cite. Then we compare against your page content and flag the gaps.

The panel shows:

• Coverage stats: "X of Y topics covered", "X of Y questions answered" - your gap at a glance
• Suggestion panels in four types, each colour-coded:
  • Content Gap (red) - subjects cited sources cover that you don't. Add this content to compete.
  • Topic Expansion (amber) - topics you mention but don't cover deeply.
  • Format Change (teal) - structural changes (FAQ schema, lists, tables) cited sources use.
  • New Content (dark teal) - standalone pages worth creating to capture adjacent queries.
• Expandable lists of all questions and topics - items already on your page show a "you" badge
• Cited domains - the actual sites AI engines are pulling from

3.5 Content (24% of total)

Category top panel opens with a plain-English diagnosis, an explainer, the category weight tag, and a horizontal bar chart of the 4 dimensions.

3.6 On-Page (20% of total)

The HTML mechanics search engines and AI crawlers read first. Less glamorous than AI Search but cheap to fix and essential to get right.

3.7 Technical (10% of total)

The behind-the-scenes plumbing search engines and AI crawlers check first. If we can't measure a dimension during the scan - usually because the page requires login, hides content behind JavaScript, or the server times out - we assign a neutral score of 50 (the exact midpoint) so it doesn't pull your overall score up or down.

3.8 Performance (6% of total)

Page load speed and rendering quality measured during the scan, split into Core Web Vitals (Google ranking signals) and supplementary speed metrics. Same fallback as Technical: if a dimension can't be measured (login-gated pages, heavy JavaScript, server timeouts), it gets a neutral score of 50 so it doesn't skew your overall score.

3.9 Keyword Opportunity

After the 5 scoring categories, the Keyword Opportunity panel lists related keywords people search for alongside your target, with a coverage score showing how many already appear on your page.

Why it matters: each related keyword is a real search query where people could find your page - but only if your content actually mentions it. Every variant in the table that your page doesn't cover is a door you're not opening. Adding the high-volume, low-difficulty ones you're missing is one of the cheapest ways to get ranked for more searches at once - you're not writing new pages, you're just making sure the page you already have is visible to more of the traffic that's already looking for what you do.

Layout: a coverage gauge (top), stats row, insight callout, variants table (best related keywords with monthly search volume, difficulty 0-100, CPC, and "On Page" yes/no - top opportunities flagged automatically), and a People Also Ask list from Google's PAA box. If there isn't enough related search data for your keyword (usually very narrow branded queries), the panel simply doesn't appear on the report.

3.10 SERP Intelligence

The "what does the SERP actually look like for this keyword" view. Pulls live Google results for your keyword + region, finds your page's position (if it ranks), and shows the top 10 with competitor metrics.

Keyword specificity: one page, two keyword variants

To show how keyword choice alone can move a ranking, here's the same Planning Direct page (/find-out-about/permitted-development/article-4-directions) scanned against two related keywords. Same HTML, same content, same domain authority - only the keyword changes.

Variant 1 - ambiguous keyword "article 4": ranks #4, but the SERP is polluted with off-topic results. Wikipedia's "Article 4 direction" (#3, planning-relevant), Ealing Council (#5), NATO's "consultation process and Article 4" (#7, completely different Article 4), Equality & Human Rights' "Article 4: Freedom from slavery and forced labour" (#9, yet another unrelated Article 4). Google doesn't know which "Article 4" the searcher means, so it serves a mix.

Variant 2 - specific keyword "article 4 directions": rank jumps to #2 (up from #4) and the competitor set cleans up completely - legislation.gov.uk, Dorset Council, London.gov.uk, Bristol Council, gov.uk. All UK planning, all authoritative, all directly relevant. Adding one qualifying word moved a real page 2 positions up the SERP AND eliminated four off-topic competitors.

The specificity lesson: Variant 1 → Variant 2 is the move most pages have waiting for them. Not a content change, not a schema change, not a link-building project - just picking a keyword that matches the page's specific intent. Now you know the perfect target, apply it and optimise for it by following OptimAI scan suggestions and get cited more by AI and rank higher in Google!

The "not found" diagnostic: competitive SERP defeat

Specificity isn't the only trap. Here's a different Planning Direct page - /find-out-about/permitted-development/loft-conversion, a page explicitly about loft conversions - scanned against the obvious keyword "loft conversion". Result: not found in the top 94.

The keyword-page fit is correct this time - the page is literally about loft conversions. But the SERP is dominated by high-authority sites (Federation of Master Builders, HomeOwners Alliance, reddit) and local-pack business listings that Planning Direct can't outweigh on domain authority alone. This is a different failure mode: the page belongs in the SERP by topic, but the competitive ceiling is too high.

The diagnostic lesson: "Not found" doesn't always mean wrong keyword. Check the top 10 competitors - if they're all huge authority domains or local-pack listings you can't outrank, the realistic path is a long-tail variant ("loft conversion permitted development rules UK", "when do you need planning permission for a loft conversion") rather than the head term. SERP Intelligence shows you the ceiling before you spend months trying to reach it.

3.11 History

The History panel only appears when you have 2+ scans of the same URL+keyword pair. It tracks score changes over time so you can see whether your edits are moving the needle.

Layout: current score gauge with delta vs first scan, sparkline of overall score over time with red/amber/green bands at 40 and 70 thresholds, and a per-scan table with every historical scan (date, overall score, rating, SERP position, per-category scores, and a View button to open that historical report).

3.12 Action Plan

Nearly last but certainly not least, the Action Plan is the original reason we built OptimAI and probably the most useful feature.

Top-right of the report has two tabs: Report (the deep view we've just walked through) and Action Plan - a stripped-down to-do list of every suggestion across the report.

The Action Plan offers two views, one sorted by category & dimensions (see above) and the other sorted by priority (see below). This helps you focus your work by moving systematically through dimensions, for example overhauling Content across a set of pages in one push, or by sweeping pages for the most critical fixes that will move the needle fastest.

You can tick off suggestions as you work through. OptimAI tracks these between scans so you can see at a glance whether each item is resolved, still needs work (the changes you made weren't significant enough to resolve the root cause), or is new - generated since the last scan.

3.13 Download PDF

Finally the Download PDF button - does what it says on the tin! The bottom-right of every report has a Download PDF button which generates a PDF copy of the full report - same scores, same explanations, same tip text - rendered as a clean print-friendly document. It also includes the full Action Plan so you have a handy cheat sheet to work from.

4. Pro Tier

Everything covered so far works on every tier. This section covers what Pro and Partner tiers add: whole-site audits, batch scanning, branded reports, and team workspace sharing. If you're scanning more than a handful of pages a month, or reselling audits to clients, these are the features that make the work repeatable and presentable.

4.1 Standard vs Pro vs Partner at a glance

The biggest line isn't a feature - it's the credit rate. Partner subscribers get 1.5x credits for their money compared to pack purchasers. £500/month on Partner buys 750 credits; £500 of credit packs buys 500. If you're running more than ~20 analyses a month, Partner is almost always the better economics and brings scan and refresh costs back to parity with Standard tier pricing.

4.2 Partner subscriptions

Partner is our self-serve monthly subscription for agencies, consultants and in-house teams running regular audits. Here's how it works:

• Credits arrive on your billing day each month
• Credits never expire - unused credits roll forward indefinitely
• Cancel any time via the Stripe customer portal - your balance stays, you drop to the Pro tier
• Had a busy month and need just a few more scans? You can always top up mid-month with regular credit packs for that bit extra.
• VAT applies at 20% on GBP transactions; USD is VAT-exempt (rest of world)
• Need more than the self-serve tiers? Contact us for custom pricing
• Keep all your historic report data regardless of tier. You paid for it so it's yours forever.

4.3 Pro Dashboard

Pro tier has a slightly different dashboard with two levels: the top level shows your scans grouped by domain with aggregated values. Click on a domain row to drill down into a table listing the scans run for just that domain - otherwise the same as the standard tier dashboard, including the filter to quickly find scans.

From the domain dashboard you can also add a domain directly to run a site crawl. This will display the Site Audit page tree (more on this below). You can also enter this view for any existing domain in the list by clicking the Site Audit button in the table.

If you've already run a single-page scan, the domain will already be listed here and you'll find the scans on the drill-down for that domain - no need to re-add via the Add Domain button.

4.4 Site Audit - Pages tab

The page tree (or Pages tab) within Site Audit is your main place to overview the latest state of your scanned pages for that domain (or subdirectory/section) and run bulk operations.

At the top of every tab within Site Audit you will see domain level metrics: backlinks, domain rank, referrers and anchor text, etc. All standard stuff, except you don't get charged extra for it, unlike with some other tools! Next there is a large amber panel explaining how to use the bulk tools - you can dismiss this once you are familiar, and if you need a reminder just pop back here to these help documents.

After the explainer panel (or after the domain stat cards once you've dismissed it, as in the image below) you'll see the page tree, which is a sitemap displayed in a similar way to a file and folder tree on your computer. You use this to select pages you want to perform bulk operations on (e.g. batch scan, summary reports or batch PDF downloads) by clicking the checkboxes.

Each row in the table shows the page path (with an icon to open the target page in a new tab), the page title we detected, and a text input for the keyword. If a scan has been performed on that page before, you will also see the Score Bubble for the most recent scan matching that page and keyword pair, and the date of that scan. Clicking the Score Bubble will take you to that specific page report.

If you have run a scan against a page for multiple keywords, the keyword input will become a drop down and allow you to select which keyword report to include. (In the screenshot above you can see this for the page /uk/media/automated-production-broadcast-software.)

For pages which have not been scanned we auto-detect a keyword suggestion (derived from the page's H1 heading, with fallback to the meta title or the URL slug if no H1 is present) which will be shown in grey text. For pages which have been scanned the keyword which was used will be shown in black text. Any text you enter here will be black and used for the batch scan when you trigger it.

4.4.1 Site Audit - Pages -> Export

Click the Export button (bottom right) to generate a Google Sheets version of the page tree sitemap. This will open in a new tab from where you can make a copy to edit or download as an Excel file. It includes checkboxes to select pages and a formula which calculates the credits required (including new scans or refreshes of previous scans at the different costs).

This is great to share with your team and plan your keyword strategy - for example listing out keyword variations you may want to run (just duplicate the original row). The file is formatted to follow the same grey/black keyword colouring as the live page tree, indicating suggested versus user-entered keywords.

4.4.2 Site Audit - Pages -> Download PDFs

Select some pages in the tree map and the Download PDFs button (next to the Export button, bottom right) will be enabled. Click it to download all the individual page report PDFs for each page selected in one zip file. (Note: if you have selected pages for which no scan has been run these will not be included, as there is no PDF generated from non-existent scans!)

4.4.3 Site Audit - Pages -> Scan

With some pages selected, the Scan and Report buttons will be enabled. They are found at the bottom of the page in what we call the Budget Bar - this does the maths for you and calculates the cost based on new scans and refreshes, showing the total cost in the Scan button itself and also your current balance (red if insufficient credits).

Click the Scan button to generate new page reports or refresh existing ones for the pages you have selected in the page tree. This runs as a batch operation which you can leave running in the background, as it may take some time. You can opt to receive an email notification when the batch is completed.

Remember to check your keywords before clicking! Once you click, the credits are deducted. (By the way, if any scans fail at any time you will get those credits automatically refunded.)

4.4.4 Site Audit - Pages -> Report

Reports run from the Site Audit page tree view are executive summary reports collating and aggregating the data from multiple individual page reports. You can generate an unlimited number of these reports for free, so you can slice and dice your site across whatever pages you like in as many ways as you like. We'll dive into further details about the summary reports below.

4.5 Site Audit - Reports tab

Clicking the Reports tab at the top right of the Site Audit page will take you to a listing of all previously-run executive summary reports. You will see a summary panel for every report showing headline results. Click the panel to view the full summary report. Click the trash can icon in the top right of a panel to delete a summary report (don't worry, all individual page reports will be preserved).

4.6 Site Audit - Keywords tab

The Keywords tab is your domain-level keyword intelligence view - what the whole site is ranking for, which pages are pulling the rankings in, and where the biggest untapped opportunities sit across the entire domain. Essentially this is the site-wide version of the Keyword Opportunity panel in single page reports (see Section 3.9 above).

There are three sections, you can jump to each from the page nav highlighted in the image above:

• Opportunities - keyword-page pairs you're already on page 2 or 3 for and could get to page 1 of Google with a push
• Unscanned - pages on your site already ranking for keywords but not yet audited in OptimAI
• Clusters - thematic groupings of keywords, a lens on which topics your domain has authority in and where the gaps sit

4.6.1 Site Audit - Keywords -> Opportunities

The Opportunities view lists keywords your domain is nearly ranking well for - typically sitting just outside the first page. Each row is a specific keyword + page pair: the phrase, where you rank today, search volume, difficulty, CPC and the page pulling the ranking.

Why this matters: moving a keyword from position 14 to position 4 is vastly cheaper than getting a new page to rank at all - Google already knows you for that query, you just need a targeted optimisation push. Click Scan on any row to run a full OptimAI audit for that page+keyword pair and get a prioritised to-do list for getting it over the line.

4.6.2 Site Audit - Keywords -> Unscanned

Unscanned lists pages on your domain that are already ranking for one or more keywords but haven't been through an OptimAI audit yet. Expand any row with the triangle to see which keywords that page is ranking for, their positions, and any opportunity badges attached. Click Scan to run a in-depth report and add this page+keyword combo to your optimisation workflow.

This view is where you catch the outliers - pages you'd never have thought to scan. That old blog post from 2021 quietly ranking #3 for something you never targeted on purpose. The news announcement that flukes on the exact right recipe for AI search readiness. Every domain has a handful of these accidental winners, and they're often the fastest pages to convert into genuine traffic assets once you know they exist. Picks up directly on the point in section 4.4: you don't need to scan every blog or news post to find the nuggets - this view surfaces them for you.

4.6.3 Site Audit - Keywords -> Clusters

Clusters group the keywords your domain ranks for into thematic buckets - related phrases that all circle the same underlying topic. Each cluster shows its label, how many keywords sit inside it, how many pages on your site contribute to it, and the best ranking position across the group. Click the triangle on the left of any cluster to expand it and see the individual keywords, positions and opportunity badges inside.

Why clusters matter: they're the clearest view of where your domain has topical authority and where the gaps sit. A dense cluster (lots of keywords, lots of pages, strong positions) is a topic Google already sees you as credible on - extending it with a hub page or supporting article typically compounds fast. A sparse cluster hinting at demand you're barely addressing is a signal to either create new content or weave the variant keywords into existing pages for additional pull without building anything new.

4.7 Site Audit - Site Health tab

The boring hygiene stuff - but more important than most people realise. Every item in this tab is an issue quietly bleeding ranking authority, wasting crawl budget, or keeping pages out of Google's index entirely. None of it is glamorous to fix. All of it tends to be overlooked - which is precisely why it's a reliable win when you don't.

Five categories, each summarised by a count tile and expanded below into the full list of offending URLs (full tooltip on each tile if you need a refresher):

• Broken - URLs returning a 4xx error. Dead links signal neglect to both Google and users, and waste crawl budget on every scan.
• Redirects - URLs that 3xx somewhere else. Individual redirects are fine, but chains (redirect-to-redirect) leak link equity and slow page load.
• Canonical issues - misconfigured or conflicting canonical tags. Google will pick one - usually not the one you'd have picked - and split the ranking weight across the others.
• Orphan pages - pages in your sitemap with no internal links pointing to them. They exist but nothing's sending them authority, so they rarely rank.
• Unlisted pages - pages discovered by crawl but missing from your sitemap. Google may never index them - and sometimes it surfaces pages that shouldn't be live at all (stray test pages, content missing from the sitemap by accident).

4.8 Summary Reports

Summary reports - also called Executive Summary reports - take a whole audit (a batch of page scans, a section of a site, a client engagement) and condense it into a single scrollable page you can share, present or hand off. Every panel is an aggregate of the pages you chose when the report was generated, so the output matches exactly the slice of the site you care about.

The top panel (shown above) is the headline view: overall weighted score across the whole set, the five category gauges, a distribution bar (how many pages sit in good / fair / poor), Improvement headline stats (how many pages moved, up or down, since the previous scan), AI Citations headline stats, and the full Dimension Heatmap - this time showing the average score per dimension across the whole set, so a red tile means the set of pages included in the report is weak on that signal, not just one page.

At the top of the page is a section nav which allows you to jump directly to that section of the report, each of which we will walk through below.

4.8.1 Summary -> Improvement

Headline stats first: how many pages have improved, declined or stayed unchanged since the last scan, and the average score delta across the whole set. Beside that, a sparkline of the page set average score over time (populated once you have more than one scan run), then a per-page list of movers with the score change per page.

4.8.2 Summary -> AI Citations

Domain-level citation intelligence across the page set. Headline stats at the top (pages cited / overall citation rate / total citations). Buttons for each engine showing total citation count which you can click to filter to Google / OpenAI / Perplexity individually. Alongside that, an average citation rate history chart shows how the rate has moved over time - builds up as you run multiple scans.

Below there's a table listing each included page with a mini citation sparkline per row, plus hit / miss / gain / loss indicators for each engine. Expand any page to see its full Best-of-Day bubble chart - same pattern as the single-page AI Search Visibility panel, so no new mental model to learn.

4.8.3 Summary -> Recommendations

Next we've reviewed all the reports for each page and distilled this to critical strategic plays - the "if you only did five things across this whole audit" list. Each recommendation is tagged High / Medium / Low priority with a one-paragraph rationale based on the specific pages and patterns that triggered it. Hand this to whoever's briefing the next optimisation sprint - it's the executive summary of the executive summary.

4.8.4 Summary -> Keyword Opportunities

Same pattern as the single-page Keyword Opportunity panel and the domain Opportunities view - just scoped to the pages in this report. Keywords with real search demand, with the usual Top opportunity badges flagging the achievable-and-valuable ones.

Click any row for a popup that reveals which pages from the report surfaced the keyword, so you can click through to the page and run a fresh report for the new opportunity keyword on that page. This is the same "from insight to action" flow as everywhere else in the tool.

4.8.5 Summary -> Cannibalization

Flags cases where two or more pages in the set are chasing the same keyword and fighting each other for rankings. The severity pill (MILD / MODERATE / SEVERE) tells you how bad the overlap is; the Strongest badge tells you which page is currently winning, so you know which page to consolidate onto.

4.8.6 Summary -> Top 5

Two panels back to back in the report flow: Top 5 Strongest Pages (your current winners - the templates, content patterns and structural choices that are already working) and Top 5 Priority Fixes (the highest-impact pages to work on next, sorted by expected movement). The strongest list doubles as a brief for replication: whatever those five pages are doing, do more of it on the rest of the site.

4.8.7 Summary -> Common Issues

The recurring suggestions across the whole page set - issues that hit a meaningful share of pages, not just one-offs. Each row shows how many pages are affected (e.g. "20/20" = every page, "16/20" = 80%). Ordered by severity then by reach. Capped at the top 10 most significant issues and filtered to anything affecting at least 25% of the set, so it's always actionable rather than overwhelming.

The panel covers all five categories (AI Search / Content / On-Page / Technical / Performance) - in the example above everything happens to be AI Search because this particular site is strong on the fundamentals and weak on AI readiness, but the list spans the full scorecard whenever issues are spread across dimensions.

Issues surfaced here update as you fix them. Once a given issue drops below the 25% pages-affected threshold it falls out, and whatever's next most common takes its place. That turns this panel into a live scoreboard rather than a fixed report - each time you regenerate the summary after a work sprint, the list reflects what's left to do.

4.8.8 Summary -> Breakdown

The per-page matrix. Every page in the set, with its SERP rank, overall score and scores for each of the five categories. Click any row to open the full single-page report underneath.

This is the panel for comparison over time. Rerun the summary report for the same page set each month and drop the tables side by side - the shape of the scores tells you which categories are moving and which are static, per page. It's invaluable for month-by-month client reporting: same visual, different numbers, clear trajectory.

4.9 Team Sharing (domain workspace)

OptimAI does team sharing differently: no invite flows, no seat management, no per-user billing. We use your email domain as the workspace boundary.

Sign in with your company email (alice@agency.co.uk). Anyone else signing in with the same domain (bob@agency.co.uk, charlie@agency.co.uk) is automatically part of your team.

To enable sharing, find the Team Sharing option from the settings menu and toggle it on there.

Sharing requires 2-way consent: if members of your team have it enabled and you don't then they won't see your reports and you won't see theirs. Once enabled both ways reports are shared (per page and summaries), branding (if enabled, see below) is applied for PDF exports and your team can operate a shared credit pool.

The shared credit pool will automatically deduct credits for reports run by any team member from the account with the highest credit balance - so, for example a manager could have an account and control budget through credit purchases as desired, while other members of the team can use the tool freely within that budget limit. There is of course nothing to prevent you or other team members adding credits too, but these won't get used until all other team members have less credits on their balance.

4.10 Branding (white-label)

Pro users can configure a brand profile that applies to all exported PDFs - individual reports, batch reports, and executive summaries. The profile is domain-level: configure once at @yourcompany.com, every team member using that domain sees the same branded output. The first user in your company to enable and save these settings will become the "brand owner" and these settings will be locked out for other members of the team to prevent un-approved editing.

Configurable:

• Logo - PNG, JPG, SVG, or WebP (up to 2MB). Replaces the OptimAI logo in the PDF header.
• Primary colour - hex value. Scores, section headings, accent bars.
• Secondary colour - hex value. Secondary accents.
• Header font - any Google Fonts family available from the built-in font-picker.
• Header font weight (100-900) - where available for the chosen font.
• Body font - any Google Fonts family available from the built-in font-picker.

Note: the OptimAI footer will also be removed completely.

5. When to Scan

OptimAI is intended as an optimisation discipline, not just a one-off audit (though obviously you can do that if you like). The real compounding value comes from running scans on a cadence that matches the work you're doing. Run scans too often and you burn credits on noise; run them too rarely and you miss regressions. Here's a few suggestions on when and how often to scan as a working starting point.

5.1 Cadence by scenario

5.2 AI Visibility is non-deterministic

AI engines don't always cite the same sources on the same query. A single "Not Found" on Perplexity for your keyword doesn't mean you're invisible - it could just be this run's roll of the dice. For AI Visibility specifically:

• Run at least 2-3 scans over a 1-2 week window for a reliable read
• Watch the Citation Timeline sparkline - trends matter more than single data points
• Don't re-scan mid-edit (you'll confuse cause and effect)
• For Mentioned → Cited work, fix the structure and refresh 2 weeks later so Google has time to recrawl and AI models have fresh content to pull

5.3 Chase movement, not scores

The absolute score matters less than the trend. A page moving from 47 → 68 over three months is winning faster than a page that hits 85 and stays flat. The History panel is the view to watch for this - it's the only way to tell whether your edits are doing anything.

5.4 What good looks like

Follow the signals OptimAI provides - we've carefully designed the tool to readily surface the highest-priority issues to the top. For most sites that have had at least some attention paid to SEO, when you first run a scan you'll probably end up with scores in the 45-65 range. This is normal, and in most cases it'll be because you haven't even begun to implement AI Search optimisation techniques (which OptimAI gives 40% of total score weight to, as that is the future). Don't panic.

As a general guide, you want to focus on getting the overall category scores into the green. You don't need to obsess about getting every dimension perfect, let alone every individual metric. That level of detail is really there to highlight what, if anything, is bringing down your overall score and holding you back. If you're stuck in red for FCP but the overall Performance category is 72, don't sweat it.

Remember, follow the Dimension Heatmap - it shows you which specific dimensions have the biggest impact, and those are the ones to focus on first.

Our score thresholds are set at 40 for fair and 70 for good, which we believe are honestly realistic. If you're getting red scores below 40, you really will be impacted and should sort that out. Getting over 70 is genuinely challenging, but that's precisely the level where you start to rise above the competition. Pass 80 and you're exceptional - that's the page that earns the traffic and the AI citations. It absolutely is possible to get green across the board - work through the suggestions systematically and you'll see your scores climb.

FAQ

Contact Support

Got a question, a bug to report, or a feature request? Send us a note. We reply from a real inbox, usually within a working day.