Getting started with AIVerdict

How to run your first audit

Two paths: run a free check on any URL to see whether AI bots can access your site (no account required), or create a free account and buy a paid audit ($29) for the full 10-dimension diagnosis — both what AI can see on your site and what AI actually says about you elsewhere.

1. Quick start — the free check (no account)

   Go to aiverdict.io, paste your site URL, and click Run Free Check. You'll get scores on two free CORE dimensions in 10–30 seconds — plus a preview of where your other scores would land. The remaining 8 dimensions are visible in the report but locked until you spend a $29 credit.
2. What the free check tells you

   Training Access — whether AI training systems can build long-term memory of your brand. When training pipelines can't reliably read you, future versions of ChatGPT, Claude, Gemini, and Perplexity simply won't know you exist. Your competitors will. Recovering takes months after fixes ship, because training data refreshes on long cycles.

   Agent Access — whether live AI assistants can actually use your site when a real user sends them. If they can't, you're invisible in real-time queries: someone asks ChatGPT or Perplexity to look up something on your site, and they either get nothing back or get incorrect information. That hurts immediately, every single query.

3. Create a free account + buy a paid audit

   Sign up free (no payment method required — just email or a social login) to save your audit history. When you're ready to run a paid audit, you'll be sent through Stripe to purchase a credit. A single audit is $29; packs of 5 ($99), 20 ($299), or 100 ($999) save up to 66% per audit. Credits never expire — spend them whenever you need an audit.
4. What the $29 audit unlocks

   The paid audit scores all 10 dimensions split into two halves:

   CORE — onsite signals. How prepared your site is for AI extraction, positioning, and trust. The two free dimensions (Training + Agent Access) plus Content Quality (does your content give AI enough to actually say something specific about you?), Answerability (when someone asks AI a question your site could answer, can AI lift a clean answer from your pages?), and Extractability (when AI does cite you, how confidently and accurately can it pull facts out — do you get verbatim quotes or vague paraphrases?).

   AURA — offsite signals. What AI actually says about you right now when probed against ChatGPT, Gemini, and Perplexity: AI Brand Recognition (do they know who you are?), Share of Voice (how often you appear on category queries vs competitors), Citation Depth (when AI cites you, does it cite your own pages or only third-party content?), and Competitive Rank (when AI lists vendors, where do you sit?).

   Alignment deserves its own line. AI builds your brand picture by synthesising your own site copy with everything third parties say about you — reviews, press, forum threads, comparison posts. When those narratives agree (you call yourself a fast-onboarding CRM for agencies AND reviewers describe you the same way), AI confidently surfaces that story and recommends you with conviction. When they disagree (you market "fast onboarding" but reviewers say "took us three weeks to migrate"), AI gets contradictory signals — and tends to default to the third-party framing, hedge with caveats, or refuse to characterise you at all. Your own messaging effectively gets overridden by what others say about you.

   Risk is a modifier that subtracts from your AURA score when AI surfaces controversies, lawsuits, security incidents, regulatory action, or sustained complaint patterns about your brand. This matters because AI assistants are explicitly safety-tuned: they avoid recommending things tied to harm. When negative signal shows up in training data or live grounding, AI doesn't just lower your score abstractly — it changes how it actually answers. AI may refuse to recommend you, attach cautionary caveats ("note that [Brand] has faced criticism for…"), or steer the user toward a competitor framed as the safer choice. Recent incidents weigh heaviest; older ones decay as new training data dilutes them.

5. Each paid audit = 3 runs over 120 days

   One credit unlocks a 120-day bundle with three audit slots — not just a single snapshot.

   Slot 1 · Baseline — runs immediately when you spend the credit. Full 10-dimension score, per-dimension recommendations with predicted score impact (so you can sequence work by ROI), the prioritized fix plan, and a complete diagnosis of why each dimension scored where it did.

   Slot 2 · Verify fixes — fire any time within 120 days from your dashboard after you ship onsite changes. CORE dimensions (Training Access, Agent Access, Extractability, Content Quality, Answerability) reflect changes within hours after deploy, so this slot is for verifying the fix actually moved the score.

   Slot 3 · Drift re-audit — fires automatically at day 60 (no action needed). Re-probes ChatGPT, Gemini, and Perplexity on the same queries to catch how your AURA shifted. Offsite signals move slowly: AI doesn't notice your new Wikipedia article, press coverage, or Common Crawl entries for weeks to months — training data refreshes on long cycles and live-grounding indexes update on their own schedule. Slot 3 catches those shifts without you spending another credit.

6. Manage everything from your dashboard

   Every audit lives in your My Account dashboard:

   • Prioritized fix plan with score-impact estimates per recommendation — sequence work by ROI
   • White-label PDF export with your accent color and logo — ready to deliver to clients
   • 30-day shareable link that anyone can open without an AIVerdict account (pinned reports never expire)
   • Automatic 90-day drift re-audit bundled with every paid audit at no extra cost
   • Bulk ZIP export, audit history, site groups, and competitor analysis

Understanding your score

Two rings, ten dimensions, one Risk subtractor. Your overall score is the geometric mean of CORE and AURA — both halves have to work for AI to actually recommend you. Each dimension is a 0–100 score weighted by site type (SaaS, e-commerce, news, local, blog, generic).

CORE — can AI extract from your site?

• Training Access Free
  Does robots.txt allow AI training crawlers like GPTBot, ClaudeBot, PerplexityBot, Google-Extended, CCBot, Bytespider, Applebot, Amazonbot, FacebookBot, and Cohere? Asymmetric vs cookie walls (training crawlers must read content uncloaked).
• Agent Access Free
  Can a live AI assistant (ChatGPT browse, Perplexity, Gemini live grounding) reach and render your pages? Tests JavaScript dependency, cookie/auth walls, response speed, and broken redirects. No robots.txt check — live agents ignore it.
• Content Quality Paid
  Can AI extract specific, citable facts? An LLM-based reviewer evaluates clarity, citation readiness, competitive positioning, and topical depth across 8–12 sampled pages. Penalises buzzword-heavy marketing prose; rewards concrete claims and data.
• Answerability Paid
  How well can AI extract direct answers? Scores definition clarity, structured lists, tables, FAQ markup, and how-to steps. The exact sub-dimension names vary by site type (SaaS, e-commerce, news, local, blog, generic) so the score reflects what AI actually looks for in your category.
• Extractability Paid
  How reliably AI can pull specific facts, claims, and entities out of your pages. High extractability = AI quotes you verbatim and confidently cites your URL. Low extractability = AI paraphrases vaguely, attributes facts to competitors, or refuses to cite you at all even when it knows you exist.

AURA — what does AI say about you?

All AURA dimensions run as live brand-blind probes against ChatGPT, Gemini, and Perplexity on category-defining queries you'd actually be asked. Brand-blind = the prompt never names your brand — AI surfaces you (or doesn't) on its own.

• AI Brand Recognition Paid
  Does AI already know who you are? Tests whether ChatGPT, Gemini, and Perplexity can describe your brand without grounding (no live web search) and whether their description matches reality. Internally tracks Wikipedia presence, Common Crawl indexation, and reviews as supporting evidence.
• Share of Voice Paid
  How often you appear in AI answers vs competitors on category-defining queries. Probes 10 user-realistic prompts across 7 intent buckets (compare, recommend, explain, troubleshoot, integrate, price-research, alternatives) per audit and counts brand mentions vs the named competitors.
• Citation Depth Paid
  When AI cites you, does it cite your own pages or only third-party content about you? High citation depth means AI quotes your docs, blog, or product pages directly — you control the narrative. Low depth means AI is reading what others say about you.
• Alignment Paid
  Does third-party content (news, reviews, forum threads) mirror your brand's positioning? If your site says "we're a fast-onboarding CRM for agencies" but reviews say "buggy, hard to set up", AI synthesises both and the misalignment shows up here. Alignment is not a sentiment score — it's a positioning-match score.
• Competitive Rank Paid
  Where AI ranks you when forced to choose. Probes prompts like "best X for Y" and "compare X vs Y" and records your ordinal position vs category leaders. A brand can have high Share of Voice but low Competitive Rank if AI mentions you but always last.

Risk — the modifier

• Risk Modifier Paid
  Subtracts from your AURA score when AI surfaces controversies, lawsuits, security incidents, or trust issues about your brand. Time-decayed — recent incidents weigh more than old ones. Risk is not a 6th AURA dimension; it tints the AURA ring orange and reduces the AURA composite score.

Want the full math? Each dimension's signals, weights, and edge cases are documented on its dedicated page. Start with Training Access for the free-tier check, or Share of Voice for the AURA flagship.

Reading your audit report

The report opens with two side-by-side rings — the blue CORE ring and the emerald AURA ring — with the Overall geometric-mean badge between them. Below that, each dimension is its own expandable card.

• circle The two rings. CORE (blue) sums the 5 onsite dimensions; AURA (emerald) sums the 5 offsite dimensions. The Overall pill between them is the geometric mean — if either half is weak, the overall drops faster than a simple average would suggest.
• unfold_more Click any dimension to expand. Each card opens to show its sub-checks, raw signals, and the per-test pass/fail. AURA dimensions show the actual prompts that were probed and the verbatim AI responses that fed the score.
• check_circle Green = passing, amber = partial, red = blocking. Red findings are the highest-impact fixes — they're the ones blocking AI from seeing or recommending you. Amber means improvements help but won't unlock visibility on their own.
• route The fix plan ties findings to score impact. Each recommendation includes a "+X CORE" or "+X AURA" estimate so you can sequence work by ROI. CSV export and white-label PDF available on every paid audit.
• auto_awesome AI Summary at the top. A plain-language overview of your strongest areas and the highest-impact improvements you can make — written so you can paste it into a Slack thread or client email.

Common actions after your first audit

Focus on the highest-impact fixes first. Here are the most common improvements, ordered by typical score lift.

• priority_high Unblock AI bots first. If your Training Access score is low, robots.txt is blocking GPTBot, ClaudeBot, PerplexityBot, or others. AI training crawlers literally can't read your content while they're blocked. Open the AI Bot Directives section of your Training Access result and update robots.txt.
• code Unblock live-agent fetches. If Agent Access is low, real-time AI queries silently fail on your pages — users asking ChatGPT or Perplexity about your site get nothing useful back. The fix is usually infrastructure-level (server, CDN, response speed, anti-bot rules) rather than content-level.
• data_object Add FAQ + structured-data markup. FAQPage, HowTo, and Product schema directly increase Answerability scores and the rate at which AI cites you in generative answers. Schema is one of the fastest CORE wins.
• edit_note Replace buzzwords with specifics. Content Quality penalises "empower your team" and rewards "trusted by 12,000 companies, 99.9% uptime, 200ms response". Concrete claims = citable claims = AI mentions.
• public Build offsite presence to lift AURA. CORE fixes are the floor; AURA is the ceiling. Wikipedia presence, press coverage, review-platform listings (Trustpilot, G2, Capterra), and developer-doc citations all feed AI Brand Recognition, Citation Depth, and Alignment.
• refresh Re-audit after 48–72 hours. Wait long enough for CDN and Cloudflare caches to clear. Every paid audit also auto-fires a 90-day drift re-audit at no extra cost so you can verify offsite improvements without spending another credit.

Your account dashboard

After signing up (free, no credit card), your My Account dashboard is where every audit lives. There's no "Pro" or "Agency" tier — you just spend audit credits whenever you need them.

• search Run Audit — Enter a URL, spend a credit, and watch all 10 dimensions populate in real time. The split layout shows the URL bar pinned on the left while the report renders on the right; click any dimension dot to jump to its section.
• history History — Every audit you've ever run, sorted by recency. Filter by source (manual, drift, competitor), assign to a Site Group, click any row to reopen the full report. Lite Checks (free CORE) and full audits live side by side.
• timeline Bundle Drift — Every paid audit auto-fires a re-audit at day 60 (the 90-day window) so you can see how your CORE + AURA shifted after fixes. Zero extra cost — comes bundled with the original credit.
• compare_arrows Competitor Analysis — Spend 0.5 credits to audit a competitor on the same 10 dimensions. Always-fresh (no cache); side-by-side CORE / AURA / Risk comparison, plus a delta column.
• picture_as_pdf White-label PDF — Every paid audit unlocks the branded PDF export. Set your accent color and logo once in the White Label tab; every PDF you generate after that uses your branding. Bulk-export multiple audits as a ZIP from History.
• folder Site Groups — Organise audits by client, project, or category. Color-coded groups; assign audits via dropdown; up to 50 groups per account. Useful for agencies managing dozens of clients.
• explore Niche Explorer — Spend 0.5 credits to test how AI describes a custom niche / category before running a full audit. Useful for validating which "best X" prompts your customers are actually asking AI.

To access your dashboard, sign in or create a free account — no credit card required to run the free access check.

Pricing — pay per audit

No subscription. Buy audit credits once, spend them whenever you need to audit a site. Each audit you spend a credit on includes 3 runs over 120 days for that domain (baseline · verify fixes · drift check).

Every paid audit includes the full 10-dimension CORE + AURA score with the Risk modifier, live brand-blind probes against ChatGPT, Gemini, and Perplexity on category-defining queries, a prioritized fix plan with score-impact estimates per dimension, a white-label PDF + 30-day shareable link, and an automatic 90-day drift re-audit at no extra cost. Credits never expire. You pick the website to audit on your dashboard after checkout.

Want to try first? The free access check on the homepage scores Training Access and Agent Access on any URL, no signup. The other 8 dimensions are visible but locked.

Frequently asked questions

• What's the difference between CORE and AURA? expand_more
  CORE measures whether AI can access, render, and extract from your site — 5 onsite dimensions: Training Access, Agent Access, Content Quality, Answerability, Extractability. AURA measures what AI actually says about you on real user queries — 5 offsite dimensions: AI Brand Recognition, Share of Voice, Citation Depth, Alignment, Competitive Rank. A Risk modifier subtracts from AURA when controversies surface. The Overall score is the geometric mean of CORE and AURA — both halves have to work.
• How long does an audit take? expand_more
  The free access check (Training Access + Agent Access) completes in 10–30 seconds. The full $29 audit — all 10 dimensions including the live brand-blind AURA probes against ChatGPT, Gemini, and Perplexity — takes 60–120 seconds depending on site size and server response time. A live timer shows elapsed time during the run.
• Which AI systems do you probe? expand_more
  AURA runs live brand-blind probes against ChatGPT, Gemini, and Perplexity. We also check robots.txt allowance for 10 training crawlers (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, CCBot, Bytespider, Applebot, Amazonbot, FacebookBot, Cohere) as part of the Training Access score. The AURA probes themselves don't include Claude as a probe target — the Anthropic chat product doesn't have a public live-grounding API the way ChatGPT/Perplexity/Gemini do.
• How often should I re-audit? expand_more
  After making fixes, wait 48–72 hours for caches to clear before re-auditing. Every paid audit also includes an automatic 90-day drift re-audit at no extra cost, so you can verify offsite improvements without spending another credit. You can also re-audit any time by spending another credit — useful after a major site redesign, CMS migration, or robots.txt change.
• Why is my score different from last time? expand_more
  CORE scores reflect your live site state — they change when you ship fixes, add structured data, or update robots.txt. AURA scores reflect what AI is saying about you right now — they can change because competitors moved, news/Wikipedia/reviews changed, or AI training data shifted. Risk drops scores when negative coverage appears in AI responses. Some AURA noise is normal between probe runs — the drift re-audit at day 60 helps separate signal from noise.
• Can I audit competitor sites? expand_more
  Yes — any publicly accessible URL. The bundled Competitor Analysis feature adds 0.5 credits per competitor and runs the same 10-dimension audit on their domain so you can compare CORE + AURA + Risk side by side. Always-fresh (no cache reuse). See Competitor Analysis for details.
• What does "crawler blocked" mean? expand_more
  Your site's infrastructure (Cloudflare bot protection, a WAF, or cookie walls) is preventing our audit crawler from reaching your content. This doesn't always mean AI bots are blocked — check the AI Bot Directives section in your Training Access result for the exact allow/deny status of GPTBot, ClaudeBot, PerplexityBot, and others. Cookie walls hit Training Access especially hard because crawlers can't pass them; Agent Access is more forgiving since live agents can sometimes navigate them.
• How do I export my audit as PDF? expand_more
  Every paid audit includes a white-label PDF export button on the report page. The PDF can be branded with your own accent color and logo (set once in the White Label tab) and contains all 10 dimensions, the prioritized fix plan, and the AI summary. Bulk ZIP export of multiple audits is available from the History tab.
• Do credits expire? expand_more
  No. Audit credits never expire. Spend them whenever you need to audit a site. Each spent credit unlocks the full 10-dimension audit on that domain plus an automatic 90-day drift re-audit. Volume discount packs (5/20/100 audits) save up to 66% per audit and there's no subscription — you only pay when you actually need an audit.