What is an AI readiness assessment?

An audit measuring how easily a website can be read and understood by AI engines and agents. It scores criteria such as technical infrastructure, structured data, content architecture and brand presence.

How do you tell a website is ready for AI agents?

Four signals: can AI bots access the site (robots.txt, WAF), is the content readable without JavaScript, is the data structured with schema, and is the brand cited in AI answers. All four are measurable; the scan does exactly that.

What is the difference between AI readiness and classic SEO?

SEO targets ranking on a results page; AI readiness targets being the source of the answer itself. The latter covers layers SEO does not: agent access, structured data, entity clarity and citability. Good SEO is a good foundation, but not enough on its own.

Why does llms.txt matter for AI readiness?

llms.txt gives models a one-file summary of your site and its key pages, increasing the odds your content is read in the right context. It is a proposed standard — no guarantee of visibility, but low cost and zero risk.

What does the AI readiness score show?

The score is a weighted total of externally verifiable criteria: access, structured data, content readability, technical foundation, API and protocol readiness. It is a 0-100 value mapping to one of five tiers.

For businesses selling only domestically with no export goal and no product catalogue, part of what is measured has no payoff today. The scan still helps with access, schema and measurement; think twice about feed and protocol investment.

What does the score not guarantee?

Sales and traffic. Agentic commerce is primarily a discovery channel today; the score measures the preconditions for existing in that channel. A high score makes you selectable — it does not by itself make you selected.

Does the whole 186-item list feed the score?

No. Only the externally verifiable items produce the score. Contracts, authority thresholds and team training cannot be scanned; they stay manual on the checklist and cannot inflate the number.

I sell services, not products. Is the deep scan still for me?

Yes. Without a feed URL the system assesses you on a services/B2B profile: the product feed and freshness categories are dropped from the denominator and you are scored on access, schema, visibility and measurement.

Will I get the same score if I scan the same site twice?

Yes, the scan is deterministic: the same input returns the same score. That is why a post-fix difference is a real difference, not measurement noise.

Will the scan strain my site?

No. A limited number of pages are sampled, requests are spaced out and the scan is one-off. It produces less load than a normal visitor session.

What if my site blocks the scan?

That is a finding in itself: infrastructure that cannot be scanned by us cannot be scanned by AI agents either. The report then shows a low access score along with which layer is blocking.

Webtures/Services/AI READINESS ASSESSMENT

AI READINESS ASSESSMENT · 50+ CRITERIA · 2026

How ready is your brand for the Agentic Web era?

GPT, Claude, Perplexity and Gemini turned search into answer generation. We scan your site across 50+ criteria, prioritise the risks and hand you a 14-day action plan.

No sign-upYour data is not storedResults in ~60 secondsSame measurement every time

50+

Audit criteria

Readiness dimensions

AI models tested

14d

Roadmap

Why it matters · AR 001

AI no longer reads content — it picks sources.

Models compare and recommend on the user's behalf. The numbers show the structural shift between 2025 and 2026.

~58%

More than half of Google searches now end with zero clicks.

SparkToro 2024

800M+

Weekly active ChatGPT users; search became an answer-generation surface.

OpenAI

+1300%

Growth in referral traffic shifting from traditional SERPs to LLM surfaces.

SimilarWeb 2025

47%

Share of B2B buyers researching purchases through ChatGPT or Perplexity.

Gartner 2025

64%

Brands budgeting AI Search Optimization as a separate line in 2026.

Forrester 2025

3.2x

How much more often sites optimised with llms.txt and structured data get cited by LLMs.

Webtures Lab

Outcomes · 6 Tangible Results

What we deliver in the AI era.

Six outcomes that start where traditional SEO ends — each one measurable.

Visibility

Stay visible in generative AI answers

Your brand is cited as a source in ChatGPT, Perplexity, Gemini and Copilot answers; customers discover you, not your competitor.

Insights

Understand real AI agent behaviour

Beyond keywords: we analyse how AI agents perceive your content and surface intent, trends and opportunities.

LLMO

Make your content readable and quotable

From schema markup to contextual optimisation; we structure your content so models recognise and cite it.

Strategy

Go beyond classic SEO

Against declining organic traffic we build durable visibility with GEO and LLMO — a strategy that still holds tomorrow.

Tracking

Measure AI visibility, prove the impact

We track how often you appear in AI answers, which pages get cited and what converts; prompts and citations become clear KPIs.

Speed

Ship fast, stay ahead of change

The AI search ecosystem changes weekly. We iterate content, schema and entity signals quickly and test across models.

Service scope · 4 Pillars · 8 Dimensions

What we measure, and how.

We analyse the site across 50+ criteria from the AI's read–evaluate–pick-a-source perspective, not the SEO one.

01 · Access

AI Bot Accessibility

robots.txt, GPTBot/ClaudeBot/PerplexityBot permissions, firewall restrictions and dynamic content access.

Çıktı · Access Map

02 · Meaning

Content & Semantic Structure

Heading hierarchy, semantic HTML, structured data, entity clarity and citability.

Çıktı · Schema + Entity Score

03 · Action

Technical Infrastructure & Actionability

Server-side rendering, API access, form & checkout automation, machine readability of the data layer.

Çıktı · API + SSR Report

04 · Authority

AI Citation & Reference

Source-selection potential, per-model visibility, reference frequency and brand mention density.

Çıktı · Citation Index

018 checks

Structured Data

Schema.org coverage, JSON-LD, Org/Product/FAQ/Review, rich properties.

0211 checks

AI Discoverability

llms.txt, AGENTS.md, sitemap, robots, canonical, OG, meta description.

034 checks

Content Readability

Heading hierarchy, semantic HTML, alt text discipline, block structure.

044 checks

Technical Foundation

HTTPS, mobile viewport, Core Web Vitals, low error rate.

053 checks

API Readiness

Public API endpoint, OpenAPI documentation, CORS configuration.

063 checks

Agent Protocol

MCP manifest, function calling readiness, Plugin/Actions structure.

074 platforms

AI Visibility

How often GPT, Claude, Perplexity and Gemini source you in answers.

085 checks

Entity Authority

Knowledge graph clarity, external reference consistency, entity links.

Deep scan · Deterministic

For those who want proof: six categories, gate rules, same input, same score.

Access, schema, product feed, freshness, visibility and measurement are scored as one chain: a site with a flawless feed that blocks crawlers cannot score well. As the scan finishes, the verifiable items in the 186-item checklist below are ticked automatically.

Agent access Gate rule

OAI-SearchBot and Google-Extended in robots.txt, HTTP status for agent user-agents, JS-free visibility, sitemap.

20 pts

Structured data

Organization, Product and Offer schema, ISO 4217, availability enum, GTIN/MPN coverage, variant model.

25 pts

Product feed

Required fields, id uniqueness, GTIN ratio, price format, product link status codes.

25 pts

Freshness and consistency Gate rule

Sampled comparison of feed vs on-site price, feed reachability.

15 pts

AI visibility signals

llms.txt, target-market language and hreflang, fact density, FAQ and comparison signals.

10 pts

Measurement readiness

GA4 and GTM detection, pixel and conversion measurement signals.

5 pts

Gate rule: If agent access is below 10 points the total score is capped at 40; if feed and site prices disagree, 15 points are deducted outright.

Score	Tier	Meaning
0–25	Invisible	Agents cannot read the site or understand the product.
26–50	Accessible	Agents can read it, but the data is not trustworthy.
51–75	Feed-Ready	Data is reliable, but visibility and measurement are missing.
76–90	Agent-Native	The channel is live and measured.
91+	Reference	Mature, including governance and trust layers.

E-exporter

The technical answer to "I sell to the US — why don't I show up in ChatGPT?"

Agencies and freelancers

A fast, evidence-grade starting point for client audits and proposals.

In-house teams

A tier assignment and a prioritised first 90 days.

Who it is not for: For businesses selling only domestically, with no export goal and no product catalogue, part of what is measured has no payoff today — ChatGPT Ads does not target Türkiye and the first wave of agentic checkout was pulled back. The scan still helps with access, schema and measurement; think twice about feed and protocol investment.

Checklist · Auditable

12 dimensions, 186 items: what should be true, how to prove it, what it is worth.

Agents read data, not design. Every item states what should be true, its proof and its weight. Your progress is stored in your browser; the scan auto-ticks the 47 items tagged auto — the remaining 139 can only be verified from inside the company.

0 / 186 completed
0 / 431 points

Access → Readability → Data → Visibility → Protocol → Transaction → Measurement

Chain rule: Investment that skips this order is wasted — a site with a perfect feed but blocked bots is never found. Critical items are worth 3 points, important 2, maturing 1; 431 in total. Drop inapplicable sections from the denominator and evaluate by percentage. A B2B firm without a product feed can skip Parts 3 and 5; Parts 1, 2, 7, 8 and 9 apply to everyone.

Percent	Tier	Meaning	Next move
0–25%	0 · Invisible	Agents cannot read the site	Parts 1 and 2
26–50%	1 · Accessible	Readable, data untrustworthy	Parts 3 and 6
51–75%	2 · Feed-Ready	Data good, no visibility or measurement	Parts 8 and 9
76–90%	3 · Agent-Native	Channel live and measured	Parts 4 and 10
91%+	3+ · Reference	Mature incl. governance	Pilot leadership

Critical rule: A tier is not earned until a part's critical items are complete — with OAI-SearchBot blocked in robots.txt, the effective tier is 0 even at 80%.

0–30 days

Access and readability

All of Part 1; 2.1–2.12; 9.1–9.3 and 9.9 (baseline); 8.1–8.2 (prompt set and first measurement).

Output: A site open to bots with validated schema, plus a measurement baseline.

30–60 days

Data and reliability

All of Part 3; 6.1–6.4 and 6.7; 5.1–5.3 (unit economics); 12.1–12.2.

Output: A validated feed, 98% stock accuracy, channel-fit SKU decisions.

60–90 days

Channel, visibility, governance

4.1, 4.4, 4.5, 4.9; the rest of Part 8; 7.3, 7.5, 7.6, 7.9; Part 10 if exporting; Part 11 review.

Output: A live product on at least one agentic channel and a written authority policy.

If any of these is true, stop everything else:

WAF or robots.txt blocks legitimate search and agent bots.
Price or stock differs between feed and site.
Product information renders only via JavaScript.
Fake GTINs generated in the feed — invented barcodes are a permanent account risk.
Fake reviews or AggregateRating present.
Stock accuracy below 90%.
No logging at all for agent transactions.
Commission plus transaction fees exceed margin but the channel stays open.
Domestic budget allocated without knowing ChatGPT Ads does not target Türkiye.
Supplier contract is hourly with no verifiable milestones.

Not a sales guarantee. Agentic commerce is primarily a discovery channel today; the first wave of embedded checkout was pulled back. Invest in feed quality and visibility.
The Türkiye asymmetry is real. Parts 10 and some of 4 have no domestic payoff today; the value is in e-export markets.
Measurement is incomplete. A meaningful share of AI traffic cannot be captured by any setup; the numbers are a floor.
Protocols are moving. ACP, UCP and AP2 change quarterly; the list reflects the August 2026 state.
Treat industry statistics with caution. "AI traffic converts X times better" claims are mostly single-source; your own baseline is the only valid evidence.

Process · 3 Steps

Not a one-off check; continuous transformation.

Analysis

Current state audit

The site is reviewed across 50+ criteria: AI bot access, content architecture, structured data and technical infrastructure one by one.

Prioritisation

Risk & impact map

Gaps are classified by impact; the areas directly affecting AI visibility are clarified and action priority is set.

Execution

Action & roadmap

Concrete recommendations for each risk: technical work, content and strategic structuring delivered as a 14-day roadmap.

RELATED RESOURCES

Go deeper

Background reading for the checklist items on this page.

AI agents and the future of SEO → Why agent readiness is becoming the new visibility battleground. Align your GEO strategy with AI Mode → How answer engines consume the signals this scan measures. GEO strategies → The optimisation playbook behind the machine-readability items. Do product reviews affect visibility? → Trust signals and review schema from the commerce sections. Agentic Web Optimization service → Close the gaps with the Webtures team. Agent Experience Optimization → Designing your site for autonomous agents end to end.

FAQ · 12 Questions

Frequently asked questions

Get started · Free

Learn your brand's AI readiness score in the time it takes to drink a coffee.

We scan 50+ criteria, measure your GPT/Claude/Perplexity/Gemini visibility and prepare your prioritised 14-day action plan.

Start the Free Analysis Talk to a strategist

Copied to clipboard