InfiniteWatchAgent Ready Web
IW AGENT READY WEB

www.prima.it
MCP & WebMCP readiness.

Discover how easily AI assistants can find your business and help customers take action.

Run a fresh check. Updates are limited to once every 15 minutes.
www.prima.it19 documents checked16:27 UTC

Your website’s AI readiness

23 checks · 6 categories

We check public information without submitting forms or taking actions on your website.

Start the conversation with your team.

www.prima.it · 59/100 · 2026-09-24

LESS FRICTION. MORE MOMENTUM.

From browsing pages to getting an answer.

Illustrative scenario
4 steps1 guided action

Connect your useful information to a guided enquiry, product choice or service request.

See the journey assumptions
  1. Explore the website
  2. Find information
  3. Compare options
  4. Take the next step

The proposed action brings these steps into one guided interaction. This is an example, not measured performance of this website.

Explore time savings and calculation assumptions
CUSTOMER TIME

How much time could your customers save?

www.prima.it · Finding information

Based on this website
Estimated time saved
6.0seconds saved

per customer task

79.9% less estimated waiting
Estimated task duration

Page-by-page journey

7.5s

4 estimated steps × 1.9s per step

With a proposed AI shortcut

1.5s

1 estimated step × 1.5s per step

Compares an AI assistant navigating pages with a proposed direct action. These are projections; we haven’t timed the task or confirmed that a discovered tool can complete it.

Adjust task steps and timing

Match these assumptions to the same completed customer task in both journeys.

Page-by-page journey
With a proposed AI shortcut

Task duration = steps × seconds per step. Each step includes assumed processing and request time.

Technical detail: token use

Tokens estimate the amount of content the AI processes.

Estimated token difference per task8,336 fewer tokens

91.6% less token use.

Page-by-page journey 9,096 tokens

With a proposed AI shortcut 760 tokens

Page-by-page journey
With a proposed AI shortcut

Tokens per task = steps × tokens per step + one-time context setup. Token changes do not change the timing assumptions.

Scan evidence: about 2,274 readable context tokens and no verified tool schema tokens.

How this website’s estimate was built
Observed page request
0.67s
Controls found
14
Forms found
0
Tool definitions found
0
  • The browser model assumes readable text plus 24 tokens per control is sent on every step without caching. Raw HTML size is not used as the default context; agents may send smaller snapshots or diffs.
  • Token sizes are approximated from retrieved characters ÷ 4; no model tokenizer was run.
  • Browser steps assume navigation, 1 input decision, submission and reading. These steps were inferred, not executed.
  • No tool definitions were verified. The MCP side models a proposed tool with 300 schema tokens and one call.
  • Each step assumes 1.2 seconds of model processing. Browser network time uses this page’s HTTP request; tool time uses an editable 300 ms placeholder. Neither is a measured customer workflow.
  • Tool argument/result tokens use an editable 400-token allowance plus 60 tokens per modeled input. Cached context, retries and real output sizes can change savings.
EXPLORE MORE COMPANIES

See how other websites compare.

Explore other companies’ AI readiness reports. Search for a website or browse by industry to find ideas for your next customer journey.

Explore the directory ↗
YOUR BUSINESS OPPORTUNITY

Turn customer intent into action.

Customer journeys

Make my website agent-ready ↗
01

Help customers find answers and take the next step

Connect your useful information to a guided enquiry, product choice or service request.

What supports this opportunity

Proposed customer journey. Public signals guide the implementation.

Customer actions listed for AI · Not yet verified

Choose a useful customer task, such as finding a product or requesting a quote, and make it available to AI assistants.

Clearly explained actions · Not yet verified

Clearly describe each available action and what it helps the customer achieve.

Required information for each action · Not yet verified

Make it clear what information an assistant needs to complete each action, such as a product name or booking date.

Action safety information · Not yet verified

Explain which actions only read information and which make changes, so assistants can handle them appropriately.

02

Get discovered with the right answer

Help assistants understand your offer and point customers towards the next useful action.

What supports this opportunity

Proposed customer journey. Public signals guide the implementation.

Clear page title · Detected

Title: Assicurazioni online per auto, moto e casa - Prima.it

Search description · Detected

A descriptive meta description is present.

AI content guide (llms.txt) · Detected

A public llms.txt document was found.

Instructions for connecting AI assistants · Not detected

Provide instructions so compatible AI assistants can connect to your available customer actions.

Content labels AI can understand · Detected

Parseable schema.org JSON-LD found. Content accuracy requires editorial review.

03

Make the next step easier

Give customers and assistants a clear route from interest to an enquiry or purchase.

What supports this opportunity

Proposed customer journey. Public signals guide the implementation.

Clear page structure · Detected

An h1 and a main landmark are present.

Easy-to-follow navigation · Detected

74 labelled, crawlable link(s) found.

Clear button and form labels · Detected

No unlabelled controls detected in this HTML sample (3 input controls). This is not a full accessibility audit.

WHAT AI ASSISTANTS CAN USE

Actions AI assistants can find

Tools let AI assistants help with tasks such as finding a product or checking an order. These are the public integration signals we found; task completion still needs testing.

0Browser integration sources
—AI tools · count unverified
UnverifiedBrowser integration signals
Explore the tools and technical evidence

WebMCP connects AI assistants to actions in a browser. MCP lets compatible assistants connect directly to a service. The details below help your technical team evaluate both.

No compatible public MCP connection confirmed (HTTP 404). View endpoint ↗

No tool names were confirmed in this sample. Tools may load after login, navigation or JavaScript execution.

0 named source declarations found. Counts cover the sampled page and advertised endpoints. Source declarations are not runtime-tested tools; entries listed through more than one route may overlap.

WebMCP directory · webmcp.com

The webmcp.com directory lookup is unavailable. Listing status remains unverified.

Open webmcp.com ↗

This is a stored directory snapshot, not a live tool test. Its count may include browser-only tools and may differ from the public MCP server count. Directory listings do not change your score.

MCP Registry

No matching listing returned. This limited name search does not prove absence from the directory.

MEET AGENTS WHERE THEY ARE

Where can customers use AI assistants?

Support varies by browser and assistant. Some connections are still experimental; this is a platform guide, not a test of your website.

Reviewed Sep 2026
See browser support and connection options

Google Chrome

149+ · origin trial

WebMCP requires an eligible origin trial or the experimental testing flag. A compatible agent is also required.

Support details

Microsoft Edge

150+ · origin trial

Available for experimentation through the Edge origin trial. Check trial eligibility and current expiry dates.

Support details

Brave / Leo

Experimental integration

Experimental Leo integration is reported by the specification project. Availability is not guaranteed for all users.

Support details

Mozilla Firefox

No confirmed release

Prototype work is tracked, but no native release is confirmed. Use standard browser automation or a remote MCP client.

Support details

Apple Safari

No confirmed native support

Native WebMCP cannot be assumed. Regular browser automation and separate remote MCP clients remain alternatives.

Support details

Remote MCP clients

Browser-independent

Compatible agents can connect directly using a supported transport. Authentication and permissions still apply.

Support details

Remote MCP is a separate route. A compatible agent can use a remote MCP server independently of the visitor’s browser. Browser automation can still use regular pages. This matrix describes platform support, not a cross-browser test of your website.

UNDERSTAND YOUR SCORE

A clear starting point for your team.

Can AI assistants find your website, understand what you offer and help customers act? Your score brings these signals together and shows where a closer look is needed.

LESS WAITING. FEWER STEPS.

Let customers get things done through AI.

Bring quoting, booking and product discovery into AI-assisted customer journeys.

A walkthrough of your website Opportunities to reduce customer effort A practical plan for your team

Make my website agent-ready

See how InfiniteWatch can help your customers do more in less time.

A conversation about your use case. No obligation.