E-commerce is changing

Is your brand on AI’s shortlist?

Buyers now ask AI assistants what to buy. See how assistants and shopping agents read your store. Get a score, the fixes ranked by impact, and a way to track it.

Run the Free Audit Get the free playbook ↓

Agentic Storefront by ForkPoint · No account for your score · No card

Best trail running shoe under $150?

Assistant

Here are my top picks:

  1. 1 Ridgeline In stock · $139
  2. 2 Northpeak In stock · $129
  3. 3Your brand?Not listed
Not on the list means no click, no session, no sale.

Free playbook

Shortlisted. The new e-commerce playbook.

Fourteen pages on how to get your store on AI’s shortlist, and how to keep it there.

Cover of the Shortlisted playbook
  • The last 90 days: agents that buy, and checkout switched on by default
  • Three simulated answers and six stores that missed the shortlist
  • The eight things agents check, with a quick win for each
  • A 30-day plan and a printable worksheet

Get the playbook

PDF, 14 pages. The download opens right after you send the form.

  • +393% AI-referred retail traffic, Q1 2026
  • +54% Conversion vs. non-AI traffic
  • $1T U.S. agentic commerce by 2030

Adobe Analytics, Q1 and May 2026 · McKinsey, 2025

01 / The blind spot

The decision happens
before the visit.

A buyer asks an assistant for the best option. The assistant answers with one to three names. If your brand is not one of them, the buyer never arrives. There is no bounce to investigate and no campaign to blame. Analytics records the absence as nothing at all.

An assistant can only recommend what it can reach, read, and trust. An agent can only buy from a store where it finds a way in. Those are properties of your store that your team can check and fix.

Platforms are opening the channel too. Shopify, for example, makes eligible products available to AI channels through its catalog. Eligibility and data mapping decide what each channel can use.

Read Shopify’s product discovery guidance ↗
  1. 1A buyer asks an assistant
  2. 2The assistant shortlists one to three names
  3. 3An agent checks price, stock, and policies
  4. 4The buyer chooses, or the agent buys

02 / See your store the way an agent does

What a shopper sees
is not what an agent gets.

Many agents read the HTML your server sends first and skip JavaScript. A price that loads late is a price they never see. The same goes for stock and for product data that lives only in scripts.

What a shopper sees

Your brand

Summit Trail 2

$145

● In stock

Add to cart

What the agent receives
<!-- /products/summit-trail-2 -->
<h1>Summit Trail 2</h1>
<span class="price"></span>  No price
<div class="stock"></div>  No stock
<!-- no Product JSON-LD -->  No schema
<script src="/app.js"></script>

Illustrative example. Your report shows the evidence from your own pages.

03 / What the audit checks

223 checks across 8 categories.

The audit tests each step, from your robots rules to your checkout signals. It runs every check on every scan, free plan included. The method is public, and the scanner behind it, Agent Lighthouse, is open source.

  1. 01

    Access & Crawl Control

    Can AI crawlers reach your pages?

  2. 02

    Content Extraction

    Can an agent read your product content without running your JavaScript?

  3. 03

    Machine Discovery

    Sitemaps, feeds, and llms.txt that tell agents what you sell and where.

  4. 04

    Structured Data

    Product, offer, and organisation schema that models quote as fact.

  5. 05

    Answer Readiness

    Copy an assistant can lift into an answer: specs, prices, policies.

  6. 06

    Agent Interfaces

    MCP, OpenAPI, and well-known endpoints that let agents act, not only read.

  7. 07

    Agentic Commerce

    UCP and ACP signals that let an agent check stock and buy.

  8. 08

    Agent Operability & Safety

    Errors, speed, and safe defaults when an agent works through your site.

How each check is scored ↗

04 / A report your team can act on

Get a score.
Fixes ranked by impact.

Enter a URL and the audit reads the page the way an agent does. You get one score for the store and one per category, so you know where you stand and what moved.

Each failed check shows the evidence found on your page, so your team can confirm it in minutes. Every finding says what to change.

  • 90+ Agent Ready
  • 70–89 Partially Ready
  • 50–69 Needs Work
  • Below 50 Not Ready
Check your store Open a real sample report ↗

Readiness report

Example report

62/ 100

Needs Work

223 checks · 8 categories

  • Access & Crawl Control 88
  • Content Extraction 41
  • Machine Discovery 70
  • Structured Data 55
  • Answer Readiness 64
  • Agent Interfaces 20
  • Agentic Commerce 35
  • Agent Operability & Safety 81

Top fixes, by impact

  1. Product pages ship no Product schema High

    No JSON-LD Product or Offer found on the scanned page.

  2. Prices render only after JavaScript runs High

    The server HTML has no price; agents that skip JS see none.

  3. robots.txt blocks GPTBot Medium

    Disallow: / under User-agent: GPTBot.

05 / Track it over time

One audit is a snapshot.
Keep watching.

Releases, theme changes, and new bot rules can undo a fix without anyone noticing. Rescan after each change and watch the score move.

  1. 01

    Run the audit

    Start with a key product page. See the score and its tier.

  2. 02

    Fix the top finding

    Give it an owner. The evidence shows exactly what to change.

  3. 03

    Scan again

    Confirm the fix held on the same page and see what moved.

  4. 04

    Keep it there

    Scheduled audits and score alerts catch the drop before an assistant does.

06 / Built for commerce teams

One report. Every owner.

Commerce leaders

See where the store stands and which fix moves it first.

Technical teams

Trace each finding to the page, template, or data source that owns the fix.

Agencies

Audit client stores and turn the findings into a scoped proposal.

07 / Go a little deeper

Practical guides.

01

E-commerce is changing

What AI-assisted research means for your product pages.

Read the guide → 02

Product facts that help buyers

A worked example for price, variants, stock, and policies.

Read the guide → 03

Read a readiness report

Turn one finding into a task and a repeat check.

Read the guide → 04

Build your business knowledge catalog

Clear prices, real customer answers, and facts AI tools can read.

Read the guide →

08 / Questions

Before you start.

Agentic Storefront is a ForkPoint product. It is separate from Shopify’s feature named Agentic Storefronts.

Is the audit free?

Yes. The Free plan runs every check on every scan, with a monthly scan allowance. Paid plans add volume, scheduled audits, competitor benchmarks, PDF export, and automation. Check current plans and limits on Agentic Storefront.

Do I need an account?

Not for your score. Enter a URL and you see the score without signing up. A free account opens the full report: every finding, its evidence, and how to fix it.

How does the score work?

Each scan gives a score from 0 to 100 and a tier: Agent Ready at 90 and above, Partially Ready from 70, Needs Work from 50, and Not Ready below that. The score measures how your page presents itself to AI agents. It does not measure search rankings or how often assistants recommend you.

Can ForkPoint fix what the report finds?

Yes. Every engagement starts from your report: an expert review that costs each fix and puts them in order, then structured data and content work, protocol implementation for agent checkout, monitoring, and managed upkeep.

Do I need to change platform?

No. The audit reads the store you have. We implement against Salesforce Commerce Cloud, Shopify Plus, BigCommerce, Adobe Commerce, and headless stacks.

Free Agentic Readiness Audit

See what AI agents see.
Then fix it.

Run the free audit for your score and the fixes that matter most.

Run the Free Audit Get Expert Help

223 checks · No account for your score · No card