Vikray · agentic commerce readiness for enterprise retail

AI is shopping. Can it find, compare and buy from you?

Shoppers now ask ChatGPT, Gemini and Perplexity what to buy. AI shopping agents need reliable access, product data they can trust and a checkout path they can complete. Most enterprise storefronts aren't ready. Vikray finds the gaps and gives your teams a practical route to fix them.

Request your CommerceGrade assessment Complimentary · 48 hours · no system access required
Product research is shifting from Google to AI assistants +31% conversion on AI-referred traffic (Adobe) Agent checkout is live in the US, UK to follow (Google, May 2026) Enterprise readiness takes quarters, not weeks

The problem

Agents don't browse. They test you, and move on in milliseconds.

A shopping agent doesn't see your campaign imagery or your brand story. It reads your structured data, tries to satisfy a real shopper's request, and either includes you or moves on. We ran live tests on major UK storefronts: agents could often find products, but stalled when a request got specific (blocked crawlers, unreadable catalogues, checkouts they can't finish), and could not complete a purchase. When they stall, they refer the shopper to a competitor who works.

$ agent fetch /robots.txt
✗ GPTBot: Disallow / catalogue invisible to ChatGPT
$ agent fetch /.well-known/ucp
✗ 404, no declared commerce capabilities
$ agent parse /p/galaxy-s26-ultra
✗ Offer.price missing · availability stale (6h)
$ agent checkout --token=vcn_•••
✗ Declined: fraud rule BOT_SUSPECT_03
$ agent select merchant
✓ competitor-b.com added to recommendation

Blocked at the door

Bot rules written years ago now block the crawlers that feed AI answers and shopping indices. You can't be recommended by a model that can't read you.

Unreadable to machines

Specs in PDFs, prices that drift from feeds, broken JSON-LD. To an agent ranking on data quality, ambiguity equals exclusion.

Revenue declined as "fraud"

Legitimate agent purchases arrive on tokenized cards from datacentre IPs, and your fraud engine declines them. The leak never shows up in any report you currently run.

See where your storefront stands

Complimentary external assessment · results in 48 hours

What "ready" means

Readiness is a ladder, not a switch. The rungs are diverging fast.

RUNG 1 · NOW

Discoverable

AI assistants can find, read and correctly represent your products when shoppers ask. Independent of checkout, and actionable today.

RUNG 2 · HAPPENING NOW

Navigable

Browser-driving agents are already clicking through your human storefront to build carts. They get there by brute force, so it breaks when your pages change, trips your bot defences, and can't be trusted or measured. Most estates aren't built for this traffic today.

RUNG 3 · LEAD-TIME

Transactable

Agents transact through a supported path, not by puppeting your UI: verified agent identity, a real cart handoff, and SCA-compliant payment for the UK and Europe. The robust end state, and the part that takes longest to build.

An agent clicking "add to bag" once is not the same as your storefront being ready for agents to transact reliably and securely. The market is early. The lead time is not. Brands that prepare now are ready when it matters; the rest start from zero.

How it works

From complimentary scan to agent-ready, in three stages.

Every stage is anchored to one number: your CommerceGrade, a 0–100 score of how ready your storefront is for AI agents, re-measured as you improve. Competitive benchmarking is introduced as the dataset grows.

STAGE 1 · FREE

External scan

We test your estate from the outside, exactly as an agent would: crawler access, UCP/ACP presence, structured-data validity, latency, checkout friction. No access, no install.

48 hours · scorecard + 3 headline findings
STAGE 2 · AUDIT

Full readiness audit

35 checks across six pillars, inside and out: data pipelines, fraud rules, PSP configuration, SCA strategy, measurement. Scored 0–100 with a prioritised roadmap.

3–4 weeks · roadmap + exec readout
STAGE 3 · ENABLE

Enablement

We run the fix list with your teams: protocol integration, agent-traffic analytics, fraud-rule tuning, SCA-ready payment flows. We get you ready, not just scored.

Ongoing · watch your grade climb

The audit

Six pillars. One score. No place for failures to hide.

The six pillars carry a weighted share of a single 0 to 100 CommerceGrade. The points below show each pillar's maximum, and together they add up to 100.

15 PTS

Discovery

Crawler access, bot-management false positives, UCP profiles, agent-surface enrolment, AI answer baseline.

20 PTS

Product data

Structured-data coverage, price and stock truthfulness, feed quality, spec machine-readability, multi-market parity.

15 PTS

Infrastructure

Latency under machine-speed load, rendering dependence, commerce APIs, verified-agent fast lanes.

20 PTS

Checkout & payments

ACP/UCP capability, friction inventory, tokenized payment acceptance, PSP flexibility, SCA strategy for agent-initiated payments.

15 PTS

Trust & governance

Agent verification, fraud rules vs. agent traffic, mandate evidence for disputes, hallucination monitoring, ownership.

15 PTS

Measurement

Agent-traffic segmentation, agent-assisted conversion and revenue, decline telemetry, competitive benchmark.

The complimentary scan covers what an agent sees from outside. The full audit goes inside, to the fraud rules, PSP configuration and data pipelines where the access-gated problems live.

Why now

Agent checkout is live in the US and heading for the UK. The window to prepare is open now.

SEP 2025

OpenAI and Stripe launch ACP, the first live agent-checkout standard, in ChatGPT.

JAN 2026

Google launches UCP with Shopify, Walmart, Target and the card networks. Agent checkout goes live for US merchants.

MAY 2026

Google confirms the rollout path. UCP checkout expands to Canada and Australia in the coming months, with the UK to follow. The rest of Europe is not yet scheduled.

UK: NEXT

The UK is named and in the queue, but not yet live. This is the preparation window. Brands that are ready when it lands get recommended first; the rest start a multi-quarter project from zero.

TODAY

Readiness takes lead time. In a complex enterprise stack, the catalogue, payments, fraud and compliance work runs into quarters. The clock started in May.

The European blind spot

Agent payments meet European banking law.

The agent-checkout standards are largely US-built and say little about Strong Customer Authentication. In the UK and Europe, when a shopper isn't present in the session to approve a payment, many agent-initiated transactions will fail or fall foul of SCA and PSD2 unless your PSP configuration, fraud rules and authentication flows are deliberately prepared for them. This is underexplored territory, and it's where our enterprise payments background sits. We map the US-built protocols to UK and European compliance, so the transactions complete legally when the rails arrive.

It is moving fast. From 15 July 2026 the FCA brings Buy Now Pay Later into regulation under the Consumer Duty, with affordability checks and clear pre-contract information required. AI assistants are already capable of selecting a Pay Later option at checkout, and most shoppers say they would let them. When an agent picks a now-regulated credit product on a customer's behalf, who controls eligibility, disclosure and consent in a flow with no human reading the screen? That is no longer theoretical, and it is exactly the kind of question your checkout has to answer.

Check your readiness

See where you stand across discovery, checkout and payments

Why Vikray

Built for the merchants the toggle-switch solutions can't help.

Enterprise-complex by design

Shopify merchants flip a switch. You run custom stacks, multiple regions, multiple PSPs and legacy fraud rules. That's the only kind of estate we audit.

UK & EU payments depth

The protocols are American-built and largely silent on SCA and PSD2 when the cardholder isn't in the session. We're not. Agent payments in Europe is our home ground.

Independent and cross-platform

We work across your stack and your vendors, not for one of them. The platforms will make their own merchants agent-ready. We make you ready, wherever you're built.

Complimentary assessment

Find out where you stand before your competitors do.

Tell us your storefront domain. Within 48 hours you'll have your CommerceGrade scorecard and the three findings that matter most.

What you'll get back:

  • Your CommerceGrade (0–100) across the six pillars
  • Pass/fail on the external checks: crawler access, structured data, latency, cart handoff, checkout
  • The three highest-impact gaps, in plain terms

Confidential and never shared or published. Delivered by senior commerce engineers, UK-based.