Paz.ai
Paz.ai is an agentic commerce platform for e-commerce brands and retailers. It monitors how products appear in ChatGPT, Google AI Mode, Perplexity and Claude, enriches catalog data, distributes optimized feeds and attributes the resulting orders.
What is Paz.ai?
Paz.ai is the trading name of AgentPass, Inc., a San Francisco company founded, according to the structured data on its home page, in 2024, and led by founder and CEO Dor Shany, previously the founder of Gigya (acquired by SAP) and OwnID. The product addresses a specific problem: shoppers increasingly start with an AI assistant rather than a search box, and most product pages expose only five to eight structured attributes where AI agents expect thirty or more.
The platform is organized around three named agents. The Monitor Agent benchmarks the catalog against competitors and finds the gaps. The Optimizer Agent rewrites titles, completes attributes and repairs structured data. The Distributor Agent publishes the resulting feeds. The surfaces monitored are ChatGPT, Google AI Mode and AI Overviews, Perplexity, Claude and Amazon Rufus; the publishing destinations are the OpenAI Product Feed, Google Merchant Center and, in the vendor's words, emerging AI commerce endpoints. Supported protocols include ACP (the Agentic Commerce Protocol, through Stripe), MCP, UCP and AP2.
One design principle runs through the product: the live store is never modified. Paz.ai works on an enriched copy of the catalog and publishes it to AI shopping agents. Changes go through a validation workflow the merchant chooses, from auto-ship to batch approval to product-by-product review. Attribution then follows the visit from landing page and page views through checkout steps to the final order, by visit and by channel.
The entry point is a free AI Readiness Report: a score out of 100 across four layers (Product Mapping, Structured Attributes, Attribute Context, Product Context), returned in about thirty seconds. For developers, the company publishes an unusual amount of machine-readable material, including openapi.json, .well-known/mcp.json, a server card, an RFC 9727 api-catalog, auth.md, pricing.md, llms.txt and llms-full.txt.
The market figures on the site come from third parties: Adobe's Q1 2026 AI Traffic Report (+48% time on site, +50% conversion for AI-referred visitors) and Shopify's Q1 2026 Commerce Trends (+14% average order value). Time to production is quoted at two weeks. The product page claims SOC 2 compliant infrastructure, a claim not backed by any published report or trust portal.
What it does
- Measure product and brand visibility across ChatGPT, Google AI and Perplexity
- Benchmark your position against named competitors on shopping queries
- Detect the missing product attributes that block a recommendation
- Rewrite titles and descriptions, complete attributes and repair structured data
- Generate a feed in OpenAI's commerce format and publish it to Google Merchant Center
- Attribute visits, carts and orders back to the AI engine they came from
When to use Paz.ai / When not to
A quick filter to help you decide if Paz.ai is the right fit.
When to use Paz.ai
- E-commerce brands and retailers running a Shopify, Adobe Commerce/Magento or Salesforce Commerce Cloud catalog
- Teams that already see traffic arriving from ChatGPT or Perplexity and have no way to measure it
- Merchants with catalogs of roughly 500 to 5,000 products, which fit the published Starter and Growth tiers
- Merchandising teams that want to industrialize product attribute enrichment, where AI agents expect 30+ structured attributes per product
- Specialty retailers of the kind already in production on the platform, such as New York occasion-wear house The Dessy Group since April 2026
When not to use Paz.ai
- Businesses looking for a storefront: Paz.ai runs no cart, no checkout, no payment and no fulfillment
- Merchants expecting the tool to set prices or handle customer service on orders
- Consumers and individual shoppers, since this is a B2B product sold to commerce teams
- Buyers who need a permanently free tier, as entry pricing starts at USD 119 per month
- Users under the age of 16, who are not permitted to use the service under the terms of use
How to use Paz.ai
A typical end-to-end flow, from setup to results.
- Test first without an account: paste a product page URL into the free AI Readiness Report and get a score out of 100 in about thirty seconds, no credit card required
- Create an account and start the 14-day trial at app.paz.ai/register, or book an Enterprise demo through the Calendly link
- Connect the catalog through Shopify, Adobe Commerce/Magento, Salesforce Commerce Cloud, CSV import or a custom API, with no code required according to the product page
- Let Paz.ai sync products, stock levels and prices, then enrich them with semantic descriptions, completed attributes and repaired structured data
- Choose how proposed changes are approved: automatic acceptance, batch approval, or product-by-product review
- Publish the resulting feeds to the OpenAI Product Feed, Google Merchant Center and Shopify
- Add the optional JavaScript attribution pixel to your own site to tie visits, carts and orders back to the AI engine that sent them
- Track visibility and attribution in the app.paz.ai console, where API keys are also created and revoked
- Install the Paz.ai - AI Readiness Checker Chrome extension to audit any product page on demand; results are emailed after a one-time code verification
Pros & Cons
Pros
- The whole chain in a single tool: measurement, correction, distribution and attribution
- Unusually complete machine-readable documentation (OpenAPI, MCP server, llms.txt, RFC 9727 api-catalog), so integration can be checked before buying
- A free diagnostic tool that needs no account, letting a buyer judge the fit before paying
- The live store is never modified, which limits operational risk
- Public prices and quotas, still rare in this segment
- A named and dated customer case (The Dessy Group, April 2026)
- Configurable approval workflow, from fully automatic to product-by-product review
Cons
- No permanently free tier: entry is USD 119 per month on annual billing, USD 149 month-to-month
- Product caps of 500 on Starter and 5,000 on Growth push larger catalogs quickly into quote-only Enterprise
- The 'SOC 2 compliant infrastructure' wording comes with no report, trust portal or date, so it cannot be read as a verified certification
- No customer data processing agreement is published or announced as available on request, and no EU representative is designated under Article 27 even though the GDPR is expressly invoked
- Hosting is announced in the United States only, with no European data residency option
- No documented way to exclude partner catalog data from model training; the only stated commitment covers Chrome extension users
- Structural dependence on third-party AI surfaces whose rules change, as the site itself documents with OpenAI's withdrawal of in-chat checkout in March 2026, from a young, English-only company (founding claimed in 2024) with no published funding round
Pricing & Plans
There is no permanently free plan. The lowest paid entry point is the Starter tier at USD 119 per month on annual billing, or USD 149 per month without commitment; annual billing is advertised as a 20% saving on all paid tiers. A 14-day free trial is offered on Starter and Growth, and the AI Readiness Report is a free diagnostic tool rather than a pricing tier. Growth is USD 279 per month billed annually or USD 349 billed monthly, and Enterprise is quoted on request. Plan changes are possible at any time: upgrades take effect immediately, downgrades at the next billing cycle.
- USD 119/month billed annually
- USD 149/month billed monthly - up to 500 products and 3 team members - AI visibility monitoring
- competitor tracking
- email alerts
- AEO Checker
- catalog optimization and quality scoring - Shopify and CSV connectors
- USD 279/month billed annually
- USD 349/month billed monthly - up to 5
- 000 products and unlimited team members - adds scheduled catalog syncs
- Brand Identity
- publishing to ChatGPT
- Google and Shopify
- and custom domains - adds the Adobe Commerce connector
- pricing on request - unlimited products and team members - adds the Agentic Storefront branded ChatGPT app
- SSO
- custom SLAs
- dedicated support and bespoke integrations - adds Salesforce Commerce Cloud
- Akeneo
- enterprise PIM and custom API connectors
Data, GDPR & hosting
A consolidated view of how Paz.ai handles your data.
GDPR overview
The GDPR is invoked explicitly, alongside the CCPA, in a privacy policy last updated in May 2026, with a separate California notice published at /ccpa. The policy lists the rights of access, rectification, erasure, objection, restriction, portability, complaint to a supervisory authority and withdrawal of consent. Rights are exercised by email to privacy@paz.ai, after identity verification, with a stated response time of 30 days. Personal data transferred out of the EEA is covered by an EU-approved mechanism such as the Standard Contractual Clauses. The minimum age is 16, or the local age of digital consent where it is higher. Notable gaps: no EU representative is designated under Article 27 despite the explicit reliance on the GDPR, no data protection officer is named, and no customer data processing agreement is published or announced as available on request.
Who owns the data?
Ownership is split three ways. The merchant keeps its catalog: Paz.ai never modifies the live store, website or fulfillment system, and works instead on an AI-optimized copy of the catalog data. AgentPass, Inc. retains ownership of the platform, its code, trademarks and documentation (terms of use, section 6), and grants the customer only a personal, limited, revocable, non-transferable, non-sublicensable and non-exclusive license to use it (section 5). Mutual confidentiality obligations are set out in section 14. Personal data remains that of the individual concerned, who keeps rights of access, rectification, erasure and portability.
Reuse rights
Paz.ai collects contact names, email addresses, country and account identifiers from partners, and online identifiers including IP addresses from site visitors. An optional attribution pixel, installed by the partner on its own site, records a session identifier, the referring AI source, the user agent, the commerce event type and the order value, currency and identifier; IP addresses captured this way are not stored, and no payment card data is collected. The Chrome extension processes the user's email, the audited page URL, a one-time verification code and the audit results, but not browsing history, passwords, cookies or keystrokes. The company states that it does not sell personal data and does not share it with advertising networks or data brokers. Named subprocessors include AWS and Cloudflare for hosting, OpenAI, Google and Anthropic for language models, plus transactional email, product analytics, browser automation, billing and security vendors. Google Analytics and Google Tag Manager set third-party cookies, with consent collected where the law requires it. The only stated commitment not to train models covers Chrome extension users; nothing equivalent is written for partner catalog data.
Data retention & training
Hosting summary
Paz.ai states that data is held on reputable U.S.-based cloud infrastructure. The named infrastructure subprocessors are Amazon Web Services and Cloudflare. Data is encrypted with AES-256 at rest and TLS 1.3 in transit. Personal data collected in the European Economic Area and transferred outside it is covered by an EU-approved mechanism such as the Standard Contractual Clauses, and no European data residency option is announced. One caveat on the technical reading: the domain resolves to 143.204.55.98, an Amazon CloudFront node (AS16509) geolocated in Zurich. That is the position of a CDN edge point of presence, not of the data storage itself, which the privacy policy places in the United States.
Where Paz.ai works
Country-level availability.
Not available in
Things to keep in mind
Risks and trade-offs to weigh before adopting Paz.ai.
- 'SOC 2 compliant infrastructure' is asserted on the product page with no report, trust portal or date behind it: treat it as a vendor claim, not as an audited certification
- No customer data processing agreement is published, so a European buyer will have to negotiate one before any procurement sign-off
- No Article 27 EU representative is designated even though the GDPR is expressly invoked, and hosting is announced in the United States only, with transfers resting on Standard Contractual Clauses
- There is no written guarantee that partner catalog data will be kept out of model training; the only such commitment covers Chrome extension users
- The market figures quoted on the site come from third parties (Adobe, Shopify, OpenAI, Salesforce, Digiday), not from measurements made by Paz.ai
- The 4.8/5 rating over 12 reviews is declared in the site's own structured data and cannot be checked on an independent review platform
- Liability is capped at the greater of USD 100 or the amounts paid over the preceding twelve months, and the paz.ai domain was registered in 2018, well before the company's claimed 2024 founding, so it is a second-hand domain; the former agentpass.ai address now redirects to it
Setup & Integrations
Technical difficulty
Low to moderate, depending on the source platform. A Shopify store connects in one click, with no code required according to the product page, and CSV import needs no development either. Adobe Commerce, Salesforce Commerce Cloud, enterprise PIM systems and custom APIs sit on higher tiers and do call for engineering. The optional JavaScript attribution pixel has to be placed on your own site, and the partner remains responsible for the notice and consent attached to it. Time to production is announced at two weeks, and under a week for a Shopify store.
Deployment
Integrations
Supported languages
Behind Paz.ai
Social
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
Which AI engines does Paz.ai monitor?
Does Paz.ai change my online store?
Which e-commerce platforms can I connect?
Is there a free plan?
How much does it cost?
How long does it take to go live?
Is there an API?
Where is the data hosted?
Is my data used to train AI models?
Who do I contact to exercise my rights, and what is the minimum age?
Should you pick Paz.ai?
Paz.ai occupies a young niche on a market that is still forming. Agentic commerce, where shoppers ask an assistant instead of browsing, is real enough that Adobe and Shopify publish figures on it, but the rules of the surfaces involved change from quarter to quarter: the site itself records OpenAI pulling in-chat checkout in March 2026.
Within that frame the product is coherent. It measures visibility on the engines that matter, finds the attribute gaps that keep a product out of a recommendation, fixes them on a copy rather than on the live store, publishes the result to OpenAI and Google, and closes the loop with order-level attribution. The technical documentation sits well above the segment average: an OpenAPI contract, an MCP server, an RFC 9727 catalog and llms.txt files make integration verifiable before anything is signed. Published prices and quotas are rarer still.
The reservations are equally clear. There is no permanently free plan, and the USD 119 entry point carries a 500-product ceiling. The SOC 2 compliant infrastructure wording is a claim, not a certification anyone can inspect. No customer data processing agreement is published, no Article 27 EU representative is designated despite the explicit reliance on the GDPR, and hosting is announced in the United States only, with transfers resting on Standard Contractual Clauses.
The practical test is traffic. A brand that already sees visits arriving from ChatGPT or Perplexity and cannot measure them has a concrete problem this tool addresses, and a Shopify catalog of a few thousand references connects quickly. Without that volume, or with strict European procurement requirements, the case is harder to make today.
- Choosing a selection results in a full page refresh.
- Opens in a new window.