Shopbox AI
Shopbox AI is an ecommerce personalisation platform for online retailers. A single AI engine powers onsite search, product recommendations, conversational shopping and merchandising, then carries the same session context into email, SMS and WhatsApp campaigns.
What is Shopbox AI?
Shopbox AI sells itself as an AI commerce layer rather than another point solution bolted onto a storefront. One engine sits on top of the retailer's existing platform and drives five onsite surfaces at once, then pushes the same understanding out into the channels that follow the shopper off the site.
Search is the first surface. Instead of matching keywords, the engine reads the whole session context — what has been browsed, what is in the basket, where the visitor arrived from — and ranks on intent. Shopbox claims this removes manual rule management entirely: no blacklists, no synonym dictionaries, no boosting spreadsheets. Comparison, filtering and guided refinement tools are built into the results.
Recommendations are the second. The pitch here is speed of understanding rather than depth of history: personalisation begins on the first click, with no prior visit data, no segments and no cohorts, and every surface updates as the session's intent shifts. That matters most for the anonymous traffic the company puts at 95% of a typical store's visits.
The third is conversational commerce, positioned deliberately against chatbots. Shopbox describes a guided buying journey that intervenes at the point of friction knowing the basket, the searches and the browsing behind it, and contrasts 13 product interactions over six weeks for a standard chatbot with 6,000 for its own approach.
Fourth is dynamic merchandising, where the team sets a business objective — protect margin, clear seasonal stock, launch a range — and the engine applies it consistently everywhere instead of following hand-written rules. Fifth, through Klaviyo, the same context reaches email, SMS and WhatsApp for browse abandonment, replenishment prompts and post-purchase cross-sell.
The company claims 25M+ products running in real time, a 4-12% revenue uplift measured with SKU-level attribution it calls deliberately punitive, and a 30-minute go-live on Shopify. The publisher, Shopbox AI, is a Dublin company founded in 2020 by Alan Gormley and Tariq Zaki, with fourteen named retail case studies published since 2021.
What it does
- Interpret what a shopper means rather than what they typed, and rank on session intent
- Recommend products from the first click, with no stored profile, segment or cohort
- Guide a hesitating buyer through conversation at the moment of friction
- Apply one merchandising objective, such as margin or stock clearance, across every surface at once
- Carry onsite context into Klaviyo email, SMS and WhatsApp messages
- Attribute revenue to SKU level under a deliberately restrictive rule
- Run built-in A/B tests and read the results in an analytics dashboard
When to use Shopbox AI / When not to
A quick filter to help you decide if Shopbox AI is the right fit.
When to use Shopbox AI
- Online retailers already trading on Shopify or Shopify Plus, where Shopbox advertises a 30-minute go-live
- Merchandising and ecommerce teams with no dedicated developer resource to spare
- Stores with very large catalogues, since the platform claims 25M+ products running in real time
- Sites carrying heavy anonymous traffic, because personalisation starts on the first click without a stored profile
- B2C and B2B retailers in fashion, food, homeware, sportswear, cosmetics and hobby verticals, where the published case studies sit
When not to use Shopbox AI
- Shoppers themselves: this is a business-to-business product sold to retailers, not a consumer app
- Teams that need a free tier or a trial before committing, as the cheapest published entry point is a paid module
- Engineering teams wanting to assess the integration from public API documentation, which the site does not publish
- Buyers with a data-residency requirement, since no hosting country or region is named anywhere on the site
- Organisations whose procurement demands formal security certification, as no ISO 27001 or SOC 2 claim appears
How to use Shopbox AI
A typical end-to-end flow, from setup to results.
- Open the public interactive demo page to see a curated homepage working before talking to anyone
- Book the 30-minute demo, which is run live on a real retailer site in your own vertical
- Answer a short brief on your platform, catalogue size and current conversion problems
- Choose the modules you actually need, since personalisation, search, conversational and marketing are priced separately
- Install the Shopbox app from the Shopify App Store, or connect through the API for other platforms
- Have the styling matched to your brand, which the FAQ identifies as where most of the effort goes
- Skip training periods, segment setup and profile building: the engine personalises from the first click
- Set a merchandising objective instead of writing rules, and let it apply across all surfaces
- Connect Klaviyo to extend the same session context into email, SMS and WhatsApp
- Follow uplift in the Shopbox Hub dashboard, with revenue attributed at SKU level
Pros & Cons
Pros
- Published per-module pricing, unusual in a market where contact-sales is the norm
- Works on anonymous traffic from the first click, with no dependence on visit history
- One engine across five surfaces means a single integration and consistent behaviour
- Fast deployment claimed: 30 minutes on Shopify, with no dedicated developer required
- Attribution rules are conservative and stated openly, tracked down to SKU level
- Light technical footprint, roughly 50KB loaded after the rest of the page
- Fourteen named, dated case studies with attributed customer quotes, from an EU-based publisher on Irish-law contracts
Cons
- An API is sold as a connection route but no public documentation exists to evaluate it
- Every performance figure is self-reported, and the results page carries its own methodology asterisk
- The FAQ still describes Automate, Growth and Enterprise packages that the pricing page does not offer
- No free plan and no free trial are announced, so the only entry point is a paid module
- No hosting country or region is named, although the privacy policy anticipates transfers outside the EEA
- No subprocessor list is published and no security certification such as ISO 27001 or SOC 2 is claimed
- A single email address appears on the entire site, with support and legal both routed through a form
Pricing & Plans
No permanent free plan and no free trial are announced. Pricing is modular and published: the lowest entry point is the AI Marketing Connector at USD 50 per month, followed by AI Search at USD 240 per month, AI Personalisation at USD 250 per month and Conversational AI at USD 400 per month. A retailer after the core personalisation capability should therefore budget from USD 250 per month rather than USD 50. All amounts are stated exclusive of VAT, and custom enterprise pricing is offered for high-volume requirements. Every call to action on the pricing page leads to a demo booking rather than to an online checkout.
- real-time behavioural signals
- homepage
- PLP and PDP personalisation
- built-in A/B testing
- analytics dashboard
- natural language understanding
- semantic and vector search
- typo tolerance and synonyms
- merchandising rules
- context-aware product guidance
- cart and checkout support
- custom persona and tone
- full conversation analytics
- SMS and WhatsApp
- AI-generated copy and timing
- smart audience segmentation
- revenue attribution
- Custom enterprise pricing
- quoted on request for high-volume requirements
- with no public amount
Data, GDPR & hosting
A consolidated view of how Shopbox AI handles your data.
GDPR overview
Compliance is claimed in plain terms. A dedicated FAQ entry asks whether Shopbox complies with GDPR, PIPEDA, LGPD or equivalent laws, and answers yes. The privacy policy names a Data Protection Officer, Tariq Zaki, reachable by email, sets out the eight GDPR rights in full, and commits to answering subject access requests within 21 days, one month at the outside and three for complex cases. Transfers beyond the EEA are anticipated and covered by a GDPR-compatible processing addendum with subprocessors, plus encryption in transit and, where possible, at rest. Being established in Ireland, the company needs no Article 27 representative. One inconsistency stands out: the policy directs complaints to the Information Commissioner's Office, the United Kingdom regulator, rather than to Ireland's own Data Protection Commission.
Who owns the data?
The terms keep two estates apart. Customer Data is defined as any data the retailer provides to Shopbox, and personal data inside it stays Customer Personal Data: the retailer remains in charge, while Shopbox acts on instruction and must pass on subject access requests it receives. Ownership of the platform runs the other way. Shopbox retains all intellectual property in the licensed services, their documentation and any improvement, and the agreement conveys no proprietary rights in the platform to the customer. One clause deserves attention: 12.6 reserves Shopbox the ability to use anonymised consumer data for analytics or to improve its own platform.
Reuse rights
The retailer keeps its own Customer Data and needs no permission from Shopbox to use it. The licence granted runs in the opposite direction, covering access to the platform rather than rights over the data, and it explicitly stops the customer acquiring ownership of the software, its documentation or any improvement made to it. Going the other way, clause 12.6 reserves Shopbox the use of anonymised consumer data for analytics and product improvement, and no mechanism is documented that would let a customer decline that use. On the visitors themselves, the company states that it excludes automated decision-making and profiling, sells nothing to data brokers and runs no re-marketing, and that the product works without collecting visitor data at all.
Data retention & training
Hosting summary
The site names no hosting country, no region and no infrastructure provider, which is an unusual gap for a platform sold to retailers. What the privacy policy does say is that personal data may be stored in, or transferred to, countries outside the European Economic Area. Those transfers are said to be covered by a GDPR-compatible data processing addendum with subprocessors in third countries, along with checks that those subprocessors keep adequate security procedures in place. Data is encrypted in transit and, in the policy's own wording, encrypted at rest where possible. No subprocessor list is published. The domain resolves to a Cloudflare anycast address, which reveals nothing about where customer data actually sits. Any buyer with a data-residency obligation will have to ask the company directly and get the answer in writing.
Things to keep in mind
Risks and trade-offs to weigh before adopting Shopbox AI.
- Every uplift number on the site is the vendor's own, so treat 4-12% revenue growth as a claim to test rather than a figure to budget against
- The pricing page and the FAQ describe two different commercial offers, so confirm which one is current before you negotiate
- Clause 12.6 lets Shopbox use anonymised consumer data to improve its platform, and no way to opt out of that is documented
- Prices are shown in US dollars excluding VAT by a euro-zone publisher, so ask which currency you will actually be invoiced in
- The privacy policy points complainants to the UK regulator rather than the Irish one, which suggests it was adapted from a British template
- Personalisation that decides what shoppers see can quietly narrow discovery, so watch catalogue breadth and long-tail sales, not conversion alone
- Handing merchandising judgement to a goal-setting engine erodes a team's own feel for the catalogue if nobody reviews what it keeps promoting
Setup & Integrations
Technical difficulty
Low, on the vendor's own account. Shopbox advertises a 30-minute go-live on Shopify through its App Store app, and up to two weeks for the most complex custom builds connected through the API. No dedicated developer resource is said to be required, and there is no training period, segment configuration or profile building to set up. The FAQ places the real effort elsewhere: getting the look and feel right against the retailer's own brand. The script adds roughly 50KB and loads after the rest of the page.
Deployment
Integrations
Behind Shopbox AI
Fundraising
Social
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
What does Shopbox AI actually do?
Does a shopper need to be logged in for it to work?
Which ecommerce platforms does it support?
How long does it take to go live?
How much does it cost, and is there a free plan?
Is there an API?
Where is customer data hosted?
Will it slow the store down?
Is it GDPR compliant?
Who is behind Shopbox AI?
Should you pick Shopbox AI?
Shopbox AI is a coherent proposition in a market that mostly sells fragments. The argument for one engine behind search, recommendations, conversation, merchandising and outbound messaging is easy to follow, and the promise of personalisation on the first click speaks directly to the anonymous traffic most stores cannot address. Two things set it apart from its peers: published per-module pricing, which is rare here, and an attribution rule the company describes as restrictive and applies down to SKU level.
The evidence is respectable without being independent. Fourteen dated case studies name real retailers and quote named people, industry awards are listed on the About page, and the results page carries its own methodology footnote rather than hiding it. But every uplift figure remains the vendor's own, and none of it can be checked from outside.
Three gaps deserve a buyer's attention. An API is offered as a connection route yet no public documentation exists, so a technical evaluation has to go through sales. Nothing on the site names a hosting country, publishes a subprocessor list, or claims a security certification, which will stall any procurement process with a data-residency or compliance requirement. And the FAQ still advertises Automate, Growth and Enterprise packages that the pricing page has replaced — a small thing, but it means the site contradicts itself about what is actually for sale.
For a mid-sized online retailer on Shopify or a mainstream platform, without a data team and with a catalogue large enough that manual merchandising has stopped scaling, Shopbox is worth the 30-minute demo. Go into it with the questions above written down, and ask for the API documentation before signing anything.
- Choosing a selection results in a full page refresh.
- Opens in a new window.