AI for BusinessNovember 12, 2026

AI Personalisation for Ecommerce in 2026: Recommendations, Email and On-Site Tools

Glass office towers photographed from street level

Written by Thomas Lovas Lokoy, Founder & Editor | NorwegianSpark SA

Researched and drafted with AI assistance from published sources; edited and published by NorwegianSpark SA · Editorial policy · Disclosure

Last updated: November 2026

The short answer: start with the personalisation you already have. On Shopify, product recommendations are generated automatically and Shopify’s own Search & Discovery app, which is free, lets you set them by hand. In email, a platform such as Klaviyo adds predictions such as expected next order date — but only once your store has at least 500 customers who have ordered and 180 days of order history. Pay for an on-site tool such as Rebuy (priced by monthly orders), LimeSpot (from $150 a month) or Nosto (custom pricing) only when you can test it against a control group and see it earn more than it costs. Prices and conditions below are from each vendor’s own pages, read on 12 November 2026, in US dollars.

What AI Personalisation Actually Does in a Store

“Personalisation” covers four jobs, and most tools do one or two of them well:

  • Product recommendations. “Related” and “bought together” blocks on product pages, in the cart and after checkout. This is where most of the money in personalisation is made, because it decides what a shopper sees next.
  • Personalised search and collections. Ordering results and category pages for the person browsing rather than the same for everyone.
  • Personalised content. Different banners, messages or offers for different segments, such as first-time visitors and returning customers.
  • Personalised email and messages. Sending based on what a customer did and is predicted to do next, such as a reminder timed to when they usually reorder.

The “AI” in all four is mostly prediction from behaviour: what people who viewed this also bought, what this customer is likely to want, when they will next order. That makes the size of your data the first limit on what any tool can do for you — a model needs orders to learn from.

The Tools at a Glance

ToolWhat it personalisesPrice on the vendor pageTrial
Shopify Search & DiscoveryRelated and complementary product recommendations, filters, search boosts and synonyms on Shopify storesFreeNot needed
KlaviyoEmail and SMS segments and flows, with predictive analytics once data thresholds are metFree up to 250 active profiles and 500 emails a month; Email plan from $20 a month (500 active profiles, 5,000 emails)Free plan
RebuyCart, checkout and post-purchase recommendations, search, A/B-tested flows (Shopify)Per package, by monthly orders: up to 500 orders, Checkout & Post-Purchase $25, Cart & Merchandising $45, Search & Collections $35, Flows & A/B Testing $35 a month14 days per package
LimeSpotRecommendations and content personalisationMax plan $150 a month for up to $50,000 of Online Store revenue per 30 days15 days
NostoRecommendations, personalised and AI search, content personalisationCustom: a base platform fee plus a fee based on GMV and traffic; no prices publishedProof of concept for qualified merchants
Dynamic YieldEnterprise personalisation and testingNo price found; dynamicyield.com now redirects to Mastercard’s websiteNot stated

We hold an affiliate partnership with Shopify, not with Klaviyo, Rebuy, LimeSpot, Nosto or Dynamic Yield; they are here because they are established names in this job.

Start Free: Shopify’s Own Recommendations

If you sell on Shopify, you already have recommendations. Shopify’s help centre says related product recommendations are generated automatically for every product, from three sources in order: products commonly bought together, products with a similar description, and products in related collections. Two limits matter for small stores. Purchase-based recommendations need previous sales of that product, and description-based ones are available only to merchants with an English storefront; when neither is available, Shopify falls back to collections.

Shopify separates two kinds: related products, which are similar to the one being viewed, and complementary products, which are often bought in addition to it. Showing complementary products needs a theme that includes a complementary-products section. You cannot edit the generated recommendations, but in the free Search & Discovery app you can add your own: go to Apps > Search & Discovery > Product recommendations > Add recommendations, choose a product, and add up to 10 complementary and up to 10 related products. You can then decide whether related products show only your picks or your picks plus Shopify’s automatic ones.

For a store with a few hundred orders a month, hand-picking complementary products for the twenty best sellers is often the highest-value personalisation work available, and it costs nothing. The same app also handles filters, search boosts and synonym groups, which is personalisation of a simpler kind: making search answer the words your shoppers actually use.

Email: Where Prediction Starts to Pay

Email is usually the cheapest place to personalise, because you already know who the customer is. Klaviyo’s pricing page lists a free plan for up to 250 active profiles and 500 emails a month, and an Email plan from $20 a month for 500 active profiles and 5,000 emails, with an “AI marketing agent” called Composer included at limited usage.

Its predictive analytics are the part most often oversold. Klaviyo’s help centre says the predictions — predicted customer lifetime value, churn risk, average time between orders and expected date of next order — appear only when your account has:

  • at least 500 customers who have placed a non-zero, non-cancelled, non-refunded order (customers who bought, not total profiles);
  • at least 180 days of order history, with orders in the last 30 days;
  • at least some customers who have placed three or more orders; and
  • an ecommerce integration or API sending placed orders.

Below those thresholds, “AI personalisation” in email means segments and flows you set up yourself — which still work well. A browse-abandonment flow and a post-purchase flow that recommends complementary products are the two most stores set up first.

Paid On-Site Tools: Rebuy, LimeSpot and Nosto

Rebuy is built for Shopify and sells four packages you can buy separately: Cart & Merchandising, Checkout & Post-Purchase, Search & Collections, and Flows & A/B Testing. Each is priced by your average monthly Shopify orders — at up to 500 orders a month the four cost $45, $25, $35 and $35 a month, rising to $425, $225, $325 and $325 at 5,000 orders — with 10% off two packages, 20% off three, 30% off all four and a further 10% on annual billing. Each package has a 14-day free trial. Its Smart Cart is described as a slide-out cart with AI-powered product recommendations.

LimeSpot lists one plan on its pricing page, Max, at $150 a month for stores with up to $50,000 of revenue per 30 days from the Online Store channel, with all features included and a 15-day free trial. Its solutions include smart recommendations and content personalisation.

Nosto does not publish prices. Its pricing page describes a modular model — a base platform fee plus a fee calculated on your store’s GMV and traffic — with AI included in every module, and a structured proof of concept on your own data for qualified merchants instead of a self-service trial. That points it at larger stores than Rebuy and LimeSpot. Dynamic Yield, long the enterprise name in this category, now sits on Mastercard’s website; we could not find a published price.

Does It Pay? A Break-Even Check Before You Buy

A personalisation tool has to earn its fee in extra profit, not extra revenue. Two numbers decide it: the monthly fee, and your gross margin. As a purely illustrative example, a $150-a-month tool in a store with a 40% gross margin needs $375 a month of extra sales just to cover itself ($150 ÷ 0.40). If your store takes $20,000 a month, that is a lift of just under 2% that must be caused by the tool — not merely sold through its widget.

That last distinction is where most vendor dashboards mislead, usually without meaning to. A shopper who clicks a recommended product and buys it is counted as “revenue from recommendations”, even if they would have found and bought the same product anyway. The only honest measure is a comparison against a control group:

  1. Write down the baseline. Average order value, conversion rate and revenue per visitor for the four weeks before you install anything.
  2. Split traffic. Show the personalised experience to part of your visitors and the plain one to the rest. Some tools include this (Rebuy sells it as its Flows & A/B Testing package); otherwise run the tool for a period and compare like-for-like weeks, accepting that this is weaker evidence.
  3. Compare revenue per visitor, not clicks on the widget. If revenue per visitor in the personalised group does not beat the control by more than the tool costs, it is not paying for itself, whatever its dashboard says.
  4. Decide at the end of the trial. Set the decision date before you start, so a free trial does not quietly turn into a subscription.

What to Fix Before Any Tool

Personalisation reorders your catalogue; it cannot improve it. Recommendations built on thin product data recommend badly, and a model cannot learn from products with no sales. The work that makes every tool here perform better is the same work that helps search and AI assistants find you: complete product attributes, consistent categories and descriptions that answer real questions. Our guide to selling through ChatGPT, Gemini and Copilot lists the product data those channels read, and it overlaps almost entirely with what recommendation engines need. If you are still choosing what to stock, a supplier catalogue such as Doba is covered in our product research guide.

One more check: if you sell to customers in the EU or UK, make sure the personalisation you add is covered by your cookie and consent setup before it goes live, because most on-site personalisation depends on tracking what a visitor does.

Disclosure: this article contains affiliate links to Shopify and Doba, plus the sponsored placements on this page. If you sign up through them we may earn a commission at no extra cost to you, and that can influence which providers we feature. We have no affiliate relationship with Klaviyo, Rebuy, LimeSpot, Nosto or Dynamic Yield.

Frequently Asked Questions

What is the best AI personalisation tool for ecommerce?

It depends on your size. A Shopify store should start with Shopify’s automatic recommendations and the free Search & Discovery app, plus email flows. Paid on-site tools such as Rebuy (priced by monthly orders) and LimeSpot ($150 a month on its Max plan) suit growing Shopify stores; Nosto and Dynamic Yield are aimed at larger retailers and do not publish prices.

Is AI personalisation worth it for a small store?

Often not as a paid tool at first. Klaviyo’s predictive analytics need at least 500 customers who have ordered and 180 days of order history, and recommendation models need sales to learn from. Hand-picked complementary products and basic email flows usually come first; buy a tool when you can test it against a control group.

Does Shopify have built-in product recommendations?

Yes. Shopify generates related product recommendations automatically from purchase history, similar descriptions (English storefronts only) and related collections. Its free Search & Discovery app lets you add up to 10 complementary and 10 related products by hand for each product.

How do I know if personalisation is increasing sales?

Compare revenue per visitor between visitors who see the personalised experience and a control group who do not. Revenue “attributed” to a recommendation widget overstates the effect, because it includes purchases that would have happened anyway.

How much do AI personalisation tools cost?

On the vendors’ own pages, read on 12 November 2026: Shopify Search & Discovery is free; Klaviyo has a free plan and an Email plan from $20 a month; Rebuy packages start at $25 a month for stores with up to 500 orders a month; LimeSpot’s Max plan is $150 a month; Nosto prices each store individually.

Related Articles

Sources

Continue reading

Continue in this collection