Nosto AI review (Oct 2026): Huginn agents, AI search, and personalization for ecommerce stores

What this Nosto page covers

This page is for an ecommerce merchandiser, ecommerce manager, or the developer who runs a store’s personalisation, especially on Shopify, Adobe Commerce, BigCommerce, or Salesforce Commerce Cloud, who uses Nosto or is comparing it. The question it answers: what Huginn and Nosto’s AI features do, what limits Nosto documents, what it says about price, and who approves anything that changes what shoppers see.

Nosto’s agentic commerce page, the help centre articles Getting Started with Ask Huginn (10 Dec 2025) and Data Analyst modules and limits (4 Mar 2026), its personalized search page, and the pricing page were read on 16 October 2026. Ask Huginn is available to all merchants but asks for consent to use generative AI and acceptance of its terms on first use; Nosto’s FAQ also tells existing customers to ask their success manager about beta and early access.

Ground rules: the merchandiser approves every merchandising rule, synonym, bundle, segment, and recommendation change; Huginn’s numbers are checked against Nosto’s own reports before anyone acts on them; no customer names, emails, or order exports are pasted into prompts; and Nosto’s conversion and order-value figures are its claims, not our measurement.

Major features (with how someone uses them)

Examples at a glance: each feature has a concrete example from a real job plus a short Try this you can run this week.

1. Ask Huginn’s Data Analyst (primary)

What it is: The Data Analyst agent answers questions about your store and Nosto data in plain words: product views, units sold, conversion rates, stock status, orders influenced by recommendations, and results for search, category merchandising, bundles, and personalised emails. Nosto’s help centre lists its limits: product questions use fixed periods of 1, 3, 7, 14, or 30 days (it picks the closest if you ask for another); stores with more than 5,000 products get answers about the top 5,000 most-viewed matching products only; and order questions cover the most recent 400,000 matching orders. Pop-ups, A/B test analytics, user-generated content, and post-purchase upsell are not yet supported.

An answer drawn from the top 5,000 products or the last 400,000 orders is a sample, not the whole store. Note which metric and period it used, and check a few products in the Nosto dashboard before changing a page.

Concrete example (merchandiser at a home decor store) The merchandiser asks which five products had the most views but the fewest add-to-carts in the last 30 days. Huginn returns five lamps and rugs and shows conversion rate as its measure. She opens each product page: two have only one photo, one shows a delivery estimate of six weeks, and two look fine. She adds photos and asks the buyer about the delivery time before changing anything else.

Try this

Primary workflow:

Ask Huginn's Data Analyst which five products had the most views but the fewest add-to-carts in the last 30 days. Have the merchandiser check each product page before changing anything.

2. Huginn and its agents: what each one does

What it is: Nosto calls Huginn its built-in central agent. The help centre (Dec 2025) lists three agents in Ask Huginn: a Docs Agent for how-to questions about Nosto, the Data Analyst Agent, and a Developer Agent for templates and implementation, whose coding abilities Nosto says are “currently limited”. The agentic commerce page lists five, adding a Merchandising agent that creates and tunes merchandising rules and a CRO agent that suggests A/B tests. Nosto says nothing goes live without your sign-off.

Two Nosto pages give different agent lists, and the FAQ points to beta and early access. Ask your success manager which agents your account has, and in which status, before planning work around them.

Concrete example (ecommerce manager at a fashion store) The ecommerce manager asks her success manager which Huginn agents are on her account and in what status, and writes the answer next to each agent. For the one marked early access, she starts on a single category and asks the merchandiser to approve every rule it proposes for the first month.

Try this

Know what you have:

Ask your Nosto success manager which Huginn agents are on your account and whether each is generally available, beta, or early access. Write the answer down before planning work around them.

3. AI Reveals: suggestions waiting for review

What it is: Nosto’s agentic commerce page describes Huginn analysing your store data in the background and surfacing opportunities “ready for you to review and activate”: AI Segments Opportunities (audience segments you may have missed), generated titles for recommendation blocks, clean-up suggestions for facets (merging duplicates, fixing misspellings, harmonising values), AI Synonyms Suggestions for search, and bundle suggestions with fallback products so bundles stay live. None of these goes live until someone activates it.

A queue of suggestions is easy to accept in bulk. Treat each one as a change to your store: a merged facet or a new synonym changes what every shopper sees.

Concrete example (merchandiser at a beauty store) The merchandiser opens the Reveals queue. She accepts a facet merge of “Moisturiser” and “Moisturizer”, rejects a synonym pairing “serum” with “oil” because they are different products, and changes a generated recommendation title that says “Best sellers” on a block that actually shows new arrivals.

Try this

Review the queue:

Open Huginn’s suggestions for facets, synonyms, and recommendation titles. Have the merchandiser accept, reject, or edit each one, and write down why any were rejected.

4. AI search and the conversational search bar

What it is: Nosto’s personalized search combines keyword and vector search (for long-tail queries with few matching words), natural-language processing, per-shopper personalisation, merchandising rules you can combine and weight, previews of results for any query or segment, AI-suggested synonyms, and support for 27 languages out of the box. On the agentic page, Huginn adds a conversational search bar (“what’s a good gift under $50?”) with follow-up suggestions and image search from a shopper’s photo.

A conversational answer can sound sure about a product that only loosely fits. Preview the queries shoppers really use, including the vague ones, before a launch.

Concrete example (ecommerce manager at a gift shop) The ecommerce manager previews ten gift-style questions in the search bar before the holiday season. “Gift for a dad who likes coffee under $40” returns mugs and a grinder that costs $65. She checks the price filter and the product data and asks the merchandiser to set a rule so gift questions respect the price a shopper types.

Try this

Preview before a launch:

Preview ten real shopper questions in Nosto’s search, including two with a price limit. Have the merchandiser mark every result that ignores the price or is out of stock before the next campaign.

5. Merchandising rules beside the AI

What it is: Across search, category pages, and recommendations, Nosto keeps rule-based controls next to its AI ranking: merchandising by seasonality, stock, conversion rate, margin, and other attributes, plus redirects and the ability to preview by segment. The Merchandising agent described on the agentic page proposes and tunes these rules; Nosto says changes need your sign-off. Its personalised search also lets you set which product attributes count and how much.

When an agent writes a merchandising rule, it is a rule like any other: it stays until someone removes it. Keep a list of agent-proposed rules with an owner and a review date.

Concrete example (merchandiser at a sporting goods store) The merchandiser accepts a proposed rule that boosts high-margin running shoes in the “running” category. Two weeks later she reviews it: conversion is unchanged, but returns on two boosted models are high because they run small. She narrows the rule to the well-fitting models and sets a review date.

Try this

Keep a rule list:

List every merchandising rule active in Nosto, including any Huginn proposed. Have the merchandiser give each an owner and a review date and remove any from a finished campaign.

6. Huginn Connect and the Nosto MCP server

What it is: Nosto says Huginn can work with outside agents through MCP or API. The agentic page lists connections for Klaviyo (marketing workflows), Omniconvert (experimentation), Shopify Sidekick, and a Nosto MCP server for developers using tools such as Claude Code, Cursor, or VS Code, plus Shopping Assistants Enrichments, which adds Nosto’s recommendations to a third-party on-site assistant (contact Nosto to check compatibility).

Connecting Huginn to an email tool or a coding assistant extends who can act on your store data. Decide which actions each connection may take and who approves them before you switch it on.

Concrete example (developer at a Shopify store) The developer connects the Nosto MCP server to her coding assistant on a staging store and asks it to update a recommendation template. The change works but removes an accessibility label. She restores the label, keeps the connection on staging, and asks the merchandiser to approve template changes before they reach the live store.

Try this

Try it on staging:

Connect the Nosto MCP server on a staging store only. Have the developer review every template change it makes and the merchandiser approve it before anything reaches the live store.

Pricing context (October 2026)

Nosto’s pricing page (checked 16 Oct 2026) publishes no prices. Pricing is modular and custom: a base platform fee plus a fixed fee calculated on your store’s volume — GMV turnover and traffic — and the level of support and scale you need. You choose modules from two suites (Product Experience Cloud and Content Experience Cloud) and can start with one. The page says AI capabilities are built into every module at no extra cost.

There is no standard self-service free trial; Nosto offers a structured proof of concept for qualified merchants, installed through the app store, to preview AI search and merchandising on your own data. All customers get the help centre and Ask Huginn; Professional and Enterprise customers get tiered priority support and customer success. Treat it as quote-only.

Uncertainty: No published prices on 16 Oct 2026, so cost depends on your quote. Which Huginn agents are generally available is unclear: the help centre lists three, the agentic page five, and the FAQ refers to beta and early access, so ask before planning. The Data Analyst’s limits (fixed periods, top 5,000 products, latest 400,000 orders) may change. Conversion and order-value lifts on Nosto’s pages, including “up to 2×” for assistant enrichments, are its claims, not our measurement.

When it’s useful (honest take)

Useful when you want recommendations, search, category merchandising, and content personalisation from one vendor with AI included in each module, and a merchandiser who will review Huginn’s suggestions and rules every week.

Less useful when you only need search and want published usage prices (Algolia AI), you want email and SMS on the same platform as search (Bloomreach AI), or the main gap is messaging rather than on-site discovery (Attentive vs Postscript).

Related options (educational)

FAQ

How much does Nosto cost?

Nosto publishes no prices. Its pricing page says pricing is a base platform fee plus a fixed fee based on your GMV turnover and traffic, that AI is included in every module at no extra cost, and that there is no self-service free trial (checked 16 Oct 2026).

What is Huginn?

Nosto’s built-in central agent. Through Ask Huginn you can question Docs, Data Analyst, and Developer agents; Nosto’s agentic page also lists Merchandising and CRO agents. Nosto says nothing goes live without your sign-off.

What are the Data Analyst agent’s limits?

Per Nosto’s help centre: fixed periods of 1, 3, 7, 14, or 30 days for product questions, the top 5,000 most-viewed products in larger catalogues, and the most recent 400,000 orders. Some modules, such as pop-ups and A/B test analytics, are not yet supported.

Do I need to agree to anything to use Ask Huginn?

Yes. Nosto says Ask Huginn is available to all merchants but requires consent to use generative AI within Nosto and acceptance of the Ask Huginn terms on first use.

Who should approve Nosto AI changes?

A named merchandiser for rules, synonyms, bundles, segments, and recommendation changes; a developer for template changes from the Developer Agent or MCP; the ecommerce manager for anything tied to discounts.

Optional next step (if seats overlap)

If you pay for a recommendations tool, a separate site search, and a store platform with its own AI that all decide which products shoppers see, that overlap is worth mapping. Map the overlap with the $19 AI toolkit audit or the $9 AI toolkit spreadsheet. Email Audit $19 or Toolkit spreadsheet $9.

Checkout is live via Stripe. After you pay, we email delivery (spreadsheet) or intake next steps (audit).

Related: Algolia AI · Attentive vs Postscript · Bloomreach AI · Attentive AI · Postscript AI · Wix AI · Yotpo AI · BigCommerce AI · Triple Whale AI · Klaviyo AI · Contentsquare AI · Optimizely AI · Gorgias AI · Gemini for marketing ops · Claude · Three checks before another AI seat · How-to: Review Nosto AI Reveals suggestions, and have the merchandiser approve every change · ops hub