What’s new in 40rty.

AMS: Agent Directives, your sales logic in plain language

New in AMS, our conversational storefront. Agent Directives are standing rules the AMS agent follows in every shopper conversation: when this happens, do that. Bundles, upsells, gift flows and free-shipping nudges, written as sentences instead of hard-coded flows or quizzes.

All of a store’s directives sit in one ordered list. Change one and the next shopper session runs on it, no deploy. Directives are set up together with the 40rty team.

Autopilot: an AI operator for your agentic channels

AgentIQ gives you the tools to win with AI shopping agents. Autopilot runs them for you. It decides what to work on, does the work, measures whether it paid off and adjusts, every day, from your Slack channel.

Each week Autopilot reads your numbers and sets one to three goals. Each day it checks its own past work first, then picks the highest-impact products from your store’s signals: health, revenue, AI agent sessions, prompt demand, inventory and the retail calendar. It fixes categories and category metafields, builds Prompt Libraries and writes Agentic Product Listings, always foundation first.

Every listing it publishes is judged at day 7 (rank), day 30 (traffic) and day 90 (conversion). The verdicts are calculated from your data, not written by the AI. A listing that misses is rebuilt on new prompts, and the prompts that failed aren’t tried again. Every Friday you get a report of what changed and what moved.

You choose how autonomous it is: fully autonomous, confidence-based or review everything. Change it any time, or just tell it what to focus on in plain language.

Tracked Prompts: see how often AI agents find you

Every prompt you use in an Agentic Product Listing is now tracked automatically. AgentIQ searches the Shopify Catalog for each one on a regular schedule and shows how often your products come back, and whether that’s getting better.

Your Tracked Prompts Rate sums it up in one number on the AgentIQ Overview, weighted toward the prompts most of your products target. Next to it, the change over the last 7 days: a green arrow means AI agents are finding you more often, an amber one less often.

Below it, every prompt gets a row: its rank rate, how many of your products use it, how well your catalog matches it and which competing stores showed up. Expand a row to see exactly which of your products came back and which competitors’ products did. Re-run a search any time, or open it in Shopify Catalog.

Bulk AI Suggestions: fix 20 products at once

A shallow product category and empty category metafields are the two gaps that most often stop an AI shopping agent recommending a product. Until now you closed them one product at a time. Now you select up to 20 products in My Store, press once and carry on working.

The run happens in the background. Progress lives in the alerts drawer, so you can leave the page, and runs stack if you want to queue another. When it finishes, every waiting product sits in one review table: the category it has today, the category AgentIQ suggests, and one line per category metafield with the values Shopify actually accepts for that field. No free text, so nothing you approve can bounce later.

You decide field by field. Accept a value, leave it pending for later or clear it. A row is Approved, Pending or Not ready, never two at once, so one press can never write half a product and store the other half.

AgentIQ now works inside Sidekick

AgentIQ is now available directly inside Sidekick, Shopify’s built-in AI assistant. Merchants can ask Sidekick to search the Knowledge Center, run playbooks, check product status and APLs, and now create or edit store FAQs.

FAQ creation is the first AgentIQ action Sidekick can actually perform, not just look up. Sidekick drafts the FAQ and shows the merchant a confirmation form before anything is saved, so nothing goes live without their approval.

Prompt Library, now ranked by Volume and Opportunity Score

Grounding your test prompts in real shopper language beats internal guesses every time. Every prompt in your Prompt Library now carries a Volume estimate and an Opportunity Score, so you know exactly which ones are worth optimizing for first.

Introducing a new product AMS

AMS (Agentic Management System) is a platform that generates a personalized, real-time storefront tailored to each individual shopper’s (or AI agent’s) intent, rather than serving a static, one-size-fits-all page.

Playbooks, ten jobs, step by step

Features tell you what a button does. Playbooks tell you which buttons to press, in which order, to get a result.

Each one names a goal, a time to complete and the features used, then walks you through the exact clicks. A few of the plays now live:

Brand Context, AgentIQ now knows who you are

AgentIQ keeps what it knows about your brand in one editable place. Brand Context compiles your Brand Identity and your Buying Personas into a single source of truth used across the platform, to create better prompts, run Operator and power other features.

It is generated automatically during onboarding, and every field is yours to edit. The more accurate it is, the closer AgentIQ gets to how your real shoppers actually talk and buy.

Agentic mapping, the step that makes agents see your APLs

AI agents only read your optimized listings after the AgentIQ fields are mapped into your Shopify Catalog. Without it, agents keep reading your raw catalog data.

It takes one visit to the Shopify Catalog Mapping screen in your Shopify admin, once per store, after your first APL exists:

Save once and every APL you create afterwards flows through automatically. Products without an APL fall back to your default Shopify fields, so creating APLs is how you expand coverage. Work your catalog by Health score: Poor first, then Mediocre, then Excellent. Allow 24–48 hours for changes to surface in the Shopify Catalog.

Product listing suggestions, fix the page humans see, too

Creating the APL is half the work. The other half is fixing your human-facing page based on what AgentIQ learned from your competitors.

Open any product in the listing editor, via Fix Product in My Store, and the suggestions are inline: titles, descriptions, image alt text and metafields, each with an impact level (High or Medium), a one-line reason and a ready-to-apply replacement. Apply, Edit or Dismiss each one, then save to commit to Shopify.

Click Optimize to run the analysis, AgentIQ scores your listing against your tracked prompts and surfaces the suggestions, so you can see at a glance which shopper queries each change helps you win. Pay particular attention to category depth, “Ski & Snowboard Goggles” performs significantly better than “Sporting Goods” for intent-driven queries. Aim for at least four taxonomy levels.

Agentic Product Listings, written for the agent

Your Shopify product has two audiences. Human shoppers read your product page. AI agents read the Shopify Catalog.

The Agentic Product Listing (APL) is the version of your product written specifically for the second audience, stored as Shopify metafields, so you can write copy that converts humans without compromising what agents need, and vice versa. The APL never touches your human-facing page.

Creating one is a three-step flow from My Store: Prompts → Run → Review. Pick up to three shopper prompts - AgentIQ’s suggestions, your own Prompt Library or typed manually. AgentIQ then searches the Shopify Catalog for each prompt, studies the top-ranked competitor listings and writes your optimized version.

1. 1
   Pick prompts

   Up to three high-intent queries. Suggested, Library or manual - mix sources freely.

2. 2
   Run

   AgentIQ studies the top competitors per prompt and writes the listing. Run in background if you’d rather not wait.

3. 3
   Review

   Read every generated field. Check Top competitors surfaced to see exactly which listings it studied.

4. 4
   Refine

   Want a different result? Rerun via Edit prompts, the APL regenerates from the prompts you pick.