AI Agents · Recommendation Engine
Recommendations — in your email, on your site, anywhere.
One recommendation engine — one brain — powers every surface. The same personalized picks show up inside your emails, on your own website and online store, and in any app you build. Each member sees products drawn from what they actually lean toward. New and anonymous visitors see trending items, so a recommendation slot is never empty.
One engine, three surfaces
The same recommendations, wherever your customer is.
You don't build a different recommender for email, for web, and for your app. It's one engine — so the picks stay consistent no matter where a shopper meets you.
In your emails
Drop the recommendation block into a campaign and every recipient gets products picked for them — filled in at send time, one block, no manual list.
The email block→On your website & store
Show live, personalized picks right on your own site or online store, in the shopper's browser — a recommendation strip on the homepage, a product page, or the cart.
In any app via API
Call the engine directly from any app or channel you run. Same recommendations, same logic — wherever you need a personalized product slot.
The API→How the personalization works
Known member? Their affinity. New visitor? Trending.
Personalization runs off each member's own affinity — the categories and products they lean toward — read straight from your unified member graph. And when the engine doesn't know the visitor yet, it never shows an empty slot.
Recognize the shopper
If the visitor is a known member, the engine identifies them and reaches for their affinity. If they're brand-new or anonymous, it falls back to trending instead.
Read their affinity
For a known member, the engine uses the categories and products they lean toward — built from their real history in your unified member graph.
Pick from your catalog
Recommendations are drawn from your live product catalog, so every pick is a real product with its current price and link.
Fall back to trending
For a first-time or anonymous visitor, the engine shows trending and popular items — so the slot is always filled and useful from the very first visit.
The picks come from your live product catalog, so prices and links are always the real ones.
Bought together & cross-sell
Related products, proven by your own orders.
The "bought together" engine scores product pairs so that genuinely related products surface — instead of the bestsellers that would otherwise show up next to everything.
Beyond bestsellers
Pair scoring corrects for popularity, so a product earns its place next to another because the two are actually bought together — not because one of them is bought with everything.
Rules mined nightly, numbers attached
Cross-sell rules are mined every night from your real orders, and each rule carries three auditable numbers: how often the pair occurs, how often buyers of one product go on to buy the other, and how much more often than average. In the product it reads like: “Buyers of X go on to buy Y in 23% of orders — 3.1× more often than average.”
Suggestions that know when to stay quiet
The rules feed a cross-sell suggestion card on active buyers' member profiles. It skips products the member already owns, and when the data can't support a suggestion, the card simply doesn't render — no fabricated picks.
Every rule has to clear a minimum-evidence bar — a pair needs multiple shared buyers and orders before it counts — so thin data never masquerades as a pattern.
Drop-in setup
Live on your site in three steps.
No data pipeline to build. To use it in the browser you create a public key — a key that's safe to use on a web page — decide which of your website domains are allowed to use it, and paste in one small snippet.
Create a public key
Make a public key — one that's safe to expose on a web page. It's designed to sit in your website's code without putting anything sensitive at risk.
List your allowed domains
Choose exactly which of your website domains are allowed to use the key. Anywhere else is turned away, so the key can't be abused from a site that isn't yours.
Paste one snippet — or call the API
Add one small script snippet to your site and the recommendation strips light up. Building an app instead? Call the API directly with the same key.
Because you list the allowed domains yourself, the browser key only works on the sites you name — and nowhere else.
Why it matters
One brain, so the story stays consistent.
The recommendations a shopper sees in an email and the ones they see on your site come from the same engine — so the experience is coherent, not two systems disagreeing with each other.
Consistent across email and web
Because it's the same engine everywhere, a member's picks stay consistent between their inbox and your website — one coherent experience, not a patchwork.
No data pipeline to build
There's nothing to wire up or maintain on your side. Create a key, name your domains, paste a snippet — the engine and your catalog do the rest.
Works for anonymous shoppers too
First-time and not-yet-known visitors still get relevant trending picks from the very first page view, so no slot ever sits empty.
It reads affinity from your member graph, picks from your product catalog, and powers the email recommendation block — all from one engine.
The same personalized product picks, drawn from each member's affinity, across email, web and app.
Related reading
Keep exploring the pages most related to this one.
Intelligent Decision
A next-best-action: who, which offer, which channel, when.
Read more CapabilityDashboard
Real-time scan-to-redemption data as KPI trends, funnel diagnostics, and a rule + LLM next-best-action panel — plus retention curves by signup cohort and a 14-day revenue outlook that's honest about being a projection, not a commitment.
Read more CapabilitySoTag for Slack
The Slack agent that hands governed, audited numbers to Claude in-thread: daily anomaly scan + human-approved coupon-send. Early access.
Read more CapabilitySoClaw — AI Win-Back
An autonomous AI teammate that wins back high-value, at-risk customers one at a time — a personalized offer or a deliberate hold — and proves the real incremental lift against an untouched control group. Its worklist is ranked by where outreach changes the outcome — revenue at risk × each member's measured persuadability from the program's own control group. Off by default, capped, instantly pausable. Early access.
Read more CapabilityAI Governance & Approvals
Every AI action runs on a per-action authorization gear — automatic, human review, or forbidden — with an approvals console, a complete decision log, one-click undo, and automatic demotion to human review when an action starts misbehaving.
Read more CapabilityBusiness Ontology
The concept behind the governed AI stack: a world model of your business — objects, properties and relationships plus governed actions — as a metadata layer mapped over the data you already have, so AI knows what a customer is, what has happened, and what it may do about it.
Read more