Inside Product Agents: how predictive and generative AI finally meet in ecommerce

At Hello Retail, we’ve spent years building what we call Product Intelligence. Quietly. Iteratively. Across thousands of stores and millions of interactions.

This week, that work took a new shape.

We launched Product Agents — and with it, a different way of thinking about email marketing, personalization, and how ecommerce actually works when you zoom in.

To unpack what’s really going on under the hood, we sat down with our own data scientist, Sarah Miguel Cournane — recorded that same week, in February, as part of our Hello Retail Conversations series.

This is not a product announcement. It’s a look at the mechanics behind it.

Watch the Hello Retail Conversation with Sarah here →

The shift: from predicting behavior to acting on it

Ecommerce has used AI for years. It just wasn’t called that.

Recommendation engines. Search algorithms. Collaborative filtering. All of it falls under the same umbrella.

What’s changed recently is not just capability. It’s speed.

“We’ve gone from using AI to support decisions to actually verifying what AI produces for us.” — Sarah Miguel Cournane

That shift matters. Because it means we’re no longer just predicting what might happen. We can act on it in real time, at scale, and in a format customers actually see.

That’s where Product Agents come in.

Product Intelligence: the foundation most people overlook

Before you can generate anything meaningful, you need to understand what drives behavior.

That’s what Product Intelligence does.

Instead of relying on who a user is, it focuses on what they interact with.

• What products are viewed together
• What gets bought together
• What never overlaps
• What gets replaced, replenished, or abandoned

This creates a completely different foundation for personalization.

“If we don’t know who you are, we focus on what you engage with. That’s enough to understand intent.” — Sarah Miguel Cournane

Behind the scenes, every product is mapped into a high-dimensional space — not just as an ID, but as a set of characteristics and relationships.

So when a new product enters a catalog, it doesn’t start from zero. It inherits knowledge.

That solves one of ecommerce’s biggest problems: the cold start.

The reality most stores don’t see

Two numbers stood out in the conversation:

• ~30% of products drive ~70–80% of revenue
• ~50% of products disappear within six months

That means half your product knowledge constantly resets. Unless you connect the dots.

Product Intelligence does exactly that — by linking new inventory to historical behavior. So even when products change, understanding doesn’t.

Where Product Agents change the game

Most email setups today rely on flows.

You build them manually. You tweak them constantly. You guess timing, triggers, and content.

Product Agents remove that layer.

Instead of building flows, you define intent. Then the system decides:

• Who to contact
• When to contact them
• What product to show
• What message to send

All based on a combination of predictive and generative AI.

The real innovation: combining two types of AI

There’s a lot of noise around generative AI. But on its own, it’s not enough.

The real value comes from combining both:

• Predictive AI — what people are likely to want and when
• Generative AI — how to communicate it effectively

“It’s not a competition between the two. The combination is where things get interesting.” — Sarah Miguel Cournane

That’s exactly what Product Agents are built on.

Timing matters — but now you get both

Take something simple like replenishment.

Buying a sofa is a one-time event. Buying dog food is not.

That difference changes everything: the frequency of communication, the type of message, the timing of outreach.

Product Agents understand that automatically. Same with price drops, product alternatives, and inventory changes.

Instead of generic campaigns, you get messages tied to actual intent.

The hidden complexity: tone of voice across languages

One of the more unexpected challenges wasn’t technical. It was linguistic.

“Friendly” is not universal.

Friendly in Spanish ≠ friendly in German. Formal in Danish ≠ formal in English. What sounds natural in one language can feel off in another.

Tone of voice becomes a matrix: language, brand identity, context, and audience expectation — all working together, at scale.

That means building systems that detect and avoid repetition, balance friendliness without becoming childish, and stay on-brand without becoming overly formal. All without manually reviewing thousands of emails.

Merchant-side agents

Most of the current conversation around AI in ecommerce focuses on shoppers. Shopping assistants. AI copilots. Chat interfaces.

Product Agents take a different angle. They are built for merchants — but they serve the shopper.

“They help the shop understand what the customer wants — and help the customer find it faster.” — Sarah Miguel Cournane

Instead of waiting for the customer to act, the store becomes proactive.

What this actually means for ecommerce

We’re moving from:

• Static flows → dynamic decisions
• Segments → individuals
• Campaigns → continuous optimization

And maybe most importantly: from shops you visit, to shops that come to you.

Not in a spammy way. But in a way that feels relevant, timely, and useful.

Final thought

There’s a tendency to see AI as a layer you add on top. That’s not what’s happening here.

This is infrastructure. A different way of structuring how decisions are made in ecommerce.

Product Agents are just one expression of that. The underlying shift is bigger.

Full details and a breakdown of each agent are available on the Product Agents page.