Product Agents explained: Types and automated emails

Product Agents combine behavioral data with Hello Retail’s Product Intelligence to schedule automated, personalized emails, deciding both timing and content based on what each customer has bought and how products relate to one another. Each agent type (post-conversion, replenishment reminder, and price drop) runs on autopilot, replacing complex manual flows with product-driven triggers. The result is emails that arrive at the right moment with a message built around the specific products each shopper cares about.

How Product Agents work

Most behavior-driven email tools fire messages based on what a shopper last did: they visited a page, abandoned a cart, clicked a link. Product Agents lead with the product itself. Hello Retail’s Product Intelligence model identifies how items relate to one another, which products customers tend to buy in sequence, and how long gaps typically fall between purchases.

That intelligence drives scheduling and content decisions. When a customer buys something, the agent doesn’t wait for them to signal interest again. It already knows, from aggregate purchase patterns, that a complementary category tends to come back into consideration after a specific number of days - whether that’s 10, 15, or 85, depending on the product type.

Product Agents sit alongside your existing email service provider as an add-on layer, working on top of the platform you already use. At launch, the integration is built for Klaviyo, so merchants already running Klaviyo flows can activate Product Agents on top of what they have today.

McKinsey’s Next in Personalization research found that companies excelling at personalization generate 40% more revenue from those activities than average players - a gap that keeps widening as consumer expectations for relevance rise.

The post-conversion agent

The post-conversion agent picks up where a purchase ends. When a shopper buys a pair of jeans, the agent looks at which products are typically bought next and how long customers wait before returning for them. If T-shirts tend to follow jeans purchases after roughly 10 days, and longer-sleeved shirts after around 15 days, the agent schedules both emails automatically, each timed to its own product-specific window.

This differs from a standard post-purchase flow in one important way. Traditional flows require merchants to define the rules by hand: if someone buys product A, send email B after X days. Product Agents derive both the timing and the product pairings from purchase data, so the logic updates as behavior shifts - without anyone editing a flow.

Salesforce’s State of the Connected Customer report found that 73% of customers expect companies to understand their unique needs and expectations. Timing an email to when a customer is actually in the market for a complementary product is one of the most direct ways to meet that expectation.

Replenishment reminders

Some product categories repeat on a predictable cycle. Mascara runs out after roughly 85 days. A facial cream might last 30. USB cables fail often enough that a proactive nudge makes sense well before a customer runs out. The replenishment agent automatically identifies which products in a catalog are replenishment items (those customers demonstrably buy again) and how long the typical repurchase window is for each one.

Once a customer buys a replenishment product, the agent schedules a reminder aligned to that window - without the merchant needing to tag products manually or configure category-specific delays. The product’s own purchase history drives the decision.

Replenishment revenue is highly predictable once the timing model is in place, and it’s revenue that often goes uncaptured when merchants rely on generic re-engagement campaigns. Litmus research on email marketing ROI puts average return at $36 for every $1 spent, but that average obscures the gap between well-timed, relevant sends and broadcast blasts. Replenishment reminders are among the clearest examples of relevance driving the difference.

Price drop agents: Three scenarios

Price drop emails are one of the more nuanced agent types because the obvious use case - shopper viewed a product, price fell, send an email - breaks down in several real-world scenarios. Product Agents address three of them directly.

Replenishment products on discount. Most price-drop systems suppress the email once a customer has already purchased the item, on the assumption that someone who just bought a flat-screen TV doesn’t need to know the price fell. That logic doesn’t hold for replenishment products. A customer who bought dog food last month and sees their usual brand drop in price is a strong candidate for an early repurchase. The agent recognizes that pattern and sends the email with a personalized message reflecting the customer’s purchase history, even though the item has already been bought.

Alternative products. A shopper browsed white sneakers, never bought them, and no price drop occurred on the exact pair they viewed. Meanwhile, a similar pair of white sneakers goes on sale. Standard price-drop logic sends nothing, because the trigger was the viewed product and that product’s price didn’t change. Product Agents use Hello Retail’s Product Intelligence model to identify alternatives and send a personalized email about the similar product that did drop - keeping the message relevant to what the shopper was actually looking for.

Upsell products on discount. A customer bought a PlayStation. A range of PlayStation games then drops in price. The agent identifies that this customer’s console purchase makes them a relevant audience for that discount and sends an email without broadcasting the offer to everyone in the database. The same principle applies to apparel: if a customer bought jeans and a compatible T-shirt goes on sale, the agent can send that specific customer a message connecting the discount directly to the jeans they already own.

Why the product-first model matters

The thread running through all three agent types is that the product itself, rather than the shopper’s last-observed action, is the primary signal. This changes what can be automated. Merchants don’t need to build and maintain increasingly complex decision trees to cover every product-purchase combination. The agent’s underlying model handles that inference, and the emails it schedules carry messages shaped around the individual customer’s purchase history.

For merchants on Klaviyo, the integration means these automated emails sit alongside existing campaigns without displacing them. Post-conversion sequences, replenishment reminders, and price-drop variations all operate as an additional layer of product-aware intelligence on top of whatever programs are already running.