The right offer. The right shopper. The right moment.
You decide which offers each shopper can get. Engage uses machine learning, trained only on your store’s data, to decide the order and the way those offers are presented, for the highest chance of a sale. Guard keeps every offer within your rules.
You decide the offer. Machine learning decides how to present it.
Its job is to make your offers as relevant as possible to each shopper.
The offer
- Which promotions exist
- Who qualifies for each one
- The rewards, timing and limits
- How offers stack
The presentation
- The order of a shopper’s eligible offers
- Which one to lead with
- How the message is worded for this shopper
- What to highlight: saving, points or the complete set
The control
- Pre-checks each offer before it is shown
- Full check on the final cart before checkout
- Margin floor, stacking, budgets and limits
- Logs every decision
Engage never decides which shopper gets which reward, and the AI never creates an offer or sets a price.
Industry research, not PerkQi results. Sources: McKinsey, Next in Personalization 2021 (US consumer survey, n = 1,013) · Twilio Segment, State of Personalization 2023.
Every shopper is different. Their offers should feel different too.
Engage’s model looks at what it knows about the shopper and what they are doing right now.
It keeps learning from your store’s own data, never from other merchants, so the model and how it behaves are unique to your store. Before your store has enough history, personalization starts from the initial model and is less precise. Predictions can improve as real shopper interactions and orders build up.
Three shoppers qualify for the same offers. Each sees them differently.
The merchant has made three offers available to all three shoppers. Engage leads with the one each shopper is most likely to act on, and words it for them.
Example. See the full example →
Personalization that happens while shoppers shop
When she arrives
She sees the offers she qualifies for, in the order most relevant to her, and the points already waiting for her.
On product pages
She sees her offer price and details, written for her and pre-validated by Guard. The final discount is confirmed against her cart at checkout, in case the cart, eligibility or budget has changed.
In the cart
A short message, written for her, nudges her to complete the offer, such as adding the second item of a set.
At checkout
Guard runs its full check on the final cart. The reward is then applied automatically, with no code to type.
Illustrative screens with example data.
Each visit teaches your store’s model a little more
Engage records which offers were shown, how shoppers responded, and what they bought. Your model learns from these results, and you see them in your dashboard.
Shown
Which offers each shopper saw, in what order and with what message.
Responded
What the shopper did next: viewed, added to the cart or moved on.
Bought
Orders and rewards used, checked by Guard.
Learned
Your store’s model updates from these results, and only from your store.
The next visit starts from what the model has learned.