AI-powered discounting app, Promi (from the team at Uber who built personalized promotions)

Shopify Apps
,

Topic summary

Promi is a newly updated Shopify app offering AI-powered discount personalization, built by former Uber team members who developed personalized promotions that generated 30%+ more revenue.

Core Features:

  • AI personalization that selectively applies discounts only to customers who need incentives to convert
  • Rule-based targeting for discount campaigns
  • Automatic price strikethroughs, sales badges, and messaging for code/automatic discounts via deep links
  • A/B testing capabilities to measure ARPU and gross profit impact

How It Works:
The machine learning model analyzes three data layers:

  1. Session context (traffic source, device, time)
  2. User context (browsing/purchase history)
  3. Shop context (product conversion rates)

Example: Traffic from Meta/TikTok ads typically receives discounts due to lower conversion rates, while direct URL visitors (higher intent) don’t receive discounts.

Pricing Model:
Fixed fee structure without commission on discount strikethrough features. Includes a 21-day free trial, allowing ~1 week for model training and time for A/B testing. Setup takes approximately 10 minutes.

Case studies available on their website demonstrating impact.

Summarized with AI on October 26. AI used: claude-sonnet-4-5-20250929.
1

Hi Shopify community!

We’ve just launched a big update for our app, Promi. Think of Promi as a personalization platform for your discounts. You can:

  • Personalize your discounts with AI (aka only give discounts to the customers who need the extra nudge)
  • Build out rule-based targeting for your discounts
  • Trigger price strikethroughs, sales badges, and messages for automatic discounts or code discounts via deep link
  • Run A/B tests to understand ARPU, gross profit, etc. impact

Check us out on the app store! Promi AI: Personalized Offers - Create personalized discounts powered by AI | Shopify App Store

AI Personalization:

Fun fact, Uber generates >30% more revenue from its promotions with personalization vs. non-personalized campaigns (we built the tech).

We’re taking a lot of our learnings from Uber to create a simplified machine learning model designed to make merchants more profit. There is quite a lot of technical details I could go into here, but the basics are that we collect data on:

  1. Session context: traffic source (Google ads, Facebook ads, direct, email, etc.), device, app vs. web, time of day, day of week, etc.
  2. User context: visitor view history, visitor transaction history, etc.
  3. Shop context: product conversion rates, product profit margin, etc.

With all this data, we run a model to predict how likely you are to purchase a product. If you are very unlikely to buy, we give you a higher discount. If you are likely to buy, we don’t give you a discount. Take a quick example: for most merchants, traffic from Meta and TikTok ads has lower conversion. Promi typically discounts traffic coming from these channels, while not discounting visitors who type in the URL of your website directly.

We’ve got a few great case studies on our website (Case Studies) to give you a sense of the impact.

Discount Price Strikethroughs:

The other, simpler reason many merchants use us is that we can trigger discount price strikethroughs and UX for any of your code or automatic discounts. Links from email, ad, or other campaigns can then show personalized sales.

Better yet, we don’t charge a commission on this feature. It’s a low fixed fee (unlike other apps).

We have a nice, long 21-day free trial so you have plenty of time for our model to train on your store (usually takes 1 week) and run an A/B test before paying anything. For most merchants, setup should only take 10 minutes.

Thoughts and feedback are very welcome!