ThinkingAI

E-commerce

Help shoppers check out. Give them a reason to return.

For teams running a shopping app, ThinkingAI shows where shoppers leave, helps you follow up and measures which offers bring in orders.

A shopper checks her phone beside a stack of delivery boxes at home.
Analytics Agent
Example funnel

Checkout started → Order paid

Down 6 pts since Nov 3

More shoppers leave during checkout on the latest app version.

years
10
companies
1,500+
applications
8,000+

From the first product view to the next order.

ThinkingAI connects what shoppers do in your app with what they buy. See where checkout stalls, bring shoppers back and test which offers earn orders.

Analytics Agent

See where checkout stalls

Track views, carts and orders by app version.

Funnel Analysis Breakdown · App version
  1. Product viewed
  2. Added to cart
  3. Checkout started
  4. Order paid

Engagement Agent

Give shoppers a reason to return

Review cart reminders and first-order follow-ups.

Journey Abandoned cart
Entry audienceAdded to cart

No paid order

App pushYour cart is ready

Pick up where you left off.

Awaiting approval

A/B Test Agent

Find out which offer earns orders

Compare each offer’s order rate and revenue per shopper.

Experiment Offer comparison

Traffic split

A Free shipping 50%
B Free gift wrap 50%
Success metrics Purchase conversion Revenue per shopper

How we help your shopping app grow. Start with a shopper problem.

Recover abandoned carts

  1. Find abandoned carts

    Find shoppers who added an item an hour ago and haven’t placed an order.

  2. Draft a reminder

    Draft a push that reminds shoppers to return to their cart.

  3. Stop after purchase

    Check for a paid order before sending another reminder.

  4. Measure recovered orders

    Compare purchase rates for shoppers assigned a reminder or no message.

Analytics / FunnelsShop app
Cart to paid order
Added to cart
1,200
Paid orders
480
Conversion
40.0%
Purchase funnelLast 7 days
  • Added to cart1,200
  • Began checkout840
  • Placed order480
App pushDraft for review
Your cart is saved.

Open the cart to pick up where you left off.

Measure each promotion. Then follow the repeat orders.

See how much each placement and offer contributed to orders, and what promo buyers did next.

Credit each placement for the order

Black Friday orders, credit shared across banner, promo code, app push and email.

Attribution Analysis · Orders, Black Friday week
  • App home banner
    34%
  • HOLIDAY15 promo code
    29%
  • App push
    21%
  • Email via webhook
    16%

Serve a different offer per audience

Set an offer for new shoppers and a different one for returning buyers.

Config Center

shopper_offer

New shoppers
Free shipping
Returning buyers
Free gift wrap

Live

Turn promo buyers into an audience

Save HOLIDAY15 buyers as a cohort. Our Engagement Agent drafts their follow-up journey.

Behavioral cohorts

Bought with HOLIDAY15

8,412 buyers

Use in Journey

Compare promo and full-price buyers

Revenue per buyer by Day 30 and days to payback, by channel.

Revenue analysis · Day 30 by channel
CohortChannelRevenue per buyerPayback
PromoMeta Ads$61Day 48
Full priceMeta Ads$74Day 37
PromoGoogle Ads$58Day 41
Full priceGoogle Ads$71Day 30

Win the second order. Bring lapsed buyers back.

Track repeat rate from the first order

Our Analytics Agent marks each buyer’s first order date, counts every order after it, and reports 30-day repeat rate by channel and cohort.

How cohorts work
A shopper receiving her parcel.
Behavioral cohorts · Repeat buyersFirst order dateOrders 2+Repeat rate · 30 days
  • Email 22%
  • Organic search 19%
  • Meta Ads 14%
Win-back journeyDraft · Awaiting approval
Journey steps
  1. Lapsed buyersAt least two past orders · No paid order in 60 days
  2. Send app pushNew arrivals are here. Take a look.
  3. Wait 3 daysListen for a paid order
  4. Show in-app reminderAt the next app open
Exit condition
Order paidEnd reminders

Buyers who order leave the journey immediately.

Win back buyers who went quiet

Reach past customers who have stopped ordering. Review their push reminder before launch. The journey ends when they place an order.

Explore journeys

A/B test your offers. See which brings in more orders.

Give two groups of shoppers different offers on the same product. Compare purchase conversion and revenue per shopper before you pick one.

AFree shipping
Product detail
A terracotta espresso cup and saucer.
Espresso Cup & SaucerTerracotta
Free shipping
Add to bag
BFree gift wrap
Product detail
A terracotta espresso cup and saucer.
Espresso Cup & SaucerTerracotta
Free gift wrap
Add to bag
A/B Test Agent

Offer experiment

Compare the two groups before choosing an offer.

  • Primary metricPurchase conversionShare of shoppers who place a paid order.
  • GuardrailRevenue per shopperRevenue divided by shoppers in each group.
  • Decision checkConfidence intervalCheck the range of the difference between offers.

Every journey and test. One report on what changed.

Compare purchase rates for the reminder and holdout groups. For offers, compare the two test groups.

Reached
18,420
Orders
1,031
Purchase rate vs holdout
5.6% vs 3.2%
+2.4 pts
Revenue
$70,100

By node

  • Sent18,420
  • Opened5,710
  • Checkout2,140
  • Paid1,031

Payback by channel

  • Google Ads
    Day 31
  • Meta Ads
    Day 46
  • TikTok
    Day 58

In this example, 5.6% of shoppers assigned to the reminder made a purchase, compared with 3.2% in the no-message holdout.

Connect your shopping app to your order data.

Bring app activity, paid orders and ad data together so you can follow a shopper from first visit to repeat order.

Your shopping app

  • Shopping app · iOS and Android SDKs
  • Orders · server SDKs or REST API
  • History · DataX or Logbus
  • Ads · Meta Ads, Google Ads, TikTok
  • Attribution · AppsFlyer, Adjust, Branch, Singular

Agentic Engine

  • Data Collection Agent
  • Analytics Agent
  • Engagement Agent
  • A/B Test Agent

Reach shoppers

  • App push
  • In-app
  • WhatsApp
  • Webhook · your own email or SMS platform

Send email and SMS through your own platform, connected through a webhook.

Your coding agent → AE CLI

Or let your coding agent run the setup, using ThinkingAI skills, through our AE CLI.

Our agents draft every journey. Your team approves the send.

  • Cap messages per shopper per week.
  • End the journey when the order is paid.
  • Edit or reject any draft before launch.
Teammates reviewing a journey together at a monitor

Journey ready for review

Drafted by Engagement Agent

Audience
Added to cart, no order in 1 hour
Channels
App push, in-app
Frequency cap
2 messages per 7 days
Exit when
Order paid