Data and operations

Website AI Reporting

Website AI reporting: analyze without Excel. Create segments, generate reports and send messages in seconds using natural language queries.

Start your 14-day trial Contact us

Customer segments

Customers who shopped in the last 90 days and have at least 3 orders

Last purchase: 90 days · Order count: 3 or more

Matching customer groupChampions (S1)

1,284customers

1,190 reachable

Net purchase value₺2,480,900

Purchase history → Audience definition → Customer list

Turn purchase history and natural-language queries into customer audiences for targeted campaigns.

Example scenario. Content and setup are tailored to your business.

How Do You Use AI Segmentation?

In the short narrated video below, you can see how to find a customer group with a single sentence, how the ready-made groups refreshed every night work, how products are chosen for each customer and how an automatic message flow is set up in the panel.

Find your customers with a sentence Type “loyal customers who have not placed an order in the last six months.” Pushouse AI turns it into a rule and immediately shows how many people match.

9 ready-made customer groups, refreshed every night Champions, Due for their next order, Starting to cool off, Waiting for a second order, Lost customers… The groups are recalculated from order history every night.

Every customer on their own schedule Messages do not go to everyone on the same day. The system calculates each customer’s own shopping interval and writes to them when their turn comes.

A storefront for every customer Product cards go out on WhatsApp, chosen from the categories each customer buys most and the price they usually pay, for example 3 dresses and 2 pairs of shoes. Products they have already bought are not shown to them again.

See it before you send Pick a customer and see in advance which cards they will receive and why. Nothing is sent from this screen.

Set it up once, let the system do the rest Messages go out when a customer joins the list, leaves it, or at intervals you set. A flow can have up to 5 steps, and it stops for a customer who places an order.

How Does Natural Language Reporting Work?

1

Describe the audience you want

For example, “loyal customers who have not placed an order in the last six months.”

2

The sentence becomes a rule

Read the rule on screen. If it is not right, edit it or try again.

3

The audience size is calculated

The total number of people, how many can receive messages and the group's net revenue are displayed together.

4

Turn the list into a campaign

Download it as an Excel file or send it directly to bulk messaging.

The workflow from writing a sentence to creating a campaign

How it works

01
Purchase history
02
Audience definition
03
Targeted campaign

For your industry

One module, tailored to the way your business works.

Beauty and Care

Build audiences around beauty categories

Use natural-language queries on purchase history to identify haircare or makeup shoppers, then prepare a campaign for the relevant audience.

Explore this industry

Choose campaign audiences by grocery category

Query purchase history in natural language to identify coffee or pantry shoppers, then prepare an offer for the relevant customer group.

Explore this industry

Sports and Outdoor

Build audiences around activity categories

Use natural-language queries on purchase history to identify camping, running or fitness shoppers and prepare campaigns for the relevant audience.

Explore this industry

Explore 2 more industry examples

Build pet-store audiences from purchase history

Use natural-language queries on purchase history to create audiences such as cat-litter or dog-accessory buyers. Prepare a relevant campaign for each group.

Explore this industry

Fashion and Apparel

Match a new collection to purchase history

Query purchase history in natural language to identify knitwear shoppers or buyers from another category. Build a relevant audience for your new-season campaign.

Explore this industry

Türkiye's First AI-Powered Reporting Engine

Pushouse Website AI Reporting instantly turns raw data into action. Without writing code or navigating complicated panels, you can create the segment you need in seconds simply by typing a sentence.

Website AI Reporting is an AI-powered Pushouse feature that turns e-commerce data into sales in seconds. Without complicated SQL queries, lengthy Excel formulas or cycles of downloading, transforming and uploading data, you can identify precise segments and turn them into campaigns just by writing a sentence.

Website AI Reporting

Why Website AI Reporting?

Natural Language Segmentation – Use a query such as “Customers who bought trousers in the last 60 days” to create a relevant audience from purchase history.
The Entire Process on One Screen – Segment → report → WhatsApp message: complete everything in minutes without leaving the Pushouse panel.

The Workflow in the Pushouse Panel

Type your request in natural language into the search box.
AI turns your request into an audience rule applied to purchase history.
A live segment list is created and appears immediately in the panel.
Send personalized WhatsApp messages through your Meta-approved WABA number with a single click.

Examples of Natural Language Queries

Customers who bought coffee in the last 30 days
Bought trousers in the last 60 days but did not buy a sweatshirt
Completed an order with a cart value over TRY 1,000
Customers whose last purchase was more than 90 days ago

Review the resulting audience and select it for your campaign’s purpose. Connect purchase history with a relevant offer and plan your message.

Platform Compatibility and Performance
Create customer audiences from purchase data imported from your store. Use details such as products, order dates and spending to select a relevant audience for your campaign.

Business Outcomes

An average of 42% higher conversion through behavior- and consent-based targeting
65% time savings by eliminating Excel and manual reporting processes
100% transparency across metrics including sending, reading, clicking and ordering

Get Started Now
Fill in the form and let's write your first Website AI Reporting query together. See live how Pushouse AI turns data into sales within seconds.

Checks That Prevent Messages Going to the Wrong Audience

Groups without consent cannot receive messages Groups made up of customers without communication consent are visible on the card but cannot become a messaging list.

Discounts are not recommended for frequent returners A permanent “no discount” warning appears on the cards for two groups, and these customers are excluded from the reachable audience count.

Your customer list is not sent to the model Only the sentence you type is sent to the AI.

See where the numbers come from Data health, historical validation and order reconciliation are available on separate screens.

Common questions

AI Customer Segmentation

Can I use it without typing a sentence?

Yes. If you prefer not to type, nine ready-made groups are available: champions, loyal customers with moderate purchase volume, customers due to order, customers starting to disengage, customers awaiting a second order, loyal customers with high returns, customers whose revenue is eroded by returns, high-value customers without consent and lost customers. Each card shows the number of people, how many are eligible for messaging and the group's net revenue; download the list as Excel or turn it directly into a bulk message.

Does the AI tell me not to send discounts to a group?

Yes. The two groups containing customers with frequent returns have a permanent “no discount” warning on their cards: sending these customers discounts is not recommended because it increases returns. The warning is more than a note; these people are also excluded from the estimated reach, so the panel does not promise a number you cannot actually reach.

Can I see the rule created by the AI?

Yes. Your sentence is first converted into a rule, then the number of people matching that rule is calculated; both steps are shown on screen. If you are not happy with the result, you can try again, return to an earlier point in the conversation or copy the rule and edit it manually. You are not asked to create a list without first seeing the rule.

Is my customer data sent to the AI?

No. The AI receives the sentence you type, not your customer list; the count and list are calculated against your own data. Groups containing customers without contact permission cannot be selected as messaging lists: they appear on the cards but are not enabled for sending. You are asked to confirm the KVKK declaration before using the module; obligations such as privacy notices and IYS registration remain yours as the data controller.

How can I trust the number shown in the panel?

The panel provides three separate checks: a data health screen showing whether your data was transferred completely, backtesting showing how the rule behaves on historical data, and reconciliation showing whether order totals in your panel match your store. You can see where a number comes from before relying on it.

Which e-commerce platforms does it work with?

It works with Ticimax, Ikas, Ideasoft and T-Soft. Shopify is currently deliberately outside its scope; if you have a Shopify store, you can build campaigns with other Pushouse modules. This module adds no code to your site and runs entirely in the Pushouse panel.

Let’s set it up for your business

Start querying your customer data in plain language today

No code is added to your site; everything happens in the Pushouse panel. Connect your store, describe your target audience in Turkish and turn the list into a campaign.

Start your 14-day trial Contact us