Industries · E-commerce and retail
An online store without the chaos,customers answered fast.
WEB-AZ builds digital and AI solutions for online stores and omnichannel retail: assistants that answer from your real data, stock and catalogue up to date on every channel, and orderly returns. Your team handles the exceptions.
Where a growing online store can save time
The same questions with every order
“When will it arrive?”, “Do you have it in a 42?”, “Can I change the colour?” Messages come in over chat, email, Facebook and phone, especially during campaigns such as Black Friday. A quick answer drawn from the order data keeps the customer close to buying.
One stock level across every channel
The same product sells on your website, on eMAG and in your physical shop. When the ERP, the warehouse and the marketplace update in step, there are fewer orders for out-of-stock products and fewer cancellations that hurt your rating.
Returns tracked from request to refund
Return requests arrive in different formats, with photos and wrong or missing order numbers. A clear workflow checks them against the withdrawal period, the product’s condition and the store’s policy, then tracks them through to the refund.
A complete catalogue that is easy to keep current
Descriptions copied from the supplier, missing attributes, photos without alt text and feeds rejected by the marketplace. With product data structured once, every new product reaches every channel correctly.
One answer about the parcel, whatever the courier
One AWB with Sameday, another with Fan Courier, a third with Cargus. Statuses gathered in one place give the customer a clear answer, with no manual searching across three platforms.
Marketing data brought together
Orders sit in the store platform, subscribers in the email tool, campaigns in the ad accounts. A shared view shows which customers come back, which products sell together and who you are allowed to contact.
Digital and AI solutions for online stores
We start with the workflow that takes up the most time and connect it to the systems you already use. Your team remains the one that decides in unclear cases.
- AI assistant for your store’s customersAnswers questions about products, sizes, delivery and returns based on your catalogue and policies, checks the order status and hands the conversation to a person when a request falls outside the usual pattern.
- Triage for after-sales supportMessages about returns, delayed parcels or faulty products are classified, enriched with the order data and sent to the right colleague, with a suggested reply they can edit.
- Sync between store, ERP and marketplaceIntegrations that move stock, prices and orders between the e-commerce platform, the invoicing software, the warehouse and marketplaces, with an error log so you know what failed to go through.
- An online store built for conversionA fast site, filters that genuinely help people choose, complete product pages and a short checkout. The foundation that automation can be added to later.
- Visual analytics in the physical shopFor retailers with physical premises: aggregated visitor flows, checkout queues and empty shelves flagged from your existing cameras, without identifying customers.
- Marketing built on your sales dataSegments built from order history, email campaigns sent only to subscribers who have given consent, and reports that link ad spend to real sales.
Use cases · E-commerce and retail
Conceptual examples showing how these solutions can be used. They do not describe delivered projects.
A campaign weekend without a chat queue
A footwear store receives many questions about sizes and delivery times during a campaign. The assistant answers from the size chart and the delivery rules, checks stock for the requested size and leaves only exchanges and complaints for the team.
Conceptual exampleOne stock level for the website and eMAG
A home appliance retailer sells on its own website and on eMAG Marketplace. An integration reads stock from the ERP at set intervals, updates both channels and flags products where the differences exceed a threshold set by the manager.
Conceptual exampleA return opened straight from the conversation
A customer writes that they want to return a jacket. The assistant identifies the order, checks whether the request falls within the withdrawal period stated in the store’s policy, generates the return form and creates the task for the warehouse. The refund is still approved by a person.
Conceptual example
Integrations for online stores and retail
A store rarely runs on a single platform. The list below shows systems that integrations can be built with; compatibility and API access are checked for each project individually.
- WooCommerce
- Shopify
- Magento / Adobe Commerce
- eMAG Marketplace
- SmartBill or Oblio
- Local ERPs (for example WinMentor, SAGA)
- Courier APIs (Fan Courier, Sameday, Cargus)
- Google Merchant Center
- Email marketing platforms (Mailchimp, Klaviyo)
Integration challenges
- Local ERPs without an API, where data can only be extracted through a scheduled export or direct database access
- Different product codes at the supplier, in the ERP and on the marketplace, which must be mapped before any sync
- Return and delivery policies that exist only in the team’s heads and need to be written down clearly before an assistant can use them
- Call limits and feed rules imposed by marketplaces, which affect how often data can be updated
Consumers, GDPR and e-Factura in e-commerce
Consumer rights
Answers about returns, warranties and delivery times must follow the published policy and consumer protection rules (ANPC, the National Authority for Consumer Protection). The assistant does not promise refunds or exceptions that only the team can approve.
Consent for marketing
Promotional emails, text messages and notifications go only to people who have explicitly agreed to receive them, and unsubscribing must be simple and honoured across all connected systems.
Customer data in conversations
Addresses, phone numbers and order history are personal data under GDPR. We define what data the assistant sees, how long conversations are kept and which AI providers process the information.
Invoicing and e-Factura
Automations that touch invoices must respect your RO e-Factura (Romania’s national e-invoicing system) workflow and your invoicing software. Tax data is not changed by an AI model, and transmission errors reach a responsible person.
Projects in this sector
Useful guides
Before you decide, read
- E-commerceWooCommerce or Shopify: which platform to choose for an online store in RomaniaWooCommerce gives you full control but needs maintenance; Shopify is easier to start with but ties you to the platform's rules. We compare costs, local integrations, flexibility and growth.Read the guide
- SEO & marketingCore Web Vitals explained simply: LCP, INP, CLS and website speedWhat Core Web Vitals measure, which values count as good, how to check them in PageSpeed Insights and Search Console, and the usual causes of a slow website.Read the guide
Frequently asked questions
Which AI solutions suit an online store?
Most often: an AI assistant that answers customers from the catalogue, policies and order data, triage of after-sales messages, stock sync between the store, ERP and marketplace and, for physical shops, aggregated visual analysis of visitor flow. We start with the workflow that takes up the most time.
Can the assistant make up information about products or delivery?
The risk exists with any language model, which is why we limit it to approved sources: the catalogue, the published policies and the order data. When it cannot find the answer there, it says it does not know and hands the conversation to a colleague instead of guessing.
What happens to my customers’ data?
The data stays in your systems, and the assistant receives only what it needs for the current question. We choose providers that do not use the data to train their models, set up role-based access and document everything for your GDPR records.
Does it work if I only sell on marketplaces?
Yes, many integrations start from the marketplace’s own API, for orders, stock and customer messages. We check at the outset what each platform allows, because their rules and limits differ.
Where is it worth starting?
With a short analysis of the messages from the last few months: which questions repeat, how long replies take and where orders get lost. Usually a single workflow stands out, such as order status or returns, that is worth automating first.
Can’t find the answer? Write to us and we’ll reply directly.
Let’s see where your store can save time
Tell us which platform you sell on, which channels you use and what takes up most of your time. We will come back with a concrete proposal for the first workflow to automate.