AI Store Assistant PRO — an AI-powered assistant for OpenCart
AI Store Assistant PRO — an intelligent AI consultant and sales assistant for OpenCart 🤖
AI Store Assistant PRO — a professional artificial intelligence module for OpenCart and ocStore, which advises customers, helps them find suitable products, compares options, answers questions about the shop and guides the customer through to the product page or adding the item to their basket.
Unlike standard chatbots, the module does not delegate control of the product catalogue to a language model. The built-in Store AI Core independently checks products, categories, prices, availability, specifications, mandatory options and compatibility. The AI provider generates a natural response based solely on verified shop data.
It is a fully-fledged virtual assistant that helps customers choose products more quickly, reduces the workload on managers and turns visitors’ enquiries into actual sales. 🚀
🧠 Store AI Core — a reliable core with no made-up products
The main problem with conventional AI chatbots is the possibility of inventing a non-existent product, an incorrect price, specification or stock status. In AI Store Assistant PRO, this problem is solved by a separate server-side verification circuit.
First, Store AI Core analyses the customer’s enquiry, searches the catalogue, verifies the results found, and only then allows the AI provider to generate a response.
- Products are selected exclusively from the actual OpenCart catalogue;
- the `product_id` cannot be generated by the AI model;
- the price and availability are double-checked before a response is given;
- product specifications are confirmed against the product data;
- incompatible options are excluded;
- mandatory options are checked before being added to the basket;
- unverified statements are blocked or replaced with a safe response;
- product cards and links are generated by the shop’s server.
The assistant must not recommend an electric pump in response to a request for a manual one, diesel equipment for petrol, or a fuel pump for water without confirmed compatibility. Such restrictions are controlled by the catalogue, the industry profile and the recommendation validator.
🔍 Smart search across the entire catalogue
The assistant can search for products throughout the entire shop, not just on the current page. It analyses names, models, SKUs, categories, descriptions, attributes, options and other available data.
The search is carried out sequentially across several levels:
- exact match by model, SKU and name;
- search by categories and descriptions;
- full-text search;
- dynamic catalogue dictionary;
- synonyms and common typos;
- rules for the selected industry profile;
- checking of critical characteristics;
- optional semantic search by meaning.
The product name and its category take high priority. Attributes and descriptions are used as additional confirmation, so a random mention of a word should not turn an unsuitable product into a recommendation.
💬 Context-aware dialogue
AI Store Assistant PRO maintains the conversation thread and understands short follow-ups.
After products have been displayed, the customer can type:
- ‘Show me something cheaper’;
- “Which one’s better?”;
- “Do you have any that run on 220 V?”;
- “Where are they located?”;
- “Give me a link to the second one”;
- “Add this one to the basket”;
- “Let’s have another option.”
The assistant links the new question to previous products, parameters and the selected option. It does not restart the search unnecessarily and does not revert to a condition that has already been rejected.
The module retains the active dialogue language and does not mix Russian and Ukrainian text in a single response. ru-ru, uk-ua and en-gb are supported.
🛒 Product cards and secure shopping basket
Relevant products are displayed directly in the chat as handy cards.
- product image;
- current name;
- price;
- availability status;
- a link to the product card;
- available purchase options.
If a single, unambiguous product is found with no mandatory options, the assistant can safely add it to the basket following the customer’s instruction.
If the product requires a choice of size, colour, configuration or another mandatory option, the customer will be directed to the product page. Where there are several suitable options, the module first clarifies the selection and does not automatically update the basket.
Protection against message reprocessing prevents the same product from being added multiple times due to repeated AJAX requests.
⚖️ Product comparisons and clear recommendations
The module can compare several models and explain the actual differences between them.
The assistant takes into account price, specifications, intended use, limitations, configuration and other verified parameters. A more expensive product is recommended only when its advantages are genuinely substantiated by the catalogue.
If a more affordable option is sufficient for the customer’s needs, the consultant can state this directly. This builds trust and helps the buyer make an informed decision without high-pressure sales tactics.
📚 The shop’s permanent knowledge base
AI Store Assistant PRO creates its own permanent knowledge base within the shop’s database.
This may include:
- products and categories;
- manufacturers;
- specifications and options;
- information pages;
- delivery and payment;
- warranty and returns;
- contact details;
- articles and news;
- ocStore Blog;
- UniShop2 Articles and News;
- verified responses from managers.
Large documents are split into meaningful fragments. When a product, category, manufacturer or information page is updated, the relevant object is placed in the update queue.
The index is permanently stored in InnoDB and is not dependent on a temporary cache. During the rebuild, old documents continue to serve the chat until they are replaced by updated versions.
✨ Semantic search by meaning
For complex queries, you can enable an additional search based on embeddings. This helps to find relevant products and information even when a customer uses unusual phrasing or does not know the exact product name.
Semantic search is used only as an additional fallback after exact search, FULLTEXT, the catalogue dictionary and industry rules.
Embedding results do not automatically become recommendations. The products found are re-checked for type, specifications, availability, compatibility and valid product_ids.
This feature is disabled by default and can utilise OpenAI, Gemini or a compatible, secure HTTPS endpoint.
🧩 16 industry profiles
The module is suitable for shops across a range of sectors. By default, the neutral ‘Universal’ profile is used, which does not introduce any unconfirmed industry-specific assumptions.
The package includes the following pluggable profiles:
- pumps and fuel equipment;
- agricultural machinery;
- electronics and household appliances;
- clothing and footwear;
- cosmetics;
- furniture and homewares;
- building materials and tools;
- household and industrial chemicals;
- car parts;
- food;
- pet products;
- sport and tourism;
- children’s goods;
- stationery;
- personalised and laser-engraved items;
- software and digital goods.
A profile may contain industry-specific terms, synonyms, typographical errors, mandatory clarifying questions, compatibility rules, restrictions, knowledge and test scenarios.
Profiles can be installed, created, edited, disabled, deleted and restored from the standard package.
🤖 Integratable AI providers
Store AI Core can work alongside popular cloud-based AI services.
- Google Gemini;
- Anthropic Claude;
- OpenAI and OpenAI-compatible APIs;
- custom compatible HTTPS gateways.
You can configure a primary model, a backup model and a second backup provider.
The fault-tolerance chain operates as follows:
primary model → fallback model → fallback provider → Store AI Core → query to the manager.
A limited number of retries are provided for temporary errors, timeouts, HTTP 429 and HTTP 5xx errors. The circuit breaker temporarily disables an unstable service and prevents unnecessary requests.
❓ Unanswered queries
If the module cannot provide a safe and verified response, the query is not lost. It is stored in a separate section for processing by a manager.
The administrator can view:
- the customer’s question;
- the full dialogue;
- the language used;
- the current page;
- recognised parameters;
- reason for escalation to a manager;
- sources used;
- related products.
The manager can write a verified response and add it to the relevant section of the knowledge base. A customer’s response or unverified AI text does not automatically become part of the knowledge base.
👥 Lead management
Customer contact details are stored in a separate ‘Leads’ section.
The following statuses are available for enquiries:
- new;
- contacted;
- qualified;
- in progress;
- order placed;
- closed;
- out of date;
- spam.
Search, filters, pagination, bulk status changes, manager notes, editing of name, telephone number and email, viewing of chat history, and export to CSV or Excel-compatible XLS are all available.
The contact form can be opened manually, via a timer, or only after a customer requests to contact a manager. The assistant should not insistently ask for a phone number during a standard product search.
📨 Email and Telegram notifications
The module can notify managers of important events:
- a new lead has been received;
- a customer has requested a call or a manager;
- a question has been raised without a confirmed answer;
- a critical error has occurred with the AI provider.
Telegram notifications can be sent to a private chat or a staff work group.
📈 Sales and customer behaviour analytics
AI Store Assistant PRO helps analyse the actual effectiveness of the chat.
- widget openings;
- product card views;
- clicks on product cards;
- additions to the basket;
- leads generated;
- ratings of replies;
- related orders.
Store AI Core’s dedicated decision log shows the customer’s intent, the number of candidates found, verification results and the confidence level.
🔐 Privacy and security
The module includes tools for the secure handling of personal and technical data.
- consent to process contact details;
- customisable retention periods;
- anonymisation of old data;
- privacy audit of important operations;
- masking of API keys and tokens;
- masking of telephone numbers, email addresses, names and IP addresses in technical logs;
- protection of user API endpoints against SSRF;
- protection of exports against spreadsheet formula injection;
- verification of user_token and access rights for administrative actions.
Full messages, prompts, keys and contact details are not recorded in the module’s technical log.
🎨 Modern floating chat design
The widget can be customised to match the shop’s corporate identity.
- separate layouts for computers and mobile devices;
- Indent settings;
- choice of primary colour;
- consultant’s photo or icon;
- an avatar next to replies;
- text next to the floating button;
- attention-grabbing animation;
- quick questions;
- single-colour, minimalist and designer icons;
- multi-coloured versions with a transparent background;
- ‘Rainbow glass chat’;
- ‘Coloured feed’;
- responsive display on mobile devices.
All designer images are scaled to a uniform visual size. The minimalist icon uses a full SVG outline and displays correctly without a cropped bottom edge.
⚡ Optimisation for large shops
The module does not load the entire catalogue into the browser and does not send it in its entirety to the AI provider with every message.
- permanent queues in the database;
- batch processing;
- table indexes;
- server-side pagination;
- search and filters;
- execution progress;
- pausing and resuming indexing;
- processing via a secure cron job;
- batch loading of products without N+1 queries.
The admin tables do not use mass Bootstrap initialisation of tooltips and popovers for hundreds of repetitive elements. Lazy delegated processing reduces the number of JavaScript instances and the load on the browser’s main thread.
🏆 Why AI Store Assistant PRO is more powerful than standard AI chatbots
- ✅ The AI does not generate product_ids or product cards;
- ✅ prices and availability are checked via OpenCart;
- ✅ Specifications are verified against the catalogue;
- ✅ Incompatible products are excluded;
- ✅ It supports the entire journey from enquiry to basket;
- ✅ the context of the conversation is retained;
- ✅ a permanent knowledge base is available;
- ✅ Industry-specific profiles can be integrated;
- ✅ Unverified responses are forwarded to a manager;
- ✅ Leads, notifications and analytics are supported;
- ✅ backup models and providers are provided;
- ✅ Semantic search does not bypass catalogue checks;
- ✅ Data and secrets are protected;
- ✅ the module is suitable for shops across various sectors;
- ✅ Integration is carried out without directly modifying the core.
💼 Benefits for businesses
AI Store Assistant PRO helps reduce the number of repetitive queries, speed up product selection and provide advice even when a manager is unavailable.
Customers receive suitable options more quickly, see real product listings and prices, understand the differences between products and can proceed to purchase without having to spend time searching through the catalogue.
For the shop, this means:
- less workload for sales assistants;
- faster responses to customers;
- more clicks on product pages;
- more items added to the basket;
- collection and processing of potential customers;
- quality control of consultations;
- building up a verified knowledge base;
- an additional 24/7 sales channel. 🌐
⚙️ Compatibility
- OpenCart 3.x;
- ocStore 3.0.4.1;
- PHP 8.1–8.3;
- MySQL and MariaDB;
- InnoDB and utf8mb4;
- UniShop2 3.6.x;
- CodeCart PRO Theme;
- ocStore Blog;
- UniShop2 Articles and News;
- Bootstrap 3;
- Font Awesome;
- Multi-Store;
- ru-ru, uk-ua and en-gb.
🚀 AI Store Assistant PRO — more than just a chatbot
This is a comprehensive intelligent consultation and sales system for OpenCart: catalogue search, product specification checks, product comparisons, product cards, shopping basket, knowledge base, leads, notifications, analytics and quality control — all in a single e-commerce module.
Help customers find the products they need faster, reduce the workload on staff and turn every conversation into a sales opportunity. 🤖🛒
Important: before installing or updating, be sure to create a full backup of your website files and database. After installation, the module is set to OFF and does not run the shopfront logic, API, cron jobs, queues, indexing or frontend resources until manually enabled by an administrator.
