Corthex · Field guides
Work through the question behind the chatbot.
Ten practical guides for the decisions that make a store assistant useful: what it knows, what it may access and what happens next.
Start with the task you want to solve. Each guide includes concrete examples, boundaries and a next step. Worksheets and simulations stay local in the browser and do not connect to your shop or certify an installation.
Use the Help Center for current product controls and the integration pages for platform setup. The guides connect those controls to a reviewable process. Examples are fictional unless a published customer story is explicitly identified.
01 / Choose your starting point
Start with the store boundary.
Follow an order question through customer identity, connector setup and the real result. Use the diagnostic path when a lookup fails, or inspect the resource matrix before granting PrestaShop access.
Whose order is the chatbot allowed to see?
Test WooCommerce order-status support with signed-in customers, guests and another customer’s order. Understand identity, access and safe fallback.
An order is missing. Find the failing link.
A diagnostic path for WooCommerce chatbot order lookup: pairing, enabled tools, signed-in identity, ownership and unavailable store data.
Read the shop. Keep the write boundary.
Inspect Corthex’s PrestaShop Webservice resources, GET/HEAD permissions, customer context and read-only limits with a practical review matrix.
02 / Choose your starting point
Make the answer explainable.
Turn approved policies into bounded answers, give every source an owner and distinguish a source citation from current store data. These guides help you decide what belongs in knowledge and what needs a tool.
Explain the return policy. Don’t invent the return.
Build chatbot answers from your own return policy. Work through missing rules, exceptions, source conflicts and the difference between advice and action.
Give every important answer a home.
Prepare a webshop chatbot knowledge base with an editable source inventory, ten review questions, source ownership and a downloadable checklist.
A source tells you where. Fresh data tells you when.
Understand what a chatbot citation proves and when prices, stock or orders need live tools. Compare evidence types with a controlled freshness example.
03 / Choose your starting point
Make a deliberate pilot choice.
Explore custom-product advice without inventing options, then calculate a budget with explicit model, context and maintenance assumptions. These are planning tools, not forecasts of sales or guaranteed savings.
A chatbot budget you can explain.
Calculate Corthex usage credits from explicit model, token and conversation assumptions. Compare plans, prepaid purchases and maintenance separately.
Help someone choose. Keep the product facts exact.
Design AI advice for custom products: clarify dimensions, verify available options and obtain consent before a supported website action.
04 / Choose your starting point
Keep evidence with the launch decision.
Choose who owns the installation and record the scenarios that passed, failed or remain untested. A demonstration, a real storefront test and an authorized production release are different steps.
Before “live”, show what actually passed.
Run a practical store-chatbot acceptance matrix: sources, prices, own orders, guests, other customers, handoff and connector failures. Export your review.
One assistant. Two installation routes.
Choose between the Corthex WordPress plugin and JavaScript widget. Compare maintenance, consent, duplicate loading and WooCommerce access.
Turn one guide into a test you can repeat.
Choose a small pilot, identify an owner and keep the expected answer beside the source or tool result. Include unavailable data and refused access in the review. Useful support is not defined by a confident answer alone; the visitor also needs an honest next step when information or authority is missing.
Choose your integration