FAQ: Faboxi and modern ecommerce customer service

Product and use case

What is Faboxi?

Faboxi is an ecommerce support workspace connecting channel conversations with customer, order, shipment, and AI-draft context in one operator view.

Who is Faboxi for?

Teams that repeatedly handle order, delivery, return, or complaint questions across several channels or commerce systems.

Does Faboxi replace WMS, ERP, or the store platform?

No. Those systems remain the source of order, payment, and fulfilment truth. Faboxi uses their data in the support workflow.

Is Faboxi a CRM?

Faboxi keeps the customer profile and contact history needed for support, but it is not designed to replace a full sales CRM.

Channels and integrations

Which channels can Faboxi support?

The architecture covers email, web chat, Messenger, WhatsApp, SMS, and voice. Publicly available and planned integrations are labelled on the product page.

Which commerce systems can connect?

The connector model covers WMS and platforms or aggregators such as IdoSell, WooCommerce, Shopify, Shoper, BaseLinker, and PrestaShop. Actual scope depends on the active connector and provider data.

Does Faboxi need store database access?

No. Data flows through versioned APIs and connectors without direct WMS or CMS database access.

What happens when an integration fails?

The operator should see a degraded state next to the affected context. Missing or stale facts must not be presented as current.

AI and automation

What does AI do in Faboxi?

AI can summarise, classify intent and risk, explain context, and draft a reply from allowed data.

Does AI send without an operator?

Not by default. An operator approves, edits, or rejects the draft. Greater autonomy requires separate policy and strict safety gates.

Can AI invent order status?

It must not. Order status comes from a current, appropriate commerce source. Missing or conflicting data limits the reply.

Can AI cancel or refund an order?

Not without a controlled process requiring preview, policy, permission, idempotency, audit, and connector confirmation.

Do operator edits automatically change the AI?

No. Feedback enters a controlled quality process. Shared-engine changes require evaluations, while store-specific settings remain tenant-scoped.

Implementation, safety, and billing

How does implementation start?

Choose one frequent process, required channels, and data sources. Launch a safe scope and verify real scenarios before expanding.

Can we start read-only?

Yes. It is the preferred way to validate commerce data and operator workflow before enabling order-changing actions.

How is SaaS customer data isolated?

Data, memberships, integrations, secrets, and background jobs are designed with workspace context. Provider secrets remain server-side.

How are AI Actions billed?

Allowance depends on the plan. AI usage is classified from the work outcome, while displaying order data is not an AI action.

How do we start?

See the demo with a specific store scenario, then choose the first process, integrations, and rollout scope.