Retail, marketplaces & DTC
Where is my order, answered before the queue forms.
Order status is the single largest contact driver in retail, and almost none of it needs a person. It is a lookup, a date, and sometimes a refund - repeated thousands of times a day, and concentrated into the weeks when staffing is hardest.
An illustrative reference architecture, not a customer
case study. No organisation is named and no deployment is described - this is
what the layers look like for a retail & e-commerce organisation,
so you can map it onto your own estate.
Channelswhere the conversation starts
Inbound numberspublished lines
SIP / contact centreexisting estate
Outboundcampaigns and callbacks
In-appbrowser WebRTC
▼
VoiceKernelthe agentic voice layer
Conversationagent, voice, turn taking
Knowledge groundinganswers with citations
Tool callsreads and writes back
Transferto a human, with context
Tenant isolationownership registry
Audit and erasureper call, per mutation
▼
Integrationyour boundary, your rules
REST + typed SDKscoped API keys
Signed webhooksretried, replayable
Live event streamSSE
▼
Retail & E-commerceyour systems of record
Order managementorders, fulfilment, shipping
Returns / RMAauthorisations and refunds
Paymentsrefunds and chargebacks
Inventorystock and availability
CRM / helpdeskticket history
Only the bottom layer is specific to retail & e-commerce. The
three above it are identical in every deployment, which is the point: the
voice layer does not need to know what industry it is in, only which systems
it is allowed to reach and what it may say.
What it changes
Mechanisms, not projections. Each of these follows from something structural
about the work, so you can check whether it holds for your own queues.
Peak capacity without peak hiringContact volume in retail is seasonal, but a trained agent is not hireable for four weeks. Voice absorbs the spike and releases it.
Hold time stops being a cost trade-offAnswering every call immediately no longer competes with the staffing budget, so service level stops being a quarterly negotiation.
Refund decisions become consistentPolicy is applied the same way at 3am on Boxing Day as at 10am on a Tuesday, and every decision carries the reason it was made.
Human attention moves up the value chainStaff stop reading tracking numbers aloud and start handling the complaints that actually retain a customer.
What the agent actually does
Order and delivery statusVerify the caller, read the current fulfilment state and the carrier estimate, and offer a text with the tracking link before the call ends.
Returns and refundsCheck eligibility against the policy, raise the RMA, issue the label, and tell the customer exactly when the money moves.
Peak-season overflowTake the calls that would have queued during a sale, and hand the genuinely unusual ones to the humans who are now free to take them.
What keeps it deployable
PCI scope stays where it isThe agent reads status, not card numbers; payment capture stays in your existing flow.
Refund limits enforcedValue thresholds live in tool availability, so the agent cannot approve beyond its authority.
Every decision auditableWhich policy rule fired, on which order, at whose request.
Order status, returns and delivery enquiries dominate retail contact volume, and all three are lookups against systems you already run.
NextRead the source
Everything above is in the open-source Community edition. Nothing on
this page needs a licence.
NextRun it yourself
Point it at your provider account and put one queue through it before
you talk to anybody.
NextThen buy the guarantees
An Enterprise agreement adds the SLA and the
support contract when it carries real volume.