Oloye.

For ecom (£200+ AOV)

WISMO tickets are 40% of your inbox. They're also 90% solvable without you.

40 percent of your support tickets are the same three questions. Where's my order. How do I return this. Does this work with the other product I bought. Each one takes four minutes of your (or a VA's) time. Multiply by 200 tickets a week and support is eating a working day, five days a week.

Ecom customer service is a well-worn workflow with clear finish lines: track the order, answer the size question, refund the small return, escalate the chargeback. The Front Desk reads every inbound ticket, pulls order data from ShipStation or Shopify, drafts the reply in your brand voice, and takes the next action under your rules. Refunds under threshold. Return labels sent. Edge cases surfaced.

Test the Front Desk on your business
Ecom brand owner at a packing table with a laptop showing a lime notification — AI support agent for ecom brands

What the Front Desk does for ecommerce brands (£200+ AOV)

  • Reads Shopify inbox, Gorgias, Klaviyo Inbox and plain email tickets.
  • Pulls live order status from ShipStation, AfterShip or the Shopify API.
  • Drafts every reply in your brand voice, trained on your last 100 messages.
  • Refunds under your set threshold via the Shopify Admin API.
  • Files returns, emails the return label, updates the order tag.
  • Escalates chargebacks, complaints, VIP customers, and any request above your threshold.

Three real workflows

What the Front Desk actually replies with and does on your behalf. Every draft is trained on your last 100 messages. Every action respects the thresholds you set.

Workflow 1

Inbound

"Where's my order? Placed it Monday, tracking hasn't moved."

Front Desk reply

"Hi Sam, just checked. Your order shipped Tuesday and is currently in transit with Royal Mail (tracking RM12345678GB). Latest scan is at the Milton Keynes hub, estimated delivery Thursday. Sometimes RM tracking is a day behind reality — I'd expect it to update tomorrow. If it hasn't moved by Friday, reply here and I'll open a trace."

Action taken

Pulls tracking from ShipStation, checks last scan date, drafts a personalised reply with real data, tags the ticket "WISMO-handled".

Workflow 2

Inbound

"This didn't fit. Can I get a refund? Order #ABC123, £42 hoodie."

Front Desk reply

"No problem. Refund of £42 is being processed to your original payment method now, should show up in 3-5 working days. I've also emailed you a free return label (no rush — you have 30 days). Sorry it wasn't the right fit."

Action taken

Verifies order is within return window and under refund threshold, processes refund via Shopify API, emails return label, tags the ticket.

Workflow 3

Inbound

"Your product broke after two weeks. This is unacceptable. I want a full refund and I'm posting about this."

Front Desk reply

[Draft held for your review] "Hi Chris, I'm so sorry to hear that. That is not the experience we want you to have. I've flagged this to the owner, who will respond within a few hours. In the meantime, can you send a photo of the fault so we can process the resolution properly?"

Action taken

Detects negative sentiment + threat of public post, drafts the empathetic reply but holds, alerts the owner immediately.

The rough numbers

Baseline

Typical £500k-£3M ecom brand: 150 tickets a week, ~4 minutes human time each = 10 hours a week.

The loss today

At £25 an hour that's £1,000 a month, before you factor in slower response killing CSAT and the chargebacks that come from ignored complaints.

What the Front Desk recovers

The Front Desk handles ~80% of tickets end-to-end and escalates the rest. Real recovery is time back plus higher CSAT plus faster refund resolution. Your numbers will differ; we log the real ones on your account.

Further reading

Independent sources on the state of ecommerce brands (£200+ AOV), why response speed matters, and what the incumbents are doing.

  • Shopify blog

    Vast published research on ecom customer service benchmarks and CVR impact of response time.

  • Klaviyo Learn

    Data on email flow performance and support-driven retention.

  • Baymard Institute

    Independent UX research institute. Widely cited on cart abandonment and post-purchase support pain.

Common questions

Can it refund without asking me?+

Only under the threshold you set (default £50). Anything over holds for your yes. You'll never see a rogue refund from the agent.

Does it know my brand voice?+

Trained on your last 100 replies plus your product docs and returns policy. Voice match is measured in the eval set before we go live.

Does it work with Gorgias, Klaviyo Inbox, and Shopify Inbox?+

All three, plus plain email and Freshdesk. Any inbox with an API or forwarding address.

Related industries

Same read-decide-act loop. Different action list.

New to agentic AI?

Start with the full pillar guide on how these systems work under the hood: architecture, thresholds, evals and how to build one.

Read the pillar: What is an agentic AI system →

Free test

Test the Front Desk on your own business.

Send 10 real messages from your last week. We show you your actual response times, what the Front Desk would have replied, and what action it would have taken. Yours to keep either way.

Book my test