Mike HeapMike is an experienced Product Manager who focuses on all the “non-development” areas of My AskAI, from finance and customer success to product design, copywriting, testing and more.
Table of Contents
- Baby textiles AI customer service results on Gorgias at a glance
- What does the brand do?
- Which helpdesk does the brand use?
- How did the brand train their AI customer service agent?
- When did they decide to turn on 'direct replies' to customers?
- What was the biggest thing they did to improve their AI agent's resolution?
- How does the brand customize their AI agent setup to work for their business?
- Setting communication, clarification and handover guidance
- Replying in every customer's language
- What impact is their AI customer service agent having now?
- Where does the brand go from here?
Created time
May 29, 2026 09:08 AM
Title length (<60)
Author
Last optimised
Sep 15, 2026
Ecomm?
Tags
Image
Publish date
May 29, 2026
Video
Slug
baby-products-ai-customer-support-case-study
Featured
Featured
Type
Case study
Ready to Publish
Ready to Publish
A DTC baby-textiles brand resolves 72% of around 1,000 monthly Gorgias tickets with AI, holding a 94% AI CSAT score and saving roughly 59 hours of agent time every month.
For a growing DTC brand, the support inbox tells the same story on repeat. Where's my order, can I still return this, which tog rating do I need for a spring baby? For this brand those questions land all day and in several languages at once, far more of them than a lean team can answer by hand without either hiring up or letting reply times slip.
Today our AI agent handles the bulk of that load right inside Gorgias. It resolves 72% of around 1,000 tickets a month, holds a 94% AI CSAT score, and hands the team back roughly 59 hours a month they used to spend on repetitive replies. The AI answers about 200 of those tickets from help articles it wrote for itself.
Here's how it came together.
Baby textiles AI customer service results on Gorgias at a glance
Fact | A DTC baby-textiles brand |
Industry | eCommerce, baby and family textiles (DTC across several European markets) |
Helpdesk | Gorgias (live chat and email) |
Support volume | ~1,000 tickets/month |
AI resolution rate | 72% (around 720 tickets a month) |
AI CSAT | 94% |
Time saved | ~59 hours/month |
Key features used | Self-Learning, Shopify integration, User Data API, Guidance across all three types, automatic language detection |
Knowledge sources | Help centers via the Knowledge Base connectors, an uploaded library of internal Word docs, Shopify and live User Data API lookups, Self-Learning auto-drafted articles |
Previously evaluated | Gorgias Automate (ruled out on price for their volume) |
Go-live mode | Internal-note mode first, then direct replies on chat and email |
What does the brand do?
The company is a baby and family textiles brand selling across several European markets. The range starts with baby sleeping bags (their hero product) and runs through crawling mats, clothing, nursery basics and full newborn starter sets.
The range is built on natural materials and positioned at the premium end. That parent-facing positioning matters for support (the tone of every reply has to match the care that goes into the products), and it has to land that way in whatever language the customer wrote in.
Which helpdesk does the brand use?
They run their support on Gorgias, across both live chat and email. Gorgias is the Shopify-native default for DTC brands, so it sits right at the center of their stack.
They'd looked at Gorgias' own native automation, Automate, first, but it was too expensive for the volume they were dealing with. So they went hunting for an alternative that could slot into the same Gorgias inbox without a rebuild, and found us in the Gorgias App Store. The pitch was simple: drop an AI agent into the helpdesk they already ran, at a price that made sense for their volume, and have it reply in every language their customers use.
How did the brand train their AI customer service agent?
An agent is only as good as what it knows, so they wired up several knowledge sources before letting it loose.
First, their existing help centers came in through our Knowledge Base connectors, which gave the AI the published FAQ content as a baseline. Then they uploaded a big library of Word docs: the internal background that had never made it into a public help center but held the answers their agents leaned on every day.
(That mattered more than it sounds: a lot of what their agents knew had never been written down anywhere public, and getting it into the AI is often what separates a 40% resolution rate from a 70% one.)
The highest-leverage piece was order data. They connected Shopify and set up our User Data API so the AI could pull live order and customer info.
For an ecommerce brand, that's the unlock: the AI answers "where's my order?" on the spot, with the customer's real order status. The Shopify integration handles multiple languages, currencies and stores natively too, which suits a brand selling across several markets.
When did they decide to turn on 'direct replies' to customers?
They didn't flip the AI to customer-facing on day one. They started it in internal-notes mode, where the AI drafts each reply as a private note and a human decides whether to send it, a low-risk way to watch the quality before any customer saw it.
Once they were happy it was getting things right, they switched it to direct replies on both chat and email (notes-first is how we'd start most rollouts too). Not every Gorgias brand makes the same call, though. A premium accessories brand on the same helpdesk keeps email in Copilot mode and only lets the AI reply direct on chat. The agent now answers customers itself in Gorgias, in their own language, with a human stepping in only when the conversation needs one.
What was the biggest thing they did to improve their AI agent's resolution?
The single biggest lever was Self-Learning.
A thin help center is the usual ceiling on resolution in ecommerce: the answers just aren't written down yet. Self-Learning gets around that. It watches what happens when a ticket goes to a human, compares the AI's draft to the reply the human sent, and where there's a gap it drafts a brand-new help article to close it, ready for the team to approve.
Over the last 30 days, articles the AI drafted this way answered roughly 200 tickets it would otherwise have escalated. That's the part that compounds: instead of someone sitting down to write dozens of articles by hand, the AI spots the questions it can't yet answer and writes them up itself, getting better every week.
Pair that self-growing knowledge base with live Shopify order lookups, and the repetitive core of ecommerce support (order status, returns questions, product and care queries, sizing) becomes something the AI handles on its own at scale. That's what carried resolution up to 72%.
How does the brand customize their AI agent setup to work for their business?
A premium brand selling across several countries can't lean on a generic bot. They shaped the agent around two things in particular.
Setting communication, clarification and handover guidance
Guidance is where they set the rules the AI follows, and they use all three types. Communication and Style guidance keeps the tone warm and on-brand (the way you'd want a baby brand to sound with a new parent), and Context and Clarification guidance tells the AI when to ask for an order number before it answers, so it doesn't guess.
Handover and Escalation guidance defines the moments it should step back and pass the ticket to a person: returns disputes, damaged items, anything sensitive.
Replying in every customer's language
They serve customers in several European languages, and the AI picks up the language of each message and replies in kind (all 95 languages we support, detected automatically). There's no separate bot per country to maintain, since one agent covers the lot. So expanding into a new market becomes a matter of adding a support address or two; the agent already handles the language.
What impact is their AI customer service agent having now?
Here are the numbers we're seeing from the last 30 days:
- 72% AI resolution rate, so the AI fully handles about seven in ten tickets that come in.
- ~1,000 tickets a month, with around 720 of them resolved by the AI, no human needed.
- ~59 hours saved a month (at roughly five minutes of agent time per ticket, those 720 resolved tickets add up fast).
- 94% AI CSAT across the tickets the AI handled.
- ~200 tickets a month answered from help articles Self-Learning wrote on its own.
For a lean team covering several countries, that keeps reply times steady as the brand grows.
Where does the brand go from here?
The obvious next step is actions. They run their returns through a dedicated returns platform wired into Gorgias, and the plan is to connect it to the AI through Tasks and Tools so the agent can check a return's status (and, with the right approvals, start a return, issue a refund or update an address) instead of just explaining the policy. Returns and order changes are the next big chunk of repetitive volume to lift off the team.
Alongside that, they'll keep widening language and market coverage as they grow across Europe, and let Self-Learning keep nudging resolution past 72% as the knowledge base fills in.
What does "fully automated" actually look like for a brand like this? Here, it means the AI owns the repetitive 70-odd percent: order status, product and care questions, sizing, the routine returns queries.
The team keeps the judgment calls (the upset customer, the damaged delivery, the return that needs a person's discretion). That's the split we'd point any ecommerce brand toward. It's how the brand gets 72% resolution and a 94% satisfaction score at the same time.
If you'd like to see more customer stories like this one, take a look at how Edel Optics reached 79% AI resolution across multiple languages and how YouGarden saved 965 hours a month. And if you're weighing it up for your own store, our pricing and plans start with a 30-day trial, and the Gorgias ROI calculator gives you a first estimate of the hours your own ticket volume would hand back.
