Ecommerce Support Is Dominated by a Few Questions
Pull a month of ecommerce support tickets and the distribution is remarkably consistent:
- Where is my order
- I want to return or exchange this
- Has my refund been processed
- Will this fit, and how does the sizing run
- Is this in stock, and when will it be back
- Can I change or cancel my order
Two of those categories, order status and returns, typically account for a very large share of total volume. They are also the most mechanical: the answer exists in your systems and requires no judgement.
That is the automation case. Not replacing support, but removing the queries that consume the team without using their skill.
Order Status Is the Obvious Starting Point
"Where is my order" is high volume, zero judgement, and the data already exists in your order management and carrier systems.
A well-configured agent verifies the customer, looks up the order, and gives the current status with the carrier tracking detail and a realistic delivery estimate.
The requirements that matter:
- Reads live order and fulfilment status, not a nightly sync
- Verifies identity before disclosing order details
- Pulls real carrier tracking rather than a generic status
- Handles split shipments and partial fulfilment correctly
- Escalates when the order is genuinely stuck, lost, or delayed beyond the promise
- Never invents a delivery date
That last point matters commercially. A confidently wrong delivery estimate generates a second, angrier contact and sometimes a chargeback.
Returns Should Be Made Easy
There is a temptation to make returns friction-heavy. Resist it.
Returns friction reliably produces more support contacts, worse reviews, and lower repeat purchase rates. The customer returning an item is not lost revenue; they are a customer deciding whether to buy from you again.
A good automated returns flow verifies the order, checks eligibility against your policy and the return window, explains the options clearly, generates the label or instructions, and confirms by email with the refund timeline.
Consumer protection law in many jurisdictions sets minimum return rights, and they override your stated policy. Make sure the configured rules reflect the law where you sell, not just your preferred terms.
Pre-Sales Questions Are Revenue, Not Support
This is where most ecommerce brands under-invest.
A customer asking "will this fit" or "is this compatible" is trying to buy. Every hour that question sits unanswered is conversion decaying.
An agent that answers sizing, materials, compatibility, dimensions, care instructions, and stock timing from your live catalogue converts browsers who would otherwise abandon.
Configure it from your actual product data: specifications, sizing guides with real measurements, stock levels and restock timing, compatibility, shipping times by destination. Keep it honest — overstating fit produces a return and a refund, which costs more than the lost sale would have.
Where Humans Must Stay
Route these to people:
- Damaged, faulty, or wrong items received
- Anything involving a safety issue with a product
- Payment disputes and chargebacks
- Repeat contacts about the same unresolved problem
- Customers who are clearly upset
- Requests for exceptions to policy
- Suspected fraud
A useful rule: if the resolution requires goodwill, judgement, or spending money outside policy, a person decides. Automation handles the cases with a correct answer.
Second-contact escalation is worth building explicitly. A customer contacting about the same order twice should go straight to a human with the full history attached. Nothing damages a support reputation faster than being re-triaged by a bot on the second attempt.
Voice and Chat Serve Different Moments
Most ecommerce contact is chat and email, but voice matters for high-value orders, older demographics, urgent delivery problems, and complex returns.
The same knowledge base should drive both, so a customer gets the same answer in both channels. Nothing erodes trust like chat and phone disagreeing about your return window.
What to Measure
| Metric | Why it matters | | Contacts resolved without an agent | The efficiency number | | First contact resolution | Whether resolutions actually hold | | Second-contact rate on the same order | The honest quality signal | | Pre-sales questions answered and conversion after | Revenue, not cost | | Return processing time | Customer experience and repeat rate | | Escalation rate and reason | Where the knowledge base is thin | | CSAT split by automated and human handling | Whether automation is hurting satisfaction |
Watch second-contact rate above all. A high automation rate with rising repeat contacts means you are deflecting rather than resolving, which is worse than not automating.
Rollout Order
Start with order status. Highest volume, lowest risk, clearest data.
Add pre-sales questions next, because that side is revenue. Add returns once your policy rules and legal minimums are correctly configured. Keep faults, damage, and disputes with humans throughout.
Final Recommendation
For ecommerce, automate order status first, treat pre-sales questions as a conversion opportunity rather than a support cost, and make returns genuinely easy.
Keep faults, disputes, and upset customers with people, escalate any second contact about the same order straight to a human, and track second-contact rate as your real quality measure.