I've spent the last 13 years building and optimising Shopify stores. In that time, I've watched one problem stay stubbornly unsolved no matter what tools merchants throw at it: customer support.
First it was email inboxes. Then live chat. Then helpdesk software. Now AI.
And yet, in the Shopify communities I'm part of — the forums, the Facebook groups, the Reddit threads — the conversation sounds almost identical to what it sounded like five years ago. Customers are frustrated. Merchants are overwhelmed. And somewhere in the middle, a support tool is generating responses that technically answer the question but somehow make everything worse.
I want to talk about why that keeps happening. And I want to hear from merchants directly — because I think the people building these tools are solving the wrong problem.
The promise was simple. The reality isn't.
The pitch for AI support tools goes like this: your customers ask repetitive questions — where's my order, what's your return policy, do you have this in blue — and AI handles those automatically, freeing your team for the complex stuff.
That's a real problem and a reasonable solution. For a specific slice of support volume, it works.
But here's what nobody talks about in the product demos: the tickets that actually destroy customer relationships aren't the repetitive ones. They're the ones that happen when something goes wrong.
A customer receives a damaged product. An order gets lost in transit. The wrong item shows up. Someone misses a birthday gift because of a shipping delay.
These are the moments that define whether a customer comes back or writes a review you'll be managing for the next two years. And they're exactly the moments where AI support, in its current form, falls completely flat.
"I've flagged this for our team to review."
That sentence. I've seen it in screenshots more times than I can count.
A customer reaches out in genuine distress — their daughter's birthday present didn't arrive, the package showed up crushed, they got charged twice for something they ordered once — and the AI responds with some version of:
"I understand your frustration. I've flagged this for our team to review and someone will be in touch within 24–48 hours."
That's not support. That's a delay wrapped in empathy language.
The customer doesn't want acknowledgement. They want the problem fixed. And the gap between those two things is where trust — and repeat business — gets permanently destroyed.
What makes this particularly painful is that most of these situations have a completely clear resolution. The answer is almost always: approve the refund, send a replacement, issue store credit, or generate a return label. These aren't complex decisions requiring senior management review. A merchant who looked at the ticket for 30 seconds would know exactly what to do.
But the AI can't do any of that. It can only talk.
There's a deeper issue with how we're measuring success.
Most support tools show you metrics like response time and ticket volume. Respond faster, close more tickets — that's the dashboard. That's what the product demo celebrates.
What they don't show you is what the customer did after the ticket was "resolved." Did they come back? Did they leave a review? Did they tell three friends about what happened?
A ticket "resolved" in 2 minutes by an AI that gave the customer a policy link and a case number is meaningfully worse than a ticket resolved in 4 hours by someone who actually fixed the problem. But the dashboard only sees the 2-minute close time and calls it a win.
CSAT scores tell part of the story, but most merchants I talk to don't have CSAT tracking set up at all. They find out a customer was unhappy when they see the 1-star review go live.
What are merchants actually dealing with right now?
I'm writing this because I'm genuinely curious about the gap between what AI support tools promise and what merchants actually experience day-to-day.
Specifically, I want to understand:
When AI support fails your customers — and I'm betting most of you have at least one example — what kind of situation is it? Is it the emotionally charged tickets, like damaged goods and lost packages? Is it situations where the customer needed something done rather than something explained? Is it the returning customer who has to re-explain their entire history every single time they contact you?
And on the other side: what would "actually useful" AI support look like to you? If you could describe what it would need to be able to do before you'd trust it to handle customer conversations without you watching over its shoulder — what is that thing?
Your experience matters more than the product demo.
I'm not asking because I'm trying to sell you anything. I'm asking because after 13 years of watching the same problem not get solved by successive generations of tools, I think the people building these products are optimising for the wrong outcomes — and the merchants living with the results are the only ones who actually know what the right outcomes look like.
The good, the bad, the "I had to personally apologise to a customer because our bot told them something completely wrong." All of it is useful. All of it matters.
Drop your experience in the comments below or reach out directly. Every story from the field is worth more than any product roadmap built in a conference room.
Shopify developer and CX systems specialist with 13+ years of experience across 50+ Shopify and Shopify Plus stores. Zendesk Certified Administrator. HubSpot Certified. Currently Marketing Automation Specialist at NZTE.
I write about ecommerce operations, customer experience, and the real-world gap between what support tools promise and what merchants actually live with.
Is your Shopify support stack making or losing you money?
I engineer Shopify Plus stores and CX systems that actually resolve tickets — not just respond to them. Zendesk, Gorgias, AI chatbots, and the full stack. Book a free 30-minute discovery call and I'll tell you honestly what's broken and what it would take to fix it.