I built a customer support inbox with AI that catches problems before your team does
hey everyone, been lurking here for a while and figured it was time to actually share what i've been building
i run a digital marketing agency in manchester and work with a lot of ecommerce brands. kept seeing the same thing over and over - support teams closing hundreds of tickets a week but nobody ever connecting the dots between them
one agent refunds a customer on monday because the product "tastes different." another agent does the same on tuesday. by wednesday there's three more. each one looks like a one-off so nobody flags it. three weeks later the founder realises they've lost £24k in refunds and it all traces back to one bad supplier batch
the data was always there in the inbox. nobody had time to read for patterns
so i built SignalCX - it's a full customer support inbox (reply, tag, assign, refund/cancel orders from Shopify) but the AI runs in the background reading every email and automatically flags patterns. defective batches, checkout bugs, carrier failures, refund surges. you don't run scans or reports manually, it just tells you
the difference from gorgias/zendesk/freshdesk is the intelligence layer and the pricing. flat monthly rate, no per-ticket fees, no per-agent fees. free plan to start
stack is PHP, MySQL, Tailwind, Claude API for the AI. built the whole thing solo
where i'm at right now:
product is live with a free plan and live demo
65+ pages on the site (comparisons, integrations, industry pages)
shopify app store submission in progress
just launched an affiliate programme (80% first payment, 50% recurring)
zero users. literally day one
would love any feedback on the site or the positioning. also happy to answer questions about the build if anyone's interested in the technical side
