AI Services Market › Use cases › Sales, Marketing & Support
Sales, Marketing & SupportDone with custom-built or unnamed agent 2026-10-27
Handle customer support with an AI agent (partial)
A merchant ran Gorgias AI Agent for months but disabled it after irrelevant answers and promises the store couldn't keep.
- What They Did
- The Gorgias team helped the merchant (Topped Toys) set up the AI Agent, which took several weeks. The agent followed logic Gorgias wrote, including editing and cancelling orders. The merchant says it worked for a few months. After the problems began, the merchant tried several support routes, then disabled the agent.
- How It Went
- The merchant says the agent first solved about 95% of issues, then failed in every case. It recommended products the store doesn't carry, couldn't edit or cancel orders, and made promises the store couldn't keep.
- Worth Knowing
- Support was the weak point. L2 replies took over 36 hours, and the team that built the agent offered no support, so ask about post-launch support before relying on it.
Try it yourself with custom-built or unnamed agent
Here are [number] recent customer support conversations where an AI agent answered for my store [store name]: [paste conversations]. Review each answer and flag any that were irrelevant or promised something we can't deliver, such as [e.g. refunds, delivery dates]. Then write a short list of rules and policy limits I should give the agent, and a list of topics that should always go to a human.
Do this with a skill.
These AI Services Market playbooks do this exact job.