Question 1

we tried a customer-facing chatbot once and it went badly — how is this different?

Accepted Answer

we don't build customer-facing chatbots by default. the work points inward: triage, retrieval, draft-and-review. agents stay on the phone or in the ticket; the ai sits behind them. customers keep talking to humans.

Question 2

how is this different from zendesk ai, intercom fin, or the gorgias suggestions we already pay for?

Accepted Answer

those tools work inside their own walls — your zendesk macros, intercom articles, gorgias replies. we build retrieval across the rest: 4 years of historical tickets, the product docs, the shipping policy pdfs, the engineering runbooks. with the mid-market e-comm cx team that cut first-response time on complex tickets by 40%.

Question 3

how do you handle the agents who think this is the first step to replacing them?

Accepted Answer

we say so on day one. the framing is 'helps agents, not replaces them' — and we measure it. the kpis we move are first-response time and resolution quality on complex tickets, not headcount. training sessions with the support team are part of the build, not an afterthought.