Is AI going to replace customer support?
No. What the Klarna walk-back, Gartner's 2027 prediction, and 86% of AI conversations needing a human tell us about where AI actually fits in customer support.
The AI-in-CX category is still being drawn. Deflection, assist, automation, copilot, agent. These words mean different things to different vendors, and the map of the category is contested. This pillar publishes our reading of the map, and where Handsom sits on it.
No. What the Klarna walk-back, Gartner's 2027 prediction, and 86% of AI conversations needing a human tell us about where AI actually fits in customer support.
Deflection rate measures whether AI touched the ticket. Resolution rate measures whether the customer got an answer. Vendors quote the first; only the second describes customer experience.
Agent assist works inside the helpdesk for your team. Deflection AI works in chat for your customers. They solve different problems, and the brands getting AI in CX right are running both.
Yes, but which job? Three different products live under the label. They solve different problems, and the cheapest is rarely the right one for the tickets that matter most.
Three things drive ongoing cost: knowledge-base drift, model quality drift, and conversation QA. Deflection AI needs all three. Team-side AI needs almost none.
Depends which kind. Deflection AI takes 3-6 months for most mid-market brands. Team-side AI takes hours. The difference is structural, not effort.
Realistic monthly spend at $5M, $20M, and $50M DTC brackets. Deflection AI versus team-side AI versus running both. List price plus the costs that don't show up on the invoice.
Vendor list prices start near $1 per resolution. A human ticket runs $8 to $12. The list-price gap is the pitch. The total-cost picture is more complicated than that.
The right deflection KPI for an ecommerce chatbot is per ticket category, not blended. Vendors quote 90%. The data quotes 8% on refunds and 60-70% on order status.
Yes, well sort of. Bad implementations hammer brand reputation, good implementations can be brand positive. What does this mean for DTC brands?