How does AI customer service change demand for support itself?

Cheaper, easier, faster customer service tends to attract more contacts, not fewer, because the friction that used to stop customers from asking goes away. AI lowers that friction sharply. So the volume saved on the contacts AI handles can be offset, in part or in full, by new contacts the system now invites. The cost case has to plan for that, not assume volume holds flat.

A retailer deploys an AI chat agent. The bot is good. Containment is high. The cost per contact in chat drops. Six months later, total support volume is up. The bot is fielding contacts that, in the old, slower, longer-queue world, the customer would have given up on. Demand has elastically expanded into the new lower friction.

What people in the field are saying

DCX Newsletter has a piece naming this directly: "AI customer support demand elasticity". The volume paradox is also a recurring theme in the trade press under "the AI volume paradox" framing.

What does demand elasticity look like in support?

Customers who would have lived with a small problem because contacting support was annoying now contact support, because the AI makes it easy. The contacts that used to be lost to give-up rate now appear in the queue. From the customer's perspective this is better service. From the support cost line, it is more contacts than the case assumed.

Why does the cost case usually miss this?

The case extrapolates from current volume. Current volume is the volume that survived the old friction. Lower the friction and the underlying demand reveals itself. The case is reading the visible part of the iceberg. The AI deployment exposes more of it.

What changes when you plan for it?

The math gets more honest. The deflection projection includes a "new contacts invited" term, not just a "contacts AI handles" term. The cost-per-contact target is lower (because there are more of them) and the team size is set for the new total, not the old residual. The CFO conversation is harder up front but more credible.

How do you measure it after launch?

Track total contacts across all channels, not only the AI channel. Track unique customers contacting per month, against a pre-AI baseline. Track the contacts the AI handles that look like they would not have been raised before (short, low-stakes, repeat customers). The increase in total contacts is the demand elasticity you absorbed.