Palona Needs Workflow Depth
Palona
The real moat here is not the voice layer, it is owning the restaurant job that happens after the caller speaks. Vapi and Retell make it cheaper to stand up a phone agent, but Palona sells into restaurants that need an agent to do specific work inside POS, catering, loyalty, menu, and store operations systems. If a cheaper stack can answer calls but cannot place the order correctly, route a catering lead, or reflect live store conditions, it stays a demo, not a durable replacement.
-
Horizontal providers are built as configurable voice toolkits. Vapi lets developers mix transcriber, model, and voice components, then connect APIs and build flows. Retell similarly focuses on prompts, knowledge bases, function calling, and telephony. That lowers build cost for a restaurant integrator, but it also means restaurant logic still has to be assembled on top.
-
The pricing umbrella is getting thinner. Vapi charges a $0.05 per minute platform fee on top of provider costs, while Retell publishes pay as you go pricing starting around $0.07 per minute and no platform fee. As those infrastructure layers get cheaper and more production ready, Palona has less room to win on raw voice novelty and more need to justify price through workflow outcomes.
-
Restaurant software incumbents are already bundling adjacent workflow pieces. Toast launched a drive thru product in 2026 that includes AI voice ordering integrations and speed of service tooling, and Palona is positioned around phone ordering, catering qualification, revenue intelligence, and operations monitoring. That makes integration breadth critical, because the vendor controlling the daily workflow can more easily absorb voice into the base contract.
The next phase favors restaurant AI vendors that look less like a bot builder and more like an operating layer for stores. If Palona keeps expanding from call handling into order execution, catering workflows, and real time store context across voice, text, and vision, it can defend premium pricing even as the underlying voice stack becomes cheaper and more interchangeable.