What are the four modes of service?
The short answer
Every service interaction now sits in one of four modes, depending on whether each side is a person or an AI. Mode 1 is human to human. Mode 2 is a customer talking to the brand's AI. Mode 3 is the customer's AI talking to a human agent. Mode 4 is the customer's AI talking to the brand's AI. Most CX investment is in Mode 2; most CX risk is in Modes 3 and 4.
Mode 1 · Human ↔ Human
The model we know: a person calls, a person answers. It is declining as a share of all interactions, which makes the human moments that remain more valuable. Done well, they are designed, coached against outcome, and carried across channels.
Mode 2 · Human ↔ Brand AI
The model most companies are building: chatbots, IVR, voice agents and self-service, with the customer still human. The design questions are whether your AI says it is an AI, states what it can help with, and hands off to a person with full context when it reaches the edge of its scope.
Mode 3 · Customer AI ↔ Human
The model almost nobody is designing for. The customer's own AI agent calls or messages your contact centre and a human agent handles it. It looks like normal traffic. It isn't. In field tests, customer-side AI has passed identity checks and worked issues through to resolution without the agent realising.
Mode 4 · Customer AI ↔ Brand AI
The model that changes everything. Two AI agents negotiate an outcome on behalf of their principals, with no human in the loop. It is the fastest, most efficient and most dangerous mode, because without shared structure every interaction is a negotiation from scratch, and policy gaps are exposed instantly.
Cost and liability
Modes 1 and 2 are a cost problem: they are where most of the budget sits, still priced in seats and minutes. Modes 3 and 4 are a liability problem. An operation that can't tell who it is talking to commits without authority, discloses to unverified parties, misses vulnerability, fails machine-speed verification, and can't defend the complaint that follows. None of it looks like a failure at the time.
Common questions
Which mode is my operation in?
Usually several at once. A contact centre with a chatbot is already running Modes 1 and 2, and is probably receiving Mode 3 traffic it can't see. The Dual CX Diagnostic scores your readiness for each mode separately.
Why are Modes 3 and 4 riskier than Modes 1 and 2?
Because the customer brings the AI. Processes designed for a human caller, such as identity verification, escalation and what an agent is allowed to promise, were never designed for an AI acting on someone's behalf.
Can I choose not to support Modes 3 and 4?
You don't decide whether customer AI reaches you; your customers do. What you can decide is whether those interactions are designed or left to chance.