What is the Service Handshake?
The short answer
The Service Handshake is an open, vendor-independent standard that sets the terms whenever AI is part of a service conversation. Before the interaction proceeds, each party declares six things: its goals, options and constraints, data permissions, fallback rules, cost parameters and authority level. It is part of the Dual CX framework, published under CC BY 4.0, and the current version is SH-1.2.
The six declarations
Every party declares:
- Goals: what it is trying to achieve, in priority order, including explicit non-goals.
- Options & Constraints: what it can and cannot do, including financial authority and actions explicitly out of scope.
- Data Permissions: what data may be used, for which service purpose, how long it is kept and how it is verified.
- Fallback Rules: what triggers clarification or escalation, how far automation can go, and who is responsible.
- Cost Parameters: limits on interaction time, exchanges and computing resources, above which the agent must escalate.
- Authority Level: who the agent represents, who is affected, what it can commit to, and the evidence behind that claim.
The rules that matter most
Agents disclose that they are AI, and whom they represent, at first contact. Missing information never counts as permission. Confirmation is not approval: irreversible actions need independently checkable delegation or approval from the person represented. Data use is limited to the service purpose. Harm routing cannot be waived, and blocked actions and reasons carry through every human handover.
What it is, and what it isn't
The Service Handshake is deliberately narrow. It does not tell you how to build your agent. It tells you what your agent must be able to say, and hear, before it is safe to proceed. Six declaration elements, open JSON schemas, vendor-independent.
Adopting it does not by itself establish legal compliance. What it gives you is a reviewable record of declared purposes, authority claims and decisions.
How it relates to the EU AI Act
From August 2026, Article 50 of the EU AI Act requires AI systems to be designed so people know when they are dealing with AI. The Service Handshake adds an operational rule on top: every AI agent discloses that it is AI, and whom it represents, at first contact, including agent to agent.
Common questions
Is the Service Handshake free to use?
Yes. It is published under CC BY 4.0: free to use and free to build on. Credit it as "The Service Handshake, Dual CX. CC BY 4.0".
What if the other party has no declaration?
Assisted participation replaces automatic rejection. The service helps record the other party's terms, and missing information never counts as permission. Participation can start as a staff-run process with no JSON needed.
Does it handle someone using AI on a relative's behalf?
Yes. Its Authority Level element separates the person represented, the person who set up the agent and the delegation claimed. Configuring an agent is never treated as authority over someone else's account.
Who maintains it?
It is maintained on GitHub with schemas, examples and tests, and archived on Zenodo with a DOI (10.5281/zenodo.23181969). Contributions are welcome; the standard belongs to whoever uses it.