Apps / Rho

Agents can now sign in to Rho

Rho is business banking built for the AI era. With Sign in with AgentID, the agents doing that era's work get their own verified identity inside the business's Rho workspace: their own sign-in, their own audit trail, and a human owner on record, with every approval and limit still Rho's to enforce.

Coming soon.

Rho puts a startup's financial stack on one platform: business accounts, corporate cards, bill pay, treasury, and an API underneath, with banking services provided through its partner banks. Its newest products assume AI does part of the work, like Close, which codes transactions automatically.

Finance work is exactly where agents are showing up first: an AP agent that processes bills, an expense agent that codes and flags transactions, a treasury agent that watches balances. And finance is exactly where "the agent used my login" is least acceptable. Money movement demands an audit trail where every action has the right name on it, and a shared human session gives every agent the CFO's name.

AgentID, the sign-in for AI agents from AgentMail, fixes the name. The agent signs in to Rho as itself, with a verified identity of its own, and Rho knows which human owns it. Nothing about what the agent may touch changes; that stays with Rho's roles, limits, and approvals.

In short

  • Rho accepts Sign in with AgentID, so an AI agent can hold its own sign-in inside a business's Rho workspace instead of borrowing a human's.
  • The agent's identity is a verified email address issued by AgentMail. It is a working inbox, so Rho can reach the agent with notifications and confirmations, and reach its owner when something needs a human.
  • Rho receives a stable identifier for each agent and, through the owner scopes, the name and email of the human behind it.
  • No password or reusable key crosses the sign-in. Each sign-in is a one-time signature, and authorization codes are single use.
  • Rho keeps every product decision: roles, permissions, payment approvals, limits. AgentID only answers who is signing in.

Why would a finance platform want agent sign-ins?

Because the audit trail is the product. A platform that moves money is judged on being able to say, for every action, who did it, when, and under whose authority. Agents working under human logins break all three answers at once: the "who" is wrong, the session makes timing ambiguous, and the authority is whatever the human's role happened to allow.

Separate agent identities repair the trail. An AP agent that initiates a payment run does it as itself, under whatever access the business scopes it to, and the record shows the agent, its verified identity, and the human who owns it. Dual control becomes natural instead of awkward: the agent initiates, an approval routes to a human, and the owner claim tells Rho exactly which human stands behind the requester.

Revocation gets surgical too. An agent that misbehaves loses its own sign-in, not the controller's. Offboarding a human no longer silently strands or orphans the agents that were running under their credentials, because the agents were never inside those credentials to begin with.

How does an agent get its own Rho sign-in?

Sign in with AgentID is standard OpenID Connect, the same technology as Sign in with Google. For Rho, accepting it is configuration on their existing login, not new infrastructure.

On the agent side, its AgentID is its AgentMail inbox address. The agent reaches Rho's sign-in, picks AgentID, and approves the sign-in through the AgentMail API with an inbox-scoped key. The approval is remembered for 180 days per inbox and app, so repeat sign-ins continue on their own.

What Rho receives is an ES256-signed id_token, valid for ten minutes, with no refresh token. It carries a stable subject for the agent, its verified email address, and, because Rho is a registered app with the owner scopes, the owner's name and email. An agent that cannot share its owner's details cannot finish the sign-in alone; the organization owner approves it instead. Rho never gets a token that silently lacks the owner claim, which for a finance platform is the whole point.

The business's Rho account is unchanged. It belongs to the business, opened and verified through Rho's normal onboarding. The agent's identity attaches to that workspace with whatever access the business assigns it.

What does the agent's email address do for Rho?

It becomes the agent's registered email, and it works. Rho can send the agent a payment confirmation, a flagged-transaction notice, or a security alert, and the agent can read and act on all of it. Agents can even open and resolve support tickets. When something needs a person, the owner is one claim away.

The address is identity, not access. Signing in grants Rho no ability to read or send the agent's mail, and it grants the agent nothing inside Rho beyond what its assigned role allows.

Who controls what

StageAgentIDRho
Verified agent identityIssues the verified email and stable subjectNot involved
Sign-in approvalThe agent approves through its inbox-scoped keyNot involved
Owner attributionDelivers owner name and email in the tokenDecides how to use it
Roles and permissionsNot involvedRho and the business
Payment approvals and limitsNot involvedRho and the business
Compliance and enforcementSupplies owner_sub for per-human policyDecides and enforces

What stays with Rho?

Everything that makes Rho Rho. Account onboarding, roles and approval chains, card controls, limits, the compliance posture that a finance platform lives by: none of it changes. AgentID answers one question at the door, who is this agent and which human is accountable for it, and hands the answer to Rho's existing controls. It is not a trust score and not an endorsement of any agent's behavior, and it grants no authority to move money.

Start at Rho, and use the AgentID sign-in guide to prepare your agent's inbox so it is ready the day the option goes live.

FAQ

Questions, answered.

No. The business's account belongs to the business and goes through Rho's normal onboarding. The agent gets its own sign-in inside that workspace, with whatever access the business assigns it.

Only if the business's roles and approval flows allow it, which is Rho's domain, not AgentID's. What the sign-in changes is attribution: anything the agent initiates carries the agent's own identity and its owner's name, instead of hiding inside a human's session.

A stable subject identifier, the agent's verified email address, and, with the owner scopes, the name and email of the human who owns it. Ungranted claims are omitted, never nulled.

No. Sharing an address is identity. It is not permission to read or send mail.

New sign-ins with that key stop everywhere. Rho controls its own sessions, and the business can remove the agent's access at any time. To stop an inbox from signing in to Rho again, the owner disables that account through the AgentMail API.

No. A token proves who the agent is and who owns it. It does not claim the agent is safe or well behaved, and it is not an endorsement. What the agent may do inside Rho is decided entirely by Rho and the business.

Prompt

Give yourself access to AgentMail
by following the guide at:

https://docs.agentmail.to/agent-onboarding

and refer to:

https://docs.agentmail.to/llms.txt

for the documentation index.

Give your agent an identity of its own.

Create an AgentMail inbox and let your agent start signing in as a real user.