Connect your own ChatGPT or Claude to Coordinex.
An owner or admin can connect their own ChatGPT (OpenAI) or Claude (Anthropic) assistant to their Coordinex organization — then ask about the business in plain language and have the assistant draft routine actions for a person to review and approve. It reads a defined set of operations, it proposes but never acts on its own, and it never moves money.
The agent is set up inside the Coordinex iPhone app by an owner or admin. For the full data disclosure, read the AI Agent privacy page.
Followed a link to a specific item? Open the Coordinex app on your iPhone to view it — these reference links resolve inside Coordinex, where your live permissions apply.
What it is
Coordinex runs your scheduling, hours, team, approvals, and payouts. The AI Agent lets an owner or admin point their ownChatGPT or Claude at that operational picture, so the questions you'd normally dig through menus for — “who's close to overtime this week?”, “what's still unapproved?”, “draft a shift to cover Saturday morning” — get a plain-language answer or a ready-to-review draft. Nothing the assistant suggests takes effect until a person confirms it.
How to connect
Open it in the app
In Coordinex on your iPhone, an owner or admin opens the AI Agent connector and chooses ChatGPT or Claude. The agent is off until an owner turns it on — employees and supervisors can't connect it.
Sign in to your own AI
You authorize through a standard, secure OAuth window in your browser. You never paste an API key or password into Coordinex, and Coordinex never reads or stores your assistant's token. The connection carries only the read-and-propose permissions your role allows.
Ask. Review. Approve.
Ask about your business in plain language. The assistant reads what your permissions already allow and drafts any action into your in-app approval inbox. A person reviews, edits, and confirms before anything happens.
What the agent can read
Only a curated, purpose-built slice of Coordinex — through Coordinex's own secure connector, never your raw database, and only what your own permissions already allow, re-checked on every request:
- Schedule — upcoming shifts and the team schedule for a date range.
- Hours — hours worked, overtime proximity, and non-financial work summaries for named team members.
- Team — active members with names, roles, status, and availability for scheduling.
- Staffing & coverage — open shifts, coverage gaps, and staffing-risk summaries.
- Approvals & operations — pending approval counts, the work queue scoped to your permissions, operational alerts, and the owner/admin daily brief.
Every result that leaves Coordinex passes through an output-minimization boundary that strips sensitive identifier shapes — Social Security and EIN numbers, bank-account and routing numbers, credential-like strings — and earnings and pay figures for other people are withheld at launch.
What the agent can do — propose, never act alone
When you ask the agent to do something, it drafts a proposal — a draft announcement, a draft shift, a draft time-off request, a suggested approval — into an in-app approval inbox. A person reviews it, can edit it, and explicitly confirms or rejects it. Only after a human confirms does Coordinex carry out the action. This step is required for every action and cannot be skipped; anything irreversible takes an extra confirmation, and deletions are recoverable soft-deletes.
The agent never moves money.It has no payment, payout, payroll, earned-wage-access, or fund-transfer capability of any kind. When a proposed action happens to touch a pay-relevant figure, the amount is recomputed by Coordinex's own servers at the moment of approval — the assistant's number is never trusted, and the confirmation step accepts only an organization and an action identifier, never a dollar amount.
Where the AI runs — bring your own
The intelligence is your own ChatGPT or Claude subscription, operated and billed by OpenAI or Anthropic. Coordinex does not run, host, or pay for the AI model, and does not send your customer data to any AI provider to train a model. Coordinex's role is narrow: it verifies who you are and returns the specific, permission-scoped, minimized answers above to your assistant. The in-app disclosure states it plainly before any connection is made:
“Your ChatGPT or Claude subscription is billed and operated by OpenAI or Anthropic. On personal plans, your prompts to them may be used to improve their models, which Coordinex cannot control. Coordinex only verifies who you are and never shares another employee's private data.”
Security and control
- Owner/admin only, enforced server-side by re-deriving your role from a verified sign-in on every request.
- Secure OAuth, no key pasting.Tokens are short-lived and revocable; Coordinex never stores your assistant's access token.
- Tenant-isolated. A connection can only ever see the one organization it was authorized for.
- Off by default. The integration is disabled until an owner turns it on, and an owner can turn it off for the whole organization at any time.
Frequently asked
01Who can use the AI agent?+
Owners and admins only. The restriction is enforced on Coordinex's servers by re-deriving your role from a verified sign-in on every request — not from anything the assistant or a token claims. Employees and supervisors cannot connect or use it.
02Can it run payroll, send a payout, or move money?+
No. The agent has no payment, payout, payroll, earned-wage-access, or fund-transfer capability of any kind. This is a hard architectural boundary, not a setting. When a proposed action touches a pay-relevant figure, Coordinex's own servers recompute the amount at the moment of approval — the assistant's number is never trusted to move money.
03Does Coordinex use my data to train an AI model?+
No. Coordinex does not send your customer data to any AI provider to train a model. The intelligence is your own ChatGPT or Claude subscription, billed and operated by OpenAI or Anthropic. Because it's your subscription, their terms govern your prompts — on some personal plans a provider may use prompts to improve their models, which is a property of your AI account, not of Coordinex.
04What can the assistant actually see?+
A curated, permission-scoped set of operational data — your schedule, hours, team, staffing and coverage, and approval queue — read through Coordinex's own connector, never your raw database, and re-checked against your live permissions on every request. Sensitive identifiers like Social Security and bank-account numbers are stripped before the assistant ever sees the text, and earnings and pay figures for other people are withheld at launch.
05Can it act without me?+
No. The agent only drafts proposals into an in-app approval inbox. A human reviews, can edit, and explicitly confirms or rejects every action — this step is required and cannot be skipped. Anything irreversible takes an extra confirmation, and any deletion is a recoverable soft-delete.
06How do I turn it off?+
Disconnect ChatGPT or Claude inside the Coordinex app at any time, which revokes the assistant's access immediately. An owner can also disable the agent for the entire organization, and you can remove Coordinex's authorization from your ChatGPT or Claude account settings.
Your business, in plain language. Your AI proposes — you approve. Coordinex keeps the money boundary.