Claude and ChatGPT both have their own marketplaces where you can connect systems/software/services to your AI assistant. Planday has a verified connector in the Claude marketplace, and plugin in the ChatGPT marketplace.

Once you connect your AI assistant to Planday you are able to:

The connector enables you to access Planday from ChatGPT or Claude, not from your Planday account. When you're in Claude or ChatGPT and you've connected with Planday, your AI assistant of choice can see your rota, your team, your shifts and your leave data. It will make recommendations, but won’t take action until you approve it.

Within this early access period, we are learning everyday how customers are using Claude/ChatGPT to take action in Planday.

From what we can see, customers are currently using Claude/ChatGPT to:

It works best for those responsible for the rota, typically shift managers, operations leads, and team supervisors. If you regularly check who's available, chase cover, approve swaps, or review rotas and you use Claude/ChatGPT today, this might be for you.

HR and operations managers with multi-site responsibility are finding it particularly useful as a fast way to compare and analyse across locations, something that currently takes a lot of manual work.

Connecting to Claude or ChatGPT is easy. You can find instructions on our Help Centre for both ChatGPT and Claude.

You don't need to learn prompt engineering. Plain language works: "Who's available to cover a Saturday morning shift?" is a perfectly good prompt.

If you are an experienced AI user, you'll quickly find ways to combine Planday data with other context: local events, the weather, business data that sits outside of Planday, to help you make scheduling decisions.

Early adopters span hospitality, care, retail and service businesses. The common thread isn't sector - it's complexity. Businesses managing multiple staff types, variable demand, or multiple sites tend to find the most immediate value. If your scheduling involves a lot of judgement calls, the connector is designed to help with exactly that.

You can ask Planday data questions and take scheduling actions, all in plain language:

Sometimes, no - particularly for simple one-step tasks you already know how to do quickly within Planday. Where it genuinely wins is on tasks that involve judgement, cross-referencing, or analysis. Checking who's available, checking if they've already done too many hours this week, cross-referencing a leave request, and drafting the cover message, in Planday those are four separate steps. In one conversation with Claude/ChatGPT, they're one prompt.

The bigger the question, the more value you get.

Regardless of what AI assistant plan you are on, Planday sends the same data to your AI assistant. However, upgraded AI plans generally provide smarter reasoning, better data analysis and faster performance. Please consult directly with your AI Assistant provider on what plan makes sense for you.

The connector only does what your Planday account already allows. A useful way to frame it for an approval process: it's not a new system with new data access, it's a new interface to the system you already have, with the same permissions you already have.

For internal approvers (IT, data leads, compliance), the key facts are:

As soon as Planday delivers your data, your AI provider (Anthropic for Claude or OpenAI for ChatGPT) takes over. From that point, their terms & privacy rules cover how your data is stored and handled.

It means you're using the connector while it's still in a monitored launch phase. The core functionality, connecting Planday to ChatGPT and Claude, making scheduling queries and taking actions, is live and working. During early access we're gathering feedback, measuring usage, and refining based on what real managers tell us.

We don't expect to remove access from early adopters. The purpose of early access is to improve the product, not to run a limited trial.

We are in an “early access” phase of our launch, this means connecting Planday to Claude and ChatGPT is currently available to all Planday customers, no matter what plan you are on. During early access we're gathering feedback, measuring usage, and refining based on what real managers tell us.

It depends on what type of plan you have with Claude/ChatGPT. Planday’s responsibility, under our Data Processing Agreement, does not extend to how the AI provider handles data once it has left us at your instruction. From then on, Anthropic (Claude) or OpenAI (ChatGPT) processes it under its own terms, including how long it’s kept and whether it’s used for training. Your employer should have a data processing agreement with that provider. If you want the specifics for your organisation, ask your IT team.

The connector can see scheduling and availability data: names, roles, shift patterns, availability, leave balances, working hours, and shift history. It cannot see personal contact details, dates of birth, payroll data, bank details, or any other sensitive personal records.

Claude or ChatGPT have no standing copy of the Planday workspace's data.

It only reads what a request needs: when a user asks a question, the AI assistant calls the specific tool for it (for example, next week's shifts for one department), and Planday returns just that answer, within that user's permissions.

Nothing gets pulled in bulk or synced in advance.

Planday’s responsibility ends when it sends the data to the AI assistant. From then on, Anthropic (Claude) or OpenAI (ChatGPT) processes it under its own terms, including how long it’s kept and whether it’s used for training. Your employer should have a data processing agreement with that provider. If you want the specifics for your organisation, ask your IT team.

Only what has already been sent to the AI assistant could be exposed. The connector doesn’t give the AI direct access to Planday’s database, and it only returns what the logged-in user can already see. So a breach of the AI provider wouldn’t reach Planday’s system directly. But any data in a conversation that was compromised could be exposed, the same as with any third-party service. We can’t tell you how secure either provider is, so check their security documentation or ask your organisation.

In addition, the Planday connection to Claude/ChatGPT is tied to a specific user account. If that user account is considered compromised, the user can be removed or reset authorisation to block any new data requests.

Planday's responsibility and boundaries: Planday fully protects your data as long as it remains within Planday's systems and is handled by Planday's approved sub-processors. Since third-party services like ChatGPT or Claude are not sub-processors to Planday, using the AI connector means you are choosing to send data outside Planday's controlled environment. As a result, Planday waives responsibility for what happens to the data within the AI tool.
​

Requirement for your own DPA with ChatGPT or Claude (Anthropic):
Because you are transferring data to a third party, you, as the data controller, must sign your own Data Processing Agreement (DPA) directly with OpenAI/Anthropic. This is offered for their Enterprise and API customers (and in these agreements, they also guarantee that they do not use your data to train their general AI models).

You can find more information in Planday’s DPA and AI connector legal notice.

In the Planday environment data residency and processing is in the EU/UK as accounted for in our data processing agreement.

Planday’s responsibility ends when it sends the data to the AI assistant. From then on, Anthropic or OpenAI processes it under its own terms, including how long it’s kept and whether it’s used for training.

You can find more information in Planday’s DPA and AI connector legal notice.

Use of an AI assistant (such as ChatGPT or Claude) to run your business may require you to update your own terms and policies associated with how you use and process data. It is best to seek the appropriate legal advice.

You can find more information in Planday’s DPA and AI connector legal notice.

No. The connector only exposes data through your existing Planday permissions. If your account setup restricts or anonymises certain fields, the connector will see exactly what your Planday account sees and no more. It does not bypass, override, or reconstruct data that your configuration has restricted.

The connector uses OAuth, the same secure standard used by Google, Microsoft, and most enterprise software integrations. It can only perform actions your Planday account is authorised to perform. If you're a scheduler who can't approve payroll, the connector can't approve payroll either. Your permissions are the ceiling, not a starting point.

Every action the connector takes is logged internally, who made the request, which tool was used and the outcome. Planday does not see the content of your conversation with Claude or ChatGPT. In the future, audit reporting in relation to the use of Claude/ChatGPT will be available in Planday, until then if you require documentation of this audit trail, you will need to request this directly from Planday.

Yes, the permissions that logged-in users can access are honored.

Admins do not have the ability to manage Claude/ChatGPT access from Planday. Important to note that Claude/ChatGPT can only see the data based on the permissions set in Planday for each user.

Since this connection sits outside Planday, we recommend setting your own guidance before connecting. If you don’t already have your own guidance in place for using AI tools, please check our AI connector legal notice to learn more about Planday AI connector.

No it only talks to backend APIs and then restricts that response to relevant data.

The exact data sent to any AI assistant connected is the same. What can change is how Claude or ChatGPT decides which of that data to present.