Right to Disconnect and AI After Hours

Since 26 August 2025, every Australian employer, including small businesses with fewer than 15 people, has operated under the right to disconnect. Larger employers have lived with it since 26 August 2024. For most businesses the law did not force an immediate change, which is exactly why it is easy to underestimate. The right to disconnect quietly reshapes a habit that many Australian workplaces built over a decade: the assumption that a text, a call, or an email after hours will simply be answered.

This guide is written for owners, people leaders, and operations managers at Australian businesses who need to understand what the right to disconnect actually requires, and where AI can help a business keep responding to customers and coordinating work without leaning on staff to be contactable around the clock. The two subjects belong together, because the businesses most exposed to the right to disconnect are often the ones that had informally solved after-hours coverage by expecting employees to pick up.

What the right to disconnect actually says

The right to disconnect is set out in the Fair Work Act 2009 at section 333M. In plain terms, it gives an employee the right to refuse to monitor, read, or respond to contact, or attempted contact, from their employer outside their working hours, unless that refusal is unreasonable. The same right applies to contact from a third party where the contact relates to the employee's work, under section 333M(2). It was introduced by the Fair Work Legislation Amendment (Closing Loopholes No. 2) Act 2024 and is administered by the Fair Work Commission.

Two features of the law matter most for planning. First, it is a right to refuse, not a ban on contact. An employer can still send a message after hours. What changed is that the employee is entitled to leave it until their next shift, and cannot be punished for doing so, unless refusing would be unreasonable in the circumstances. Second, whether a refusal is unreasonable depends on a set of factors the Act lists, including the reason for the contact, how it is made and how disruptive it is, whether the employee is compensated for being available, the employee's role and level of responsibility, and their personal circumstances.

Disputes follow a defined path. Under section 333N, the parties must first try to resolve the matter at the workplace level. If that fails, either party can take it to the Fair Work Commission, which can deal with the dispute and, if needed, make orders, including an order that an employer stop certain conduct. The Fair Work Ombudsman publishes plain-language guidance for employers and employees on how the right works in practice.

Nothing in this article is legal advice, and the reasonableness test is fact-specific. Where a situation is genuinely unclear, get advice from an employment law specialist. What follows is about the operational reality the law creates, and how to meet it without asking your people to absorb the gap.

The operational gap the law exposes

For years, plenty of Australian businesses handled after-hours demand informally. A customer calls at 7pm, the owner or a senior staff member picks up. An urgent supplier email lands at 9pm, someone deals with it. A shift falls through and a manager works the phones to fill it. None of that was ever a real system. It was goodwill, and often unpaid.

The right to disconnect removes the assumption that goodwill will always be there. An employee is now entitled to switch off, and the business that quietly depended on them not switching off has a coverage problem it never properly named. It is a prompt to build the coverage a business actually needs as a deliberate design choice, rather than a favour extracted from staff.

There are three honest options for genuine after-hours demand. Pay for availability properly, through on-call arrangements and allowances that make the reasonableness test work in your favour. Decide that some contact can simply wait until the next business day, and set customer and internal expectations accordingly. Or put a system in place that handles the routine load without a person needing to be on their phone at all. Most businesses end up with a mix, and this is where AI has a legitimate, specific role.

Where AI helps you respect the law, not evade it

AI is not a way to pressure employees or to route around a right the Parliament deliberately created. It is a way to reduce the volume of after-hours contact that lands on a human at all, so that switching off becomes realistic rather than a source of conflict. Used well, AI lowers the temperature the law is responding to.

Three uses are genuinely useful here.

Answering routine calls and enquiries. A large share of after-hours contact is routine: a booking, a question, a status check. An AI phone or chat layer can capture these, answer the simple ones, and hold the rest for the next business day. That means an employee is not the fallback for a 9pm booking enquiry. Our guide on how to answer business calls 24/7 without night staff covers the mechanics for businesses that genuinely cannot let routine calls go unanswered.

Triaging what is genuinely urgent. The reasonableness test hinges partly on whether contact concerns a real emergency. An AI triage layer can separate a true safety or service emergency, which may justify contacting an on-call person, from a routine matter that can wait. That protects staff from being pulled out of their evening for things that were never urgent, and it creates a record of why an emergency contact was made.

Preparing a clean morning handover. Instead of staff monitoring inboxes overnight so they are not overwhelmed in the morning, an AI email assistant that drafts and sorts in your brand voice can organise overnight enquiries so the team starts the day with a prioritised, summarised queue. The work waits, but it waits in order, which removes the pressure that pushed people to peek at their phones after hours in the first place.

These outcomes are directional rather than a guaranteed metric, because the right mix depends entirely on your industry, your customers, and how much genuine after-hours demand you carry. The principle holds regardless: the less routine load that reaches a human after hours, the easier it is to respect the right to disconnect without losing responsiveness.

What a compliant, humane policy looks like

A business that takes the right to disconnect seriously needs more than a tool. It needs a written position that managers and staff understand, so that the reasonableness test is not being argued from scratch every time. The policy does not need to be long. It needs to be clear.

A workable policy covers a few things. It states the default expectation that employees are not required to monitor or respond outside working hours. It defines, as concretely as your business allows, what counts as a genuine emergency. It explains how on-call arrangements and any allowances work, so availability that is expected is also compensated. And it sets out how customer-facing coverage is handled after hours, whether that is a next-day commitment, a paid roster, or an AI layer that catches routine contact.

Managers also need training, because the most common breach is not a formal policy failure. It is a well-meaning manager who texts a team member at night out of habit and treats a delayed reply as a problem. The right to disconnect protects the employee who waits until morning. A manager who penalises that, even subtly, is where disputes start. This is a change-management task as much as a compliance one, and our guidance on managing AI and workplace change without losing your people applies directly to rolling out any new after-hours expectation.

Common situations and how to handle them

The reasonableness test lands differently across industries, so it helps to think through the patterns Australian businesses actually face.

Consider a professional services firm where a client emails a request at 8pm expecting a same-evening reply. The firm cannot compel a staff member to answer, and treating a next-morning response as a failure would put the firm on the wrong side of the law. The better design is to set a clear client service standard for after-hours enquiries, back it with an automated acknowledgement that sets expectations, and let the substantive work begin at the start of the next business day. The client gets certainty, the employee gets their evening, and the firm carries no hidden reliance on someone choosing to log back in.

Consider a business that genuinely runs on urgent after-hours demand, such as a services operator where a real fault cannot wait until morning. Here the honest answer is a paid on-call arrangement, structured so that the people expected to be available are compensated for that availability. Compensation is one of the factors that shifts the reasonableness assessment, and a documented on-call roster is far more defensible than an informal expectation that a particular person will always answer. An AI triage layer sits neatly alongside this by making sure the on-call human is only pulled in for the genuine emergencies, not the routine calls that could have waited.

Consider the internal case, where the after-hours contact is between colleagues rather than from a customer. A manager who messages a direct report at night to check a detail may believe it is harmless, but a pattern of it erodes the boundary the law protects. The fix is cultural as much as technical: agree that non-urgent internal questions wait, use shared systems so answers can be found without pinging a person, and reserve after-hours contact for matters that genuinely cannot wait. A well-configured internal knowledge assistant reduces the reflex to message a colleague simply because information is hard to find.

The through-line across all three is that the right to disconnect rewards businesses that name their after-hours needs honestly and meet them by design. Vague expectations create disputes. Clear standards, paid availability where it is genuinely required, and automation for the routine load keep the business responsive without asking staff to carry the gap.

Fit this into your wider workplace obligations

The right to disconnect does not sit alone. It joins a set of Australian workplace obligations that businesses already manage, from award and pay compliance to broader Fair Work compliance that AI can help automate. Where AI touches employee communications or monitoring, it also intersects with privacy. If you introduce a system that records or analyses after-hours contact, do it within the Australian Privacy Principles and Privacy Act obligations, and be transparent with staff about what is captured and why. An AI layer meant to protect people's evenings should never quietly become a surveillance layer over them.

For businesses building broader internal automation, the same governance discipline applies as anywhere else. An AI HR assistant handling policy questions, leave and onboarding can help staff self-serve during business hours, which reduces the after-hours pinging that happens when people cannot find an answer. As always, the deployment should follow the responsible AI governance pattern of clear data controls and human oversight.

The real opportunity

It is tempting to read the right to disconnect as a constraint, one more rule that makes running a business harder. The more useful reading is that it forces a question worth answering anyway: does your business have a real system for after-hours demand, or has it been quietly running on the goodwill of people who could not switch off?

Building that system deliberately, with paid coverage where availability is genuinely needed, next-day handling where contact can wait, and an AI layer for the routine load that never needed a human in the first place, is better for customers and better for staff. It removes the hidden reliance the law has now named. The businesses that treat 26 August 2025 as a prompt to design their coverage properly will end up more resilient than the ones that keep hoping someone picks up.

Related reading

• Answer business calls 24/7 without night staff
• The AI email agent that drafts client replies in your voice
• Fair Work compliance automation with AI
• Managing AI and workplace change without losing your people
• An AI HR assistant for policy, leave and onboarding
• Deploying AI agents responsibly: governance and human override