Andrea Chan on balancing AI and human judgement in workplace mental health

When employees go quiet or appear “just fine and happy”, it does not always mean they are not struggling. Workplace engagement scores and counselling feedback can paint a positive picture on the surface, yet mental health remains complex and difficult to decode.

As AI reshapes how organisations work and measure productivity, it is also entering conversations around workplace mental health. But can technology genuinely help employers recognise when people need support without reducing their experiences to data points?

In this timely conversation ahead of World Mental Health Day 2026, Andrea Chan, BTL Rep and Group Head of TOUCH Counselling & Psychological Services, explores where technology can complement human judgement, why psychological safety remains critical, and how organisations can use AI to support, rather than surveil, employee wellbeing.

Read the detailed insights from the interview below:

I recall supporting an organisation which contacted TOUCH Counselling and Psychological Services (TCPS) after the loss of one of their long-serving employee to suicide.

Many colleagues were very shocked and ladened with guilt, wondering whether they had missed the signs that she was struggling. They know that the particular colleague had been skipping lunch with everyone but no one asked how she was and she did not share how she was feeling with anyone either. 

The experience underscored the importance of psychological safety in the workplace, where people feel safe not only to speak openly about their challenges, but also check in on one another with care and concern. One of the biggest barriers to seeking help at work is still the fear of being perceived differently. 

Employees may worry that admitting they are struggling will cause colleagues or managers to see them as less capable, less resilient or less ready for greater responsibility. Meanwhile, others may tell themselves that they should simply cope on their own, especially if everyone around them appears to be managing well.

This is where the gap between mental health awareness and psychological safety often becomes evident. An organisation can have mental health benefits, workshops and resources in place, but employees may still hesitate to use them if they are unsure how speaking up will affect the way they are perceived in the workplace. 

Building psychological safety requires consistent reinforcement that people are supported, and it is often demonstrated in everyday moments, such as how managers respond when someone says they are struggling, asks for help or needs time to recover.

Over time, people are more willing to speak openly when they see that asking for support is met with empathy, confidentiality and, importantly, no negative consequences for their career or standing within the organisation. This is why efforts such as Beyond the Label remain important. By challenging stigma and encouraging open conversations about mental health, they help create environments where people feel safer seeking support before they reach a crisis point.

We need to move away from the idea that mental health support is something people turn to only when they have reached a breaking point. Looking after our mental well-being should be seen as a normal part of working life, just as we take steps to look after our physical health.

Organisations can help by making support visible and accessible, but culture is ultimately shaped by how people experience the workplace day to day. This is where managers have a particularly important role because they are often the people employees interact with most closely.

Managers do not need to be mental health experts, nor should they be expected to diagnose what someone may be going through. What they can do is listen without immediately judging or trying to solve the problem, check in when someone appears to be struggling and help connect them with the appropriate support when needed.

They can also set the tone through their own behaviour, whether it’s respecting boundaries outside working hours, encouraging people to take breaks or showing that asking for help does not make someone any less committed to their work. These small, consistent actions can help make conversations about mental well-being feel much more welcome.

Lastly, managers can also show vulnerability and encourage psychological safety by being open about their own experiences and vulnerabilities. When leaders demonstrate that they are human, acknowledge that they experience stress and have moments when they struggle, it helps to create a culture where employees feel safer being open about their own challenges.

AI can be useful when it helps organisations see patterns across the workplace that may otherwise be difficult to spot. For example, survey or engagement data might show that a particular team is experiencing sustained pressure, or that workloads have increased over time. This type of insight gives an organisation an opportunity to investigate what may be happening and determine whether something needs to change.

Where we need to be much more careful is when those signals are used to make conclusions about an individual. Mental health is complex. Workplace data can never capture everything that may be happening in someone's life, nor should behaviours such as working late, taking leave or becoming less engaged automatically be interpreted as signs of a mental health condition.

The role of AI should be to help organisations ask better questions about the workplace, rather than attempt to predict who may or may not be struggling. Used in that way, it can support prevention without reducing a person’s well-being to a set of data points.

Responsible use starts with being very clear about what the technology is there to do and what it should never be used for. At TOUCH, we believe technology can support insights and decision-making, but it should never replace human judgment, empathy and context.

If an organisation is looking at workforce-level trends, the purpose should be to identify where working conditions or support may need to improve. It should not become a way to single out employees, label someone as being “at risk”, or influence decisions about their performance, progression or suitability for a role.

Wherever possible, employers should work with aggregated and anonymised information rather than tracking individuals. There should also be clear limits around who has access to the information, how long it is kept and what decisions it can be used to inform.

Additionally, it is important to remember that an alert is treated as a prompt for further inquiry, not as a conclusion. It should encourage leaders to ask what may be happening within a team or workplace and have meaningful conversations where appropriate, rather than assume the technology already has the answer. 

Ultimately, employees are more likely to trust and benefit from these tools when they are designed to support well-being rather than monitor behaviour. The distinction is critical in ensuring that early-warning systems remain a source of support, not surveillance.

Trust has to come first. If employees are unsure about what is being collected or worry that information about their well-being could somehow be used against them, even a well-intentioned initiative can create more anxiety rather than support.

Employers need to be upfront about what information is being collected, why it is needed, who can see it and how it will be used. They should also collect only what is genuinely necessary and put strong safeguards around sensitive information.

There should always be appropriate human oversight as well. If technology surfaces something about an employee, that person should never be left in a position where an automated interpretation becomes the final word. They should be able to ask questions, raise concerns and, where appropriate, correct or challenge how their information has been understood.

And where personal or sensitive mental health information is involved, employees should have meaningful choice over whether they participate, without fearing that opting out will disadvantage them. Ultimately, trust comes from giving people clarity and agency over information that is deeply personal to them.

Preventive care starts with listening. Technology may help an organisation recognise that something needs attention, but understanding what needs to change comes from the people experiencing the workplace every day.

That is why lived experience needs to be part of the conversation from the beginning. Employees who have experienced mental health challenges can offer perspectives that a survey or dashboard cannot capture. Rather than designing programmes first and asking employees for feedback afterwards, organisations can involve them in shaping those approaches from the outset.

Listening also needs to lead to action. Prevention is not identifying someone who may be struggling earlier and asking them to cope better – It’s having a meaningful conversation and truly listening, as well as looking at the environment around them to determine whether workloads are sustainable, whether managers are equipped to respond with empathy, whether expectations are clear and whether people feel comfortable speaking up before they are overwhelmed. 

This effort begins with all of us. The upcoming BTL Fest, taking place from 23–25 October 2026, is themed “Mental Wellness Begins With All of Us” and is a a gentle reminder that each of us has a part to play, whether it is by checking in on a colleague, listening with care, or helping someone feel safe enough to ask for support.

When technology helps us see where to look, human conversations help us understand what is happening and lived experience helps shape what we do next, we move closer to a genuinely preventive approach. Most importantly, employees remain participants in that process rather than becoming subjects of it.