Research by Omni HR points to a contradiction in the AI-era workplace: Companies expect employees to contribute more strategically, but their management practices may be teaching them to keep their heads down and follow instructions.
In an interview with SMEhorizon, CEO Brian Ip argues that AI can reduce the administrative work surrounding people management, but it can’t replace the coaching, feedback, and judgement needed to help employees take greater initiative.
How is AI adoption changing workforce expectations and revealing gaps between APAC leaders and employees?
Leadership expectations are changing quickly. Earlier this year, in Omni’s “State of AI in HR” report, 82% of HR leaders said they’re already using AI in their HR functions or plan to do so within the next year. Most see it as a way to free up employees’ time for more strategic work. The workforce hasn’t yet caught up with that expectation.
In our “Tiger Boss” report, published in August, we found a 30-percentage-point gap between employees’ confidence in following instructions and their confidence in innovating or influencing decisions beyond their roles. Leaders want employees to think bigger, but managers need to provide coaching for that transition to succeed.
How can managers address the “robot employee” problem in the AI era?
When a manager offers criticism only when something goes wrong and a quick pat on the back when something goes right, employees quickly learn that the safest approach is to keep their heads down and do exactly what’s asked of them, and nothing more. Our “Tiger Boss” report asked employees what motivates them, with respect, recognition, communication, and support for self-improvement all making the list. Employees want managers who coach them, explain clearly what good performance looks like, and give them room to try something and get it wrong without being penalised.
The data also offer an encouraging finding: Companies using five types of performance review recorded an attrition rate of 13.5%, compared with 38% among companies using only one type, regardless of pay or promotions. Recognition and substantive guidance don’t require a bigger budget. They require managers to treat feedback as an ongoing conversation rather than a once-a-year formality.
What did Omni HR learn from observing APAC HR teams when developing its agentic AI tool, Mino, and how does its infrastructure work?
We developed Mino after examining how HR teams were spending their time. Previously, a manager seeking to understand attrition trends within their team would typically ask HR to compile data manually, often from several systems, into a report. That process could take hours, although answering the question itself could take minutes, and this was not an isolated case.
When we surveyed HR leaders across the region, fragmented data, rather than AI capability, emerged as the main barrier. Most companies stored employee data across three, four, or sometimes more disconnected systems, and AI tools cannot analyse information effectively if they cannot access it in one place.
This indicated that the underlying data foundation, rather than an individual AI feature, was the issue to address. Omni therefore built its architecture around a connected system spanning every market in which a company operates, with Mino developed on top of that system rather than added later.
With that foundation established, we then considered how Mino should use the data.
That was particularly important because HR data are among the most sensitive information a company holds. Many AI tools send these data to third-party models for processing, but we didn’t want Mino to use that approach.
When someone asks Mino a question, the system doesn’t send employee records to an AI model. Instead, it reads the structure of the data, creates a targeted query, and checks that query against the user’s existing permissions. The AI then works only with the query results, not the raw records. Every query is visible and downloadable, allowing an administrator to check exactly what was requested from the database and determine whether the response matches the stored information. Mino doesn’t grant additional access when acting on a user’s behalf: It can’t access data that the user couldn’t view directly.
How can HR use AI to improve performance reviews, coaching, and learning and development?
Performance reviews are a good example of where AI can be useful, but it shouldn’t write the review for the manager. In our “Tiger Boss” report, we found that employees who received detailed feedback of around 98 words left at a lower rate than those whose comments were around 10 words.
AI can help managers provide that level of detail more quickly by summarising a quarter’s worth of notes, surfacing specific examples instead of generic praise, and flagging when someone hasn’t been recognised in months. Another Omni survey found that learning and development and performance management are among the least AI-enabled HR functions, with adoption rates of 28% and 49%, respectively.
The judgement still has to come from the manager. What AI can do is clear away the busywork around it, so managers can spend more time on that judgement instead of pulling data together.
How should HR leaders rethink employee performance metrics in the AI era?
Most performance metrics measure only output, such as sales figures or productivity. But by the time those numbers fall, employees have usually been checked out for a while.
Leaders need to focus on leading indicators instead, including manager stability, how often employees receive meaningful feedback, and whether they’re recognised consistently rather than once a year. Our research on employee attrition found that manager changes and unpaid leave predict departures more reliably than pay does. Our performance management research found that employees pull back gradually, sometimes 12 to 18 months before it shows up in engagement scores.
AI can make these signals more practical to track by identifying patterns at a scale that no HR team could manage by hand.
