|
Getting your Trinity Audio player ready...
|
I spent years working in employment services before I moved into learning design. What stands out most about that work is the people on the other side of the desk – clients navigating circumstances that were often fragile, urgent, and complex. If the AI tools that exist today had existed then, the temptation to use them in that environment of volume, documentation, and relentless throughput pressure would have been real and entirely understandable.
It is something I think about in my consulting work and in my graduate research on AI training in human services. This is what I have come to call Unanchored Efficiency.
The risk leaders are watching for – and the ones they miss
When leaders think about AI risk, the concern is usually focused on visible misuse. The assumption is that risk will show up through struggling or resistant staff who produce low-quality outputs, make errors or avoid using the tools altogether.
That scenario exists, but there is also an overlooked risk. Given how quickly AI is being taken up across the sector, it is probably growing. That is the risk I have come to call Unanchored Efficiency – and it may already be present in your organization or practice.
What Unanchored Efficiency looks like
Unanchored Efficiency is what happens when AI use in an organization outpaces the frameworks, guidance, and shared understanding needed to keep it aligned with professional standards. It is not a product of poor judgment or weak values. It tends to surface when capable people are operating in conditions that reward speed, where leadership has not defined what should stay constant.
None of this is a reflection of the practitioner’s values. It is a reflection of what the organization has and has not made clear.
The issue is not what the tool produced but what happened to professional judgment along the way.
This distinction matters particularly in employment services, where clients are often navigating circumstances that make them specifically vulnerable. A newcomer managing an immigration process, a client returning to work after a health disruption or someone sharing something sensitive because they trusted the person in front of them are all situations where professional judgment is not an added benefit. It is the work.
When efficiency begins to displace that judgment, even incrementally, something important changes.
The scale of the gap
The conditions that give rise to Unanchored Efficiency are well documented, even if the phenomenon isn’t named.
A 2025 IBM Canada study found that while 79% of full-time office workers reported using AI at work, only 25% were using enterprise-grade tools. The remainder were relying on personal applications or a mix of personal and employer-provided tools, operating largely outside organizational oversight. A KPMG Canada survey from the same year found that just over half of Canadian workers are now using generative AI at work, but 83% say they want or need further training to use it effectively, and fewer than half feel their organizations provide adequate support.
A separate KPMG global study, conducted with the University of Melbourne and surveying more than 48,000 people across 47 countries, found that 57% of employees reported hiding their AI use and presenting AI-generated work as their own. Only 40% said their workplace had any policy or guidance on generative AI use at all.
These figures are not from employment services contexts specifically, but there is little reason to think the sector is insulated from dynamics that appear this consistently across industries. If anything, the combination of high caseloads, limited organizational resources and work that depends heavily on professional judgment gives those dynamics more room to take hold.
Where organizations tend to land
The research points to a gap. In practice, that gap tends to show up in a few recognizable ways.
- High use, low visibility. Some organizations have significant AI use already underway but limited visibility into how it is occurring. Tools are in use across teams, but there is no shared understanding of which ones, for what purposes, or how client information is being handled. The efficiency gains are visible, but what is happening underneath them is not.
- Low use, low awareness. Others have very little AI use but also very little awareness of what that means. A few quiet workarounds may already be in place. The conversation about AI has not yet happened in any meaningful way, and the longer it waits, the more ground there is to cover.
- Aware, not yet moving. Some organizations have thought carefully about the risks but have not yet found a clear path forward. The concern is real, but the response has not yet taken shape.
- Deliberate and intentional. And some have moved carefully, establishing shared expectations around appropriate use, data handling, and professional judgment before gaps become problems. These organizations are not necessarily the most advanced AI users, but they are intentional ones.
Unanchored Efficiency is most visible in the first pattern, but the second is where it tends to quietly take root before anyone has thought to look. The fourth is worth holding onto as a horizon – it does not require sophisticated technology or extensive resources, but it does require clarity about what AI should and should not do, and a shared understanding among staff about what that means in practice.
Understanding which of these comes closest to your organization is a reasonable starting point. It shapes what needs to happen next and in what order.
The questions worth asking
Unanchored Efficiency does not emerge as a single identifiable issue. It develops through a series of small, individually reasonable decisions that accumulate over time. Responding to it begins with building visibility into current practice.
For management:
- Where is AI already shaping how work gets done in your organization, and how visible is that in practice?
- Which parts of your team’s work rely most heavily on professional judgment, and how is that being protected or altered through AI use?
- What assumptions are you making about how staff are using these tools, and how confident are you that those assumptions reflect reality?
- When uncertainty arises, do your teams have a shared understanding of what requires human judgment and what can be supported by a tool?
For practitioners:
- At what point in your workflow are you relying on AI, and what role is it playing in your decisions?
- When an output looks complete, what do you use to determine whether it is actually appropriate for this client?
- What information are you entering into these tools, and do you understand where it goes and how it is handled?
- Where has AI made your work faster, and what, if anything, has changed in how you apply your judgment as a result?
Unanchored Efficiency is not a reason to step back from AI. In many cases, these tools provide real value by reducing administrative burden and creating space for more focused work.
What matters is not whether AI is used, but how its use is understood.
Developing a clear view of current practice of AI readiness in organizations, including who is using these tools, how they are being applied, and what assumptions are shaping their use, is often the most practical place to begin. The organizations that will serve their clients well are not those that move the fastest. They are the ones that remain deliberate about what should and should not change.
