A Happy Medium

Like most opinions on the internet, public response to AI is only becoming increasingly polarised, especially when it comes to discussing AI’s impact on the labour market. Major companies have been undergoing ‘strategic restructuring’ in which employees are ‘displaced’ in favour of a more ‘digital-first’ approach. Underneath the various euphemisms the sentiment holds true: AI is replacing jobs.

Whether that’s right or wrong, a good thing or not, it can be really frightening. Feeling any amount of uncertainty in your job is not a comfortable place to be and the overarching message that you either ride the train of progress or get left in the dust, is not a very helpful one.

The solution likely lies somewhere in a murky grey area, which is not very marketable, so I’ll lead with the term: A happy medium!

Working in a private specialist company, as opposed to a large enterprise business, the effect that AI is having on us is much more tangible and controlled. With the managing director and our head of technology on board, here’s how we use AI in our roles to evolve the platform and stay competitive in the world of competency assessments…and we still get to keep our jobs (yay!)

 

A TOOL THAT SERVES THE SOFTWARE ENGINEER

In software development AI removes the small, repetitive friction points that typically slow developers down. Tasks such as writing boilerplate code and refactoring simple functions are not technically difficult, but they’re boring. By using AI in a meaningful way, our software engineer can spend less time on these things. A 2025 study carried out by Google surveyed 5,000 technology professionals and found that 80% reported that using AI increased their productivity. Over half of the participants believed that AI also had a positive impact on the code quality. Globally the shift towards using AI as a tool for software engineers has already been made, and it’s how companies adjust to the change that counts.

Now, if your software engineers are able to work quicker then you mightn’t need so many of them (tempting, as labour costs are the biggest cost) Alternatively, you could look into reclaiming the time saved and reinvesting it into building better, more innovative solutions (this one takes our vote!)

With AI influencing our product roadmap, we are able to push for greater innovation: designing adaptive systems, solving complex behavioural problems and building a better user experience. Our developers spend less time in the weeds and more time brainstorming the future of our platform and the direction we want to go in.

I also think it would be remiss of me to not acknowledge that AI has proven to be less than reliable at times. We still rely daily on our software engineer’s deep expertise in competency assessment, thoughtful design, and good engineering judgement. AI is an invaluable tool for us, but a tool nonetheless, and after all, a tool is only as good as the person using it.

All that being said, here’s what we’re doing with AI (please don’t count how many times I say AI)

 

ADAPTIVE 360

Adaptive 360 takes a competency assessment and transforms the rating process into a two-sided conversation with adaptive questions. Rather than participants answering a fixed set of questions, they instead engage in an AI driven conversation which evaluates the input in real time and adapts the questions accordingly.

Think of it a little more like a skilled interview. Adaptive 360 asks follow up questions, prompts for greater clarity, asks for examples and uncovers behaviour patterns. However, unlike an open-ended AI chat system, Adaptive 360 operates within a clients structured competency framework, and because the questioning is adaptive, the experience can be shaped to suit a wide range of contexts and objectives.

Without going into too much detail (here’s our Adaptive 360 article that does) I think it’s important to note that the AI isn’t the one making decisions or drawing conclusions. The AI surfaces patterns, while Halo’s structured assessment engine governs scoring and ensures consistency. The system also provides a reasoning statement for its recommendation, so users can understand how and why conclusions were reached. In other words, behaviours, not conversations, remain the unit of assessment.

We’re also working on:

 

AI-PULSE

An upgraded experience to our traditional pulse check in’s.

AI-pulse can generate short, targeted mini surveys or realistic role-play scenarios for the participant to respond to, adapting questions based on behaviour trends and previous assessment data.

The embedded AI can then go one step further and provide personalised coaching guidance and nudge participants towards improvements on an ongoing basis.

The hope is that instead of a static ‘check-in’ participants are supported in continuous development.

One more for you to chew over:

 

AGENTIC AI CAPABILITIES

This one makes a little more sense if you already know the difference between agentic AI and generative AI (which I did not) but essentially, it’s an AI model that operates independently and proactively to maintain long-term goals and track progress over time.

Utilising agentic AI will serve as a neat little feature when it comes to rolling out larger projects, where it might not always be ideal to wait for the end of a project to receive feedback and data insights. The agentic AI system can use the emerging data immediately as it flows in; identifying trends, key themes and qualitative feedback throughout the running of a survey project.

It provides another level of insight and adaptability to the survey projects and allows for more agile development strategies in response. Leaders can see what’s happening, as it happens. To use a phrase much loved by AI itself, these valuable metrics will be delivered in real-time!

 

Additionally, we’re also undergoing a full revamp of our platform and website (busy busy) but that’s the point, AI see’s us looking beyond the hum drum of a daily workload and thinking big picture.

The caveat, not all work that can be outsourced to AI should be done by AI.

In a company designed to help measure and support people’s behaviour and development, there would be something inherently dystopian about attempting to replace the people in the equation.

We have two support teams, both in Australia and Ireland that are dedicated to resolving user issues and ensuring smooth roll out of assessments and projects. Surveys themselves are built by our team and proof checked for errors, and we employ graphic designers to make all the cohesive branding specific visuals on client portals.

Every AI output is reviewed, every decision owned by a human, and every new feature is grounded in real expertise.

AI also serves as a bad replacement for human communication, if you’ve ever been sent an unedited AI response in lieu of a regular email comm then you’ll know what I mean. If its people speaking to other people, then there’s no need for AI to butt in, which is to say, we’re not being encouraged to use AI to outsource writing (phew!)

I also think there’s no replacement for human creativity…but I digress.

For the small fish in a big pond, AI isn’t so much the villain. It sees us doing more engaging work, feeling less stressed and expanding our roles, whilst delivering smarter products and faster innovations. A win-win all around.

 

Click here to reach out and request a demo.

 

 

Published by Halo Feedback, written by Erin Power