Selectio blog - insights on hiring, testing & HR

I write here about topics that come up in my work and might be useful to you, whether you're running a business, managing a team or working through your own career. You'll find a mix of practical advice on recruitment, HR, employment law and getting ahead at work.

Where it helps, I bring in what the research and I/O psychology tell us. If there's a topic you'd like me to cover, let me know using the Suggest a topic button.

Your staff are running their employment agreements through AI - here's what that means 

By Lynne Allison · October 2026

Something has changed in my HR advisory work over the last year. More and more often, a new employee or an existing staff member comes back to their employer with questions about their employment agreement, and it's clear they've run it through an AI tool first. They arrive with a list of clauses they want to discuss, change or push back on, based on what AI has told them.

Here's a recent example. I drafted individual employment agreements for several new starters. A couple of them came back with queries, and all were identical. They were written the same way, raised the same points, and in the same order. None of the points actually related to New Zealand employment law as it stands. It was pretty clear AI had written them.

This is happening here in New Zealand, and it's worth understanding, whether you're an employer or an employee.

It's not just here

Overseas, it's becoming common. In the UK, employment tribunal judges have noted that AI is contributing to more complex claims and to longer, more legal-sounding grievances. In Ireland, workplace complaints rose 44% in 2025, and employment lawyers there believe AI tools are part of the reason, because they make the process easier for people without a lawyer.

I'd expect New Zealand to follow the same path.

Why it isn't all bad

Employees have always been entitled to understand their employment agreement and to seek advice before signing it. For a lot of people, especially those who can't afford a lawyer, AI makes that feel much more accessible. A staff member who reads their agreement properly and asks questions isn't being difficult. In many ways that's a good thing.

A concern is still valid even if AI helped someone put it into words.

Where it goes wrong

The problem is that AI isn't always right, and it doesn't know your situation. In the conversations I've been part of, I've seen:

  • Overseas law applied to a New Zealand agreement. AI often draws on Australian, UK or US rules that don't apply here.
  • Standard, reasonable clauses flagged as unfair. Things like trial periods, confidentiality or reasonable notice, just because they sound restrictive.
  • Confident answers that are simply wrong. AI can sound very certain, and sometimes refers to cases or laws that don't exist.
  • Advice that tells people what they want to hear. If you ask AI whether a clause is unfair, it's inclined to agree with you.

There's also a privacy risk. Pasting an employment agreement, or letters about a workplace issue, into a public AI tool means that information is no longer in your control.

What employers can do

  1. Make sure your agreements are clear and up to date. They're now being read more closely than ever. With the 2026 law changes, and the Holidays Act being replaced in 2028, now is a good time for a review.
  2. Stay calm and respond to the substance. Work out what the person is actually worried about. Often there are only one or two real issues under a long list of points.
  3. Explain the reasons. If a clause is standard and lawful, say why it's there. Most people are reassured once they understand it.
  4. Be open to reasonable changes. Negotiating terms is normal, and good faith applies in both directions.
  5. Have a simple AI policy. Let staff know what they shouldn't put into public AI tools.

What employees can do

  1. Use AI to help you understand things, not to make decisions for you.
  2. Check what it tells you against Employment New Zealand's free information at employment.govt.nz.
  3. Be careful what you paste in, especially other people's personal information.
  4. Get proper advice before you take action, particularly if you're thinking about challenging your employer or raising a grievance.

How I can help

I work with both employers and employees. If you're an employer and a staff member has come back with a long list of AI-generated concerns, I can help you work out what matters and respond fairly. If you're an employee and you'd like a proper check of your employment agreement or a workplace issue, I'm happy to talk it through.

Personal grievances just got harder to win - here's what that means for you... 

by Lynne Allison - September 2026

If you run a business or manage people, you'll want to know about this one.

In February this year, the Employment Relations Amendment Act 2026 came into force. It changed one of the most important parts of how personal grievances work in New Zealand: what happens when the employee's own conduct played a part in the situation.

What's actually changed

Under the old rules, an employee's conduct was one factor the Employment Relations Authority weighed up when deciding what to award. It could bring the amount down, but it rarely changed the outcome altogether.

That's no longer the case. Under the new law:

  • If an employee's conduct contributed to the situation but doesn't amount to serious misconduct, they can no longer be awarded reinstatement or compensation for hurt feelings, humiliation or loss of dignity. Lost wages may still be considered, but that's at the Authority's or Court's discretion.
  • If the conduct amounts to serious misconduct, no remedies are awarded at all. That means no reinstatement, no lost wages, nothing.
  • More broadly, the Authority can now reduce any remedy by up to 100% where the employee's actions contributed to what happened.

It's a big change, but it isn't a free pass for employers. The Authority and Court will still look for a genuine, direct connection between what the employee did and the situation that led to the grievance. If your own process was the real problem, such as a rushed disciplinary meeting or a decision made before the employee had a proper chance to respond, the employee's conduct on its own won't necessarily save you.

Why this matters more than it might seem to

It's tempting to read this and think "great, less risk for me". And there is less risk, if you get it right.

The law hasn't removed the need to do things properly. A sound process is what lets you benefit from these changes if a grievance does come up. That means your documentation, your investigation and your decision-making all need to stand up to scrutiny.

Where to start - and why it doesn't have to start with a lawyer

Here's what I want business owners and managers to know: you don't need to go straight to an employment lawyer every time a tricky situation comes up. A lot of the risk in personal grievances gets built in early, during a disciplinary process, a restructure or a difficult conversation, long before anything reaches a lawyer's desk.

Getting good HR advice at that early stage is almost always more cost-effective than dealing with the legal side afterwards. A short conversation about how to run a process properly, what to document, and what "genuine contribution" needs to look like can save you a lot of time, stress and money down the track. It costs a fraction of what legal representation does once things have escalated.

This is a big part of my HR advisory work: practical, evidence-based advice on getting people decisions right before they turn into a bigger problem.

Get in touch

If you're not sure how these changes affect the way you handle disciplinary processes, dismissals or difficult conversations with staff, I'm happy to talk it through. A short conversation now is a lot cheaper than a legal bill later.

Using AI in your job search: what helps and what hurts  

by Lynne Allison - August 2026

Job hunting is hard work, and AI tools are now everywhere. I want candidates to stand out because of their own experience and their own story. AI can help with parts of your search, but it shouldn't write your application for you.

Where AI helps

AI can be useful for finding job listings, spotting roles you might have missed, and working out which of your skills match a particular position. Used this way, it can make the early part of your search easier.

When it comes to your CV, cover letter and application questions, though, your own words will always do better.

Why your application should sound like you

I don't recommend using AI to write your application. Recruiters read a lot of CVs, and AI-written ones are getting easier to spot. They tend to be generic and repetitive, and they often miss what's actually interesting about your experience.

A recent LinkedIn post by hiring manager Jeff Grbelja (Director, Verse) shared what happened when he advertised a role. Out of 304 applications:

  • only 32% followed the instructions and included a cover letter
  • 28% were from overseas applicants without the right to work here
  • 24% had no relevant experience
  • 16% didn't attach a CV
  • more than 70% were clearly written by AI.

That matches what I've seen. I've managed more than 17,000 applications a year, and recruiters quickly learn to recognise the patterns: the same phrases, the same structure, and sometimes bits of the AI prompt that people forget to delete. When seven out of ten applications look the same, it's very easy to say no.

So be yourself, and write it yourself. That alone will help you stand out.

Follow the instructions

Recruiters aren't trying to trick you. They're trying to work out whether you can do the job, which is why they ask for particular documents or questions.

  • If you're asked behavioural questions, never write "refer to my CV". It tells the recruiter you're not taking the role seriously.
  • Screening questions are there to show your capability. If you don't answer them properly, you'll almost certainly be declined.
  • More employers now ask for a cover letter in your own words, often to filter out AI-written applications. If you ignore that, you're putting yourself at a disadvantage.

Interview preparation

AI can help you practise. You can use it to come up with likely questions or to rehearse your answers. But it can't replace a real conversation.

From coaching hundreds of candidates, I've found that people who practise out loud with a friend, mentor or recruiter feel more confident and do better on the day.

Using AI sensibly

  • Use AI to find opportunities, not to write your application.
  • Be careful with your personal information, and only use platforms you trust.
  • Read your application through before you submit it, and make sure it sounds like you.

How I can help

I really enjoy working with candidates. If you're not sure how to approach an application, get in touch. I can help you work out your strengths and how to show them in your own words.

The bottom line

AI can make parts of your job search easier, but the application needs to be you. Take your time, use your own words, and let your real experience do the talking.

How to write good interview questions, and why it matters 

by Lynne Allison - July 2026

Interview questions are one of the best tools you have for working out who will do well in a role. In I/O psychology, good questions start with a clear understanding of the job and what it takes to succeed in it.

Start with the role

Before you write a single question, get clear on the main tasks, responsibilities and challenges of the position. Two good places to start:

  1. Talk to the person doing the job now. They'll know what the role really involves day to day.
  2. Use the position description. It should set out the key tasks and the competencies, skills and abilities a successful person needs. If it's out of date, talking to the current person in the role will help you update it before you recruit.

Use a structured behavioural interview

I recommend a structured behavioural interview, where every candidate is asked the same questions. The questions focus on the key competencies for the role, and ask candidates for real examples of when they've shown those skills.

A structured interview:

  • is one of the best predictors of who will do well in the job
  • helps you avoid costly hiring mistakes
  • makes the process fairer for every candidate
  • can also point to training and development needs once the person starts.

It's also worth opening the interview by asking why the candidate is interested in the role and what made them apply. It gives you a feel for their motivation and interest in working for you.

Writing the questions

Once you know the core competencies, you can use AI to help draft questions. Just check that each question really matches your competency definition.

Here are two examples for problem solving:

  1. "Tell me about a time when you identified a problem that others hadn't noticed yet. What was the problem, and how did you address it?"
    This looks at whether the candidate spots issues early, thinks through their impact, involves the right people and takes the initiative to fix them.
  2. "Describe a situation where you had to solve a problem with incomplete information. How did you approach it?"
    This looks at how the candidate handles uncertainty, finds the information they need, weighs up risk and makes decisions under pressure.

How many questions?

It depends on the level of the role. As a guide:

  • an administrative role usually needs around four behavioural questions
  • a management or executive role usually needs five to seven.

These are in addition to your opening question about why they're interested in the role.

Consider a work test

A work test, based on a key task the person would do in the job, can be a strong predictor of performance, especially alongside a structured interview. For example, you might ask candidates to prepare a short presentation on a topic related to the role. There's more on work tests on my [Recruitment page].

Test your questions

Before the interviews, try your questions out in two ways:

  1. Read them out loud. How do they sound? Are they too long or wordy? Written and spoken language are different, so make sure you're comfortable asking them.
  2. Try them on someone you trust, and who won't pass them on to candidates. Ask whether the questions make sense, whether the language fits your organisation, and whether they understood what you were asking.

Build a rating scale

This is probably the most important step. A rating scale is how you measure the level of competency a candidate shows in each answer.

I usually use a 1 to 5 scale, and "anchor" each number with a description so everyone scores the same way. For example:

  • 1 = poor
  • 2 = limited
  • 3 = acceptable
  • 4 = strong
  • 5 = very strong evidence of the competency

It helps to add a sentence or two describing what a strong or weak answer looks like for each question.

Get your panel on the same page

For your scores to mean anything, your panel need to rate answers consistently. That means:

  • your competencies are clearly defined
  • everyone on the panel knows the competencies, the questions and what the role needs
  • everyone agrees what a "5" looks like, and what a 1 to 4 look like too.

A short briefing before the interviews makes a big difference.

Why it matters

A structured approach, with questions based on the competencies that matter and a rating scale your panel understands, gives you a fair, consistent way to compare candidates. It also means you're much more likely to hire the right person first time.

I help organisations design structured interviews that measure what really matters for the role. If you'd like help with your questions, rating scales or the whole recruitment process, get in touch.