How AI surprises caught me out.
I’m a big fan of AI, and use it in some form, most days. But recently I found myself in a situation where it resulted in my wasting a lot of time.
A client had asked me to be involved in a recruitment exercise to interview for a new marketing assistant. All the preliminary stages were completed and we were down to the short list. The CVs had been checked, and we were about to start online interviews to triage the list down to two or three for face-to-face interviews.
My role was to observe and assess the candidates technical marketing skills.
My client had created a great pre-interview questionnaire. Candidates were asked to comment on the client’s current marketing strategy, media choices and collateral – and were also asked how they would approach the same issues if they were appointed.
All candidates put in really good responses so I was looking forward to some great interviews.
The first candidate started off strongly as my client asked some general questions, but as we started to probe the answers to the questionnaire, it was obvious that something was wrong. They answered with a string of buzzwords – often inappropriately. I jumped in and asked for some explanations and it was clear the candidate actually had only a sketchy understanding of what marketing actually is.
As we moved through the interviews similar patterns were repeated. One candidate used a lot of acronyms, which they pronounced wrongly, and when challenged, could not say what they stood for. Another had suggested using certain metrics, but could not explain how they could be used.
As we discussed the interview afterwards, it became apparent where we had gone wrong. Using the questionnaire had been a mistake. It was too good and too explicit. The candidates had obviously put it straight into ChatGPT (and who could blame them?), and it had generated some very sensible (and similar) responses. Of course, the argument is that they should have used the AI results to prepare better for the interview – except that foolishly, we had set the tripwires for ourselves.
We’d wasted the best part of a day. Had we not provided the questionnaire and gone straight to the face-to-face interviews, we could have achieved better results, faster.
So, what have I learned from this experience? It’s to be aware that we now live in an AI world now. Others will use AI, in fact I would be disappointed if they didn’t. But we must prepare for the behaviour of other players. To ensure our own efficiency, we must consider what AI responses might be to our business actions, and plan accordingly.
to employ a new marketing assistant for them. All the preliminary stages were completed and we were down to the short list. The CVs had been checked, and we were about to start online interviews to triage the list down to two or three for face-to-face interviews.
My role was to observe and assess the candidates technical marketing skills.
My client had created a great pre-interview questionnaire. Candidates were asked to comment on the client’s current marketing strategy, media choices and collateral – and were also asked how they would approach the same issues if they were appointed.
All candidates put in really good responses so I was looking forward to some great interviews.
We should have seen it coming
The first candidate started off strongly as my client asked some general questions, but as we started to probe the answers to the questionnaire, it was obvious that something was wrong.
They answered some questions with a string of buzzwords – often inappropriately. I jumped in and asked for some explanations, and it was clear the candidate actually had only a sketchy understanding of what marketing actually is.
As we moved through the interviews similar patterns were repeated. One candidate used acronyms that they pronounced wrongly, and when challenged, could not say what they stood for. Another had suggested using certain metrics, but could not explain how they could be used.
As we discussed the interview afterwards, it became apparent where we had gone wrong. The questionnaire we used had been a mistake. It was too good and too explicit.
The candidates had put the document into ChatGPT (and who could blame them?), and it had generated some very sensible (and similar) responses. Of course, the argument is that they should have used the AI results to prepare better for the interview – but no excuses, we had to admit we had set the tripwires for ourselves.
The problem was not that the candidates had used AI, but that we had not been prepared for the result and not thought the scenario through.
We’d wasted the best part of a day. Had we not provided the questionnaire and gone straight to the face-to-face interviews, we could have achieved better results, faster.
So, what have I learned from this experience?
It’s to be aware that we now live in an AI world now. It’s not enough to have our own developed strategies in place. Others will use AI. Clients, suppliers and the public – in fact I would be disappointed if they didn’t. But we must prepare for the behaviour of other players. We must consider how AI responses to our output might look, and structure our activities accordingly.
You might also be interested in:
Marketing for more
Marketing for more is a normal reaction – but what ‘more’ do you really want? And how much more? And by when? Get under the skin of ‘more’. What does it mean?
7 Old-school marketing approaches that still pack a punch.
In our AI-driven, digitally dominated world, it’s easy to overlook old-school marketing approaches. However, these time-tested strategies add impact and results.
