Image by Marco Bianchetti
By Liz
We’ve been talking a lot about AI at work lately. And at family braais and weddings and over coffee with friends.
It feels like every digital tool we have now also has an AI assistant.
Generative AI has brought some interesting opportunities to our work. We can speed generate a large number of options and tackle time-intensive tasks more efficiently.
It also brings undeniable challenges. We’ve received 20-page AI-generated briefs that could have been one paragraph written by a person. Our testing has shown that AI can confidently produce work that looks impressive but doesn’t actually speak to the problem or take into account any nuance.
While it can save time in one area, checking and refining its outputs can take more time in another.
So how do we make the most of this tool without compromising the quality and thoughtfulness that good work requires? I think we need to lean into our humanness.
AI is only as good as the thinking behind it
One of the most important things to understand about generative AI is that it amplifies whatever input it is given.
If you provide a hastily written prompt, it will produce a lot of poorly considered outputs, very quickly. But give it a clearly defined problem, strong objectives, excellent strategy and a good sense of what success looks like, and it can be a powerful tool.
Here are ten human skills that are critical to directing, interrogating and refining generative AI’s capabilities.
1. The ability to ask the right question
Crafting a useful prompt requires fully getting to grips with the challenge at hand in all its nuance, navigating assumptions and surface-level solutions. The more thoughtfully the challenge is defined and the more curiously AI’s answers are probed, the more useful the technology becomes.
2. Critical thinking
One of the biggest risks of generative AI is how convincing it can be. An answer can sound completely authoritative without it necessarily being correct, strategically sound or appropriate.
AI can create options. Human judgement determines whether an option is worth using.
3. Taste
AI can generate 50 versions of something but only a skilled person can recognise the one that I can work with. This skill comes from developing a strong internal understanding of aesthetics, tone, relevance, cultural nuance and authenticity.
4. Strategic thinking
AI can do a lot of things but only we can answer the question: “Should we do this?”. To answer this requires an understanding of organisational objectives, audiences, competing priorities, trade-offs and long-term consequences.
5. Area expertise
There’s a misconception that AI will make specialist skills available to everyone. I think the opposite is true.
It’s only with appropriate expertise that a person can discern if a generated piece of work is truly any good, if it’s appropriate, what is missing and what needs to be challenged.
Human expertise and experience is how we know if AI is executing the right thing, in the right way.
6. Creativity and imagination
There is an important distinction between generating and creating.
AI can generate combinations of ideas, images, words and concepts at extraordinary scale but humans bring all those things that make art magical and rich and weird and brilliant and original.
7. Communication
Using AI effectively relies heavily on us to be good communicators. We need to translate between our intention when creating the input, the AI system and the outcome we’re working towards.
That means being able to articulate particularly well at every level of an interaction with AI: what we want, why we want it, what success looks like, what needs changing and what should be rejected.
8. Editing and refinement
It takes a slight perspective shift in the beginning, when it seems dazzling what AI can produce almost instantaneously, but I now always see the first output as raw material. The real work begins after the initial generation: selecting, fact-checking, simplifying, and infusing nuance.
It requires editorial skills to turn something generic into something meaningful and effective.
9. Emotional intelligence
AI can simulate empathy, but only humans understand humans.
This distinction is particularly important in our work, where success depends on much more than delivering a technically sound website. It also involves reading the room, managing misalignment, building trust and reading between the lines.
As more of our lives become AI-assisted, genuine human relationships will become even more valuable. Technology can make interactions more efficient, but it can’t replace the trust that comes from feeling seen and understood.
10. Ethical judgement
Just because AI can do something doesn’t mean it should. We are the custodians of what is ethical and responsible.
We need to think about whose interests are being served, who might be excluded, and what the potential consequences of using a particular tool might be.
What does this mean for the future of our work?
I don’t think the answer is a blanket rejection of generative AI, nor do I think the answer is to build it into every area of our business. For every tool that we come across, we assess if it serves us, our work and our clients. This means understanding where AI can add value and being clear about its limitations.
In many ways, managing AI is like managing an extremely capable junior collaborator. It can be productive, but it needs extensive direction, context and feedback.
If anything, putting AI through its paces has highlighted for me that our work is not just building technically sound websites, it’s understanding organisations, their objectives and their audiences, and translating this understanding into meaningful digital experiences.
