Forget AI Experts. I'm Hiring People Who Can Talk to AI.

The stack of resumes on my desk here in Delhi all started to look the same. A sea of sameness, really. Every other candidate was suddenly an ‘AI Expert,’ a ‘Prompt Engineer,’ or a ‘Generative AI Specialist.’ Six months ago, this would have thrilled me. In the frantic, early days of the AI gold rush, I was convinced that we needed to hire our way into the future. My agency, Pulse AI, felt the pressure to have these specialists on the payroll, like a badge of honour proving we were on the cutting edge.

So I hired a few. They came with impressive certifications and spoke a lingo filled with terms like ‘token optimization,’ ‘parameter tuning,’ and ‘zero-shot prompting.’ They could build complex chains of prompts that looked more like code than conversation. The problem was, they couldn't actually create anything great. Their output was technically precise but creatively sterile. It lacked soul.

I remember one project vividly. We were developing a campaign for a new, high-end brand of artisanal masala chai. The client wanted visuals that felt authentic, warm, and distinctly Indian—a modern take on a timeless ritual. Our newly hired ‘AI Expert’ took the lead on generating concept art. He approached the AI, one of the leading image models, like a vending machine. His prompts were surgically clean: ‘Photorealistic image of a steaming cup of masala chai on a rustic wooden table, shallow depth of field, warm morning light, 4K resolution.’

The AI delivered exactly that. A flawless, generic, and utterly boring stock photo of a cup of tea. It could have been taken in a cafe in Stockholm or Seattle. There was nothing of India in it, nothing of the monsoon rain outside the window, nothing of the fragrant steam carrying notes of cardamom and ginger. My AI expert was frustrated. He tweaked the parameters. He changed the lens type. He kept adding technical jargon. The images got sharper, the lighting more dramatic, but the feeling remained absent. He finally threw his hands up, blaming the limitations of the model. ‘It just can’t grasp the cultural nuance,’ he declared.

He was wrong. The tool wasn’t the problem. His approach was.

At the same time, one of our junior copywriters, a young woman named Meera, had been quietly experimenting on her own. She had no certifications, no formal AI training. What she did have was a novelist’s sensibility and a boundless curiosity. One afternoon, I saw her hunched over her screen, whispering to her monitor. I walked over and saw she was ‘talking’ to the same image generator.

Her process was completely different. It was a conversation. She started with a story. Her first prompt was something like: ‘An old soul in a young body, sitting by the window of her small apartment in a leafy lane of Bandra. It’s raining outside. She’s wrapped in a hand-spun shawl, reading a dog-eared book of poetry. Next to her, a beautiful clay kulhad holds steaming chai, the steam catching the silvery light. The mood is pensive, peaceful, and a little bit magical.’

The first image the AI produced wasn't perfect. The cup was wrong, the woman looked too Western. But the mood—the mood was there. Instead of getting frustrated, Meera responded. She continued the conversation. ‘That’s a lovely start,’ she typed. ‘Let’s make a few changes. The cup should be a simple, unglazed terracotta kulhad, the kind you find at a railway station. And the woman should look more distinctly Indian, maybe with a small nose ring. Let’s also add a stack of old books on the floor next to her chair. The mood is right, just deepen the shadows a bit.’

She went back and forth with the AI, five, six, seven times. Each iteration was a gentle course correction, a new layer of detail, a refinement of the narrative. She wasn't commanding the machine; she was collaborating with it. She was guiding its vast, non-human imagination toward a specific human feeling. The final image was breathtaking. It was everything the client wanted and more. It wasn't just a picture of tea; it was a story in a single frame. It was the feeling of home on a rainy day.

That was the moment everything clicked for me. The skill we needed wasn't ‘prompt engineering.’ It was ‘AI fluency.’ It’s the ability to have a patient, iterative, and creative dialogue with a non-human intelligence. It’s understanding that the AI is not a search engine or a calculator. It’s a very powerful, very eager, and very literal-minded junior colleague. It has access to all the information in the world but has no lived experience. It has no taste, no intuition, no sense of irony or culture of its own. Your job isn't to give it the perfect command. Your job is to be its guide, its curator, its creative director.

This revelation changed how I hire. I stopped scanning for ‘AI’ on resumes. Instead, I started designing my interviews around conversation. I now give every promising candidate a simple, open-ended creative brief. ‘We need to launch a new eco-friendly scooter in a crowded market like Mumbai. The target audience is young, environmentally conscious, but also short on time. Give me a campaign concept.’ Then I sit them in front of a laptop with ChatGPT, Midjourney, and a few other tools, and I say, ‘Show me how you’d start. Talk me through your process.’

I’m not looking for a polished final product. I’m watching their process. Do they get frustrated when the AI misunderstands them? Do they give up and call the tool stupid? Or do they get curious? Do they rephrase their questions? Do they try new angles? Do they feed the AI’s bad idea back into it to see if it sparks something new? Do they know how to layer ideas, to critique an output, and to build on it? I'm looking for the conversationalists, the storytellers, the patient collaborators.

This isn't just a phenomenon I'm seeing in my Delhi office. I talk to friends and colleagues in London, New York, and Singapore, and they’re all arriving at the same conclusion. The initial rush to create siloed ‘AI departments’ filled with technical wizards is proving less effective than empowering everyone to become AI-fluent. The real transformation happens when the strategist in your Bengaluru office can use AI to model market trends, the designer in Mumbai can use it to brainstorm visual identities, and the copywriter in Gurgaon can use it to break through writer’s block. The value compounds when it’s a distributed capability, not a centralized specialty.

AI fluency is built on a foundation of soft skills that we’ve always valued: curiosity, patience, critical thinking, and the ability to articulate an idea clearly. It requires you to know your own domain deeply, so you can effectively critique the AI’s output. A great art director can guide an image model better because she has a deep well of visual knowledge. A great writer can coax a better story out of a language model because he understands narrative structure and tone. The AI doesn’t replace their expertise; it amplifies it, but only if they know how to talk to it.

We are moving from an age of instruction to an age of conversation. The most important user interface for the next decade of technology will not be a keyboard or a mouse, but our own natural language. The people who thrive will be those who can use words to guide these incredibly powerful new systems toward interesting, beautiful, and useful outcomes. They will be the translators, the whisperers, the creative partners to artificial minds.

So, I’m not hiring AI experts anymore. I’m hiring curious people. I’m hiring patient teachers. I’m hiring poets and storytellers and critics. I’m hiring people who know how to have a good conversation, and who have the grace to guide a brilliant, tireless, and clueless new colleague toward making something wonderful.