Beyond the Prompt: Owosina Hands NAC 2026 A Practical Framework for Working With AI

Adebayo Owosina, Co-Founder and Chief Creative Officer of The Hook Creative Agency, has told delegates at the 2026 National Advertising Conference that generative AI has raised the floor of creative production without raising the ceiling, and that the industry must learn to divide labour between what machines can do and what only people can.
Speaking at a breakout session at the ongoing National Advertising Conference titled “Beyond the Prompt: Balancing Generative AI Efficiency with Authentic Human Stories” on the final day of the conference at the Abuja Continental Hotel, Owosina presented a three-part framework of Hand, Hold and Prove for deciding what to hand over to AI, what to hold back, and how to account for its use in the work agencies release.
The session drew heavily on his own practice at The Hook, including campaigns for Meristem and Airtel and the “Fear Us” campaign that was exhibited at Google Creative Lab in New York. Owosina argued that the industry’s current enthusiasm for prompting has produced competence without distinction. “Everybody is a prompt engineer now,” he said. “And you’d think that because the floor is getting higher, the roof is supposed to get higher. Personally, I have not seen really great work in a while.”
“…The school of advertising is dying”
Owosina had his first diploma was in Animal Health and Production, followed by a second in Animal Production and Health. He spent almost twelve years going through university strikes before finding his way into advertising through a first job at an advertising agency, where he was asked for a portfolio and presented work designed on a BlackBerry Curve 2. That unlikely route, he noted, showed that the real school of advertising is the first two to three years of hands-on agency work.
That school, he warned, is under serious threat from AI. The learning came from the frequency of revisions, the back-to-back rewrites, the repetitions that build judgement. “My guys will just prompt now,” he said, “whether their prompt is right or not.” He regrettably noted that the craft that emerged from the old grind like the ability to sit with a problem, to generate options, to discover things on the way to a solution is being skipped. He described the loss in terms of what the research journey used to yield including the tangential websites, the programmatic ads, the accidental discoveries that fed creative instinct. He noted that prompting delivers an answer without the journey, and the journey was where the learning lived.
“Insight cannot be prompted”
Noting that Generative AI learns from existing data to create new text, images, videos and code, Owosina noted that Storytelling, by contrast, is the art of using words, images and sound to connect with an audience. Without connection, there is no storytelling.
He returned repeatedly to the “Fear Us” campaign as evidence. The idea came from a personal experience where he was singled out from a queue of about thirty travellers at a European airport and detained for nearly ninety minutes, he realised the officer was not investigating him so much as performing fear of him. The campaign turned that stigma into a statement of capability which an AI would not have been able to produce.
A copywriter, he added, would have argued that “Fear Us” sounded negative and needed whitewashing. The campaign was exhibited for three weeks at Google Creative Lab in New York, and plans to take it to schools with Nigerian populations in Brooklyn, Queens, etc, were underway before COVID intervened.
The “O’to ge!” campaign in Kwara State is another example made the same point from a different angle. The campaign required a week of listening to communities, including people without phones, to understand their pain points. AI would have produced generic material, he argued, because the people in question are not in the training data. Nigeria has more than 120 million citizens who are offline. “The models have read everything except the people you want to sell to,” he said.
The framework: Hand, Hold, Prove
Owosina’s practical contribution was a framework he described as the product of his own working practice. The Hand category covers what can be delegated to AI with limited risk like sifting through CVs and portfolios, reviewing interview calls, quick research, transcribing recordings, generating mockups, mood boards and variants, and reporting. Even here he urged caution, noting that AI is a sycophant, …“whatever you feed to it, it’s going to agree with you” and is therefore only useful to people who already know what they are looking for.
The Hold category is the creative core that must not be delegated. Insight generation heads the list. “It’s not even up for debate.” The final call on whether work is good enough to release also stays with people, because AI will approve whatever it is asked to approve. Language checks must be done by native speakers, not by translation tools otherwise, he said, campaigns created in Lagos end up cursing audiences in Sokoto or Katsina. And the value of a person in the room who disagrees with you cannot be replicated. “A person who disagrees with you in the meeting is gold,” he said, describing The Hook’s flat structure, with no corner offices for managing directors, as a deliberate attempt to make challenge easy. If a junior’s alternative gets the work to 85 per cent and the boss’s preferred option gets it to 87 per cent, he said, let the junior win. AI, by contrast, will simply agree that the work is fine.
The Prove category concerns accountability. He cited this year’s D&AD awards, where the Yellow Pencils shortlist contained no AI-executed work because it did not meet the storytelling standard. Agencies, he argued, need to start tagging their work with how exactly AI was used, who signed it off, and what was done with the time saved. Clients pay for time, so if AI is genuinely saving time, the return should show up either in better margins or higher quality. If neither is happening, the use of AI should be questioned.
Going further, he defined great taste as the ability to discern quality, intent and emotional resonance…knowing what works, why it works, and what to leave out. That cannot be acquired through prompting, he noted, but comes from time spent with audiences, understanding the intention behind every action.
In closing, he charged the industry on its responsibility for the next frontier: “The weight of the future of advertising in Nigeria is on our shoulders,” he said. “Everybody has a stake in it whether you’re a client or an agency. The task is how to preserve the creative integrity of our dear profession. AI cannot save us. At least we know that now.”
