How to choose the right AI for the task

A creative director is using AI to speed up the generation of advertising jingles. He produced multiple flawless melodies in minutes, but something wasn’t right. “It sounded perfect, but it felt too perfect, like it did not breathe,” he said when interviewed for a recent study. The jingles were technically excellent quality, but lacked warmth and emotional authenticity.

At times, that missing ‘warmth’ from AI output is hard to define. But it’s important to understand the limitations of generative artificial intelligence (GenAI), given that more and more businesses are using it for creative work and business operations.

In a paper published in Business Horizons, Colin Campbell, Sean Sands, Catherine Ferraro, H. Y. J. Tsao, and Esade professor Alexis Mavrommatis argue that businesses should stop treating AI as a single, all-purpose capability. Instead, managers need to understand the different kinds of intelligence it can offer, and that some are more useful than others depending on the scenario and task at hand.

AI has more than one kind of intelligence

To help us understand AI, the researchers draw on psychologist Howard Gardner’s theory of multiple intelligences. They propose a framework that distinguishes eight forms of GenAI intelligence: logical-mathematical, linguistic, audio, spatial, interpersonal, contextual, intrapersonal, and existential. Each refers to a different capability, from analyzing information and generating text to understanding social situations and reflecting on human values.

That’s not to say that AI thinks like a human. It doesn’t. But looking at GenAI through this lens offers managers a clearer way to understand what different AI systems can and can’t do. For example, a system that excels at writing content may be less capable of picking up on cultural nuances or responding properly to complicated emotional scenarios.

Not all of GenAI’s capabilities are developing evenly. Text generation, analysis, and image creation are proving useful for many businesses already. In contrast, GenAI’s understanding of social situations and its contextual sensitivity are less reliable.

The researchers advise managers to avoid labelling AI as a single tool. They highlight the need to assess its specific capabilities. “AI’s effectiveness depends on how well its intelligences are matched to the demands of a given task,” they argue.

Fast and convincing does not always mean effective

AI can help businesses increase productivity by speeding up creative work. One product designer interviewed for the study spoke about a dramatic change in production speed: “It used to take 6 or 8 hours for one photoreal render. Now, I see 10 variations in 10 minutes.”

Professionals can explore more ideas in less time, but producing realistic images or eloquent audio doesn’t always guarantee that the result will be appropriate, distinctive, or emotionally engaging. A campaign can be professional but generic; a message may be grammatically correct while missing the cultural context that makes it meaningful.

A recent example is a YouTube ad for Voyager Broadband in New Zealand, featuring AI-generated content, that was pulled following complaints about its portrayal of Māori and Pasifika people. Intended to be humorous, the ad showed three men watching a sports game online. When the internet started to lag, the two Māori/Pasifika men made animal-like noises and attacked their white friend. The ad illustrates how AI-generated content can reproduce harmful stereotypes when cultural and societal sensitivities are not adequately considered.

These issues aren’t limited to the creative industry. AI can generate an answer that sounds confident and complete, but it may fail to address the actual business problem. This is why it’s imperative to understand whether an AI system’s capabilities match the task it’s assigned. The question for managers deciding how to use AI is: Can AI produce an answer? Yes. But can it produce the right answer for the situation?

Why human oversight is needed

So now AI is used daily in business; have we come to trust it? A 2025 survey of 125 US advertising executives, by the Interactive Advertising Bureau (IAB), found that more than 70% had encountered an AI-related incident, including inaccurate, biased, or off-brand content. Yet despite this, fewer than 35% planned to increase investment in AI governance or brand-integrity oversight over the following year. Faster production can bring clear opportunities, but if content isn’t reviewed by humans, those opportunities can be lost.

Increased output from AI isn’t necessarily a win if audiences respond negatively. In an August 2026 Bentley University–Gallup survey, 49% of US adults felt negatively about businesses using AI to create advertisements, compared with 19% who viewed it positively. Respondents were more in favor of AI helping employees develop ideas rather than AI replacing people’s images and voices. The way businesses use AI can matter as much as whether or not they use it.

How managers can make better AI decisions

The researchers propose a framework of three practical priorities for managers to help them determine the best use for AI.

First, match AI capabilities to the task. Don’t merely decide to use AI without first defining the specific problem to be solved and then choosing an AI with the correct capabilities. A tool that efficiently generates initial concepts may not be appropriate for making sensitive decisions that require contextual judgment.

Second, experiment and evaluate. Managers should assess AI outputs against clear objectives. Speed and fluency don’t always result in quality. Human oversight to compare results, check for errors, and assess whether outputs meet customers’ and employees’ needs can help reveal where AI adds value and where it falls short.

Third, manage human and AI capabilities side-by-side. The researchers advocate a shift from traditional personnel management to intelligence management: understanding what people and AI systems each contribute, and how they can complement one another. Crucially, rely on humans for judgment, ethics, cultural understanding, and managing brand reputation.

The advantage lies in choosing well

Generative AI is accessible to most businesses, so the competitive advantage isn’t found in merely using AI. Successful use of AI in business is more likely when leaders consider where to apply it, how to assess the results, and how they combine machine capabilities with human judgment.

The researchers’ framework helps guide those choices. To enhance business success, managers should ask: which kind of intelligence does this task require, and how can we bring it together with human expertise to create something genuinely valuable?