As nations and organisations continue to reap early dividends of new-age technology Artificial Intelligence (AI), as well as explore additional use-case potential, industry experts have argued that prompt engineering will evolve to be the prime point of focus, especially in generative AI (GenAI).
Prompt engineering is the process of structuring or crafting an instruction in order to produce the best possible output from a generative AI model.
The process could involve specifying parameters such as tone, length, level of detail, format or context, guiding the model to provide responses to better align with the user’s objectives.
By learning how to tailor prompts effectively, analysts say, users can unlock the full potential of AI thus creating more sophisticated workplace outputs from minimal input.
Pundits observe that while this skill was initially viewed as a technical necessity, it has since crept into business strategy, marketing, customer service and a host of other business processes.
Anthony Muiyuro, who serves as the East Africa regional director of IT infrastructure firm Syntura Group, explains that prospects of mass mastery of the skill have been eased by the fact that it is equated to coding, only that this time it’s with use of a natural language.
“The beauty about it is that it’s a natural language. Prompt engineering is basic English, but it is fundamentally knowing how to position the words right. I understand there are even tutorials to guide people on how to input statements that will likely give the most accurate output,” he says.
Syntura Group, global Managing Director Chris Evans points at flexibility, explaining that users will attain the prompting skill through multiple trials.
“Rather than having a pre-conception of what you need to ask, you have to try doing it differently because nobody really understands how all of this works and how phrase your question is important. So, try a few different ways,” notes Evans.
“It is the openness to change; trying things is the key thing here.”
While the duo acknowledges that AI responses are never going to be 100 percent aligned, they stress that prompt engineering is about taking the accuracy to as close as possible to the mark.
Tech entrepreneur Mbugua Njihia observes that the rise of prompt engineering will likely come with some unintended consequences, including enhancing human-to-human interactions as people will learn to be more precise in communicating their needs.
“On the one hand, businesses that master the art of structured questioning can extract immense value from AI, automating processes with precision and optimising decision-making,” wrote Njihia in a recent Business Daily commentary.
“On the other hand, as we become more adept at formulating clear, goal-oriented prompts, we are also forced to confront a deeper reality; most of us are inept at asking the right questions or packaging our asks succinctly,” he added.
Mr Muiyuro further highlights the role of user feedback in advancing this capability, saying the dynamic engagement aids greatly in sharpening AI answers.
“For instance, you’ve asked it for something, and it brings it in very illustrative language and then you can tell it, ‘Give me clear and precise language’ or you can tell it to dramatise the answer or to add data and statistics,” he says.
“So, I think it’s also fun when you’re actually engaging it and telling it the extent to which you want it to tweak the response, and this could get you sounding as natural as possible.”
Among the factors identified as hindrances to effective prompt engineering include AI model limitations, which serve as a blocking wall beyond which a user’s creativity cannot be applicable, as well as unintended biases where prompts inadvertently lead to biased or inappropriate responses especially in sensitive areas.
