Unlocking AI’s Potential: Generative AI, Prompt Engineering, and Workplace AI Adoption

Sprout Solutions recently hosted a webinar called Practical AI: Generative AI, Prompt Engineering, and Workplace AI Adoption. It was aimed at demystifying AI adoption for organizations and offering valuable insights into generative AI and prompt engineering. 

Experts from Sprout were invited as speakers, such as Gian Dela Rama, Chief Product Officer and Head of Sprout AI Labs, Head of Strategy and Operations (AI Products) Betty Margulies, and Mid AI Engineer Edward Vendivil.

Read more to learn more about some of the webinar highlights!

One of the primary objectives of the webinar was to address common misconceptions surrounding AI. Many professionals fear that AI will replace human jobs or become “too intelligent.” 

Dela Rama remarked that a few years ago, people thought about super-advanced robots (a la Terminator) when talking about AI, highlighting the fear brought by the technology to many.

It’s a good thing, though, that people are becoming more aware and adept at using the technology, shared Dela Rama. He also noted that over 86% of working Filipinos think of ChatGPT and its competitors when talking about AI. He also emphasized Sprout’s AI use manifesto.

There are also a lot of people, particularly those who heavily use AI now, that think that the technology is infallible. However, there are  certain risks involved when using AI for work. 

According to Dela Rama, generative AI has the tendency to “hallucinate” because it may tell the user false information if its source data is not enough or if the prompt is unclear.

AI adoption is dramatically increasing worldwide, especially within the last couple of years. In fact, about 72% of businesses use the technology in at least one business function, according to a survey report cited in Margulies’s discussion. 

She noted that while the Philippines might be lagging behind in the use of AI, it’s still a fact that many companies based in the country are starting to explore the advanced technology.

This is understandable considering the benefits of AI. A study cited by Marguilies showed that workers’ performance can improve by up to 40% if they use AI to complete some of their tasks. The problem, however, is that their performance can also significantly drop (by 20% to 25%) if they ask AI to complete tasks that it isn’t designed to do.

The solution is simple: learning how to use AI effectively by addressing the most common challenges related to AI adoption:

Margulies emphasizes that companies should not blindly mandate the use of AI tools. To achieve the best results, careful planning and understanding how it relates to business growth is crucial. 

For example, companies should assess areas where repetitive or time-intensive tasks exist—such as data entry, customer support, or supply chain management—and determine whether AI can enhance efficiency or free up employees for more strategic work. Organizations should also consider the long-term implications of AI adoption, including cost, scalability, and integration with existing processes.

Proper training is necessary to help people understand and harness the potential of AI to the fullest.

This means offering employees workshops, step-by-step guides, or ongoing support to help them use AI tools effectively. Providing hands-on practice sessions where employees can learn how to use AI for specific tasks will help create a more AI-literate workforce.

Employees may resist adopting new technologies due to fear of job loss or discomfort with unfamiliar tools. Change management strategies are essential to address these concerns and create a culture that is more open to innovation. 

You can start by clearly communicating the benefits of AI adoption for both the company and the employees. It’s also important to stress that AI will not replace humans but rather just make their jobs easier by reducing mundane tasks and creating opportunities for upskilling.

Prompt engineering plays a huge role in enhancing operations and customer interactions But what is it?

Prompt engineering refers to the process of crafting detailed commands (or “prompts”) to help GenAI come up with the best possible answers to your queries. 

“If GenAI is like our smart assistant, then prompt engineering is how you tell our assistant what to do in the most effective way. So, it’s all about how you phrase your input, how you ask, how clear you ask it, and what context you are giving it,” Vendivil said. A few examples:

• Crafting email and follow-up emails
• Summarizing meeting minutes
• Improving slide decks or presentations

Vendivil also shared a few prompt engineering tips:

Few-shot prompting is a technique where you provide a language model with a small number of examples within the prompt to guide its responses. This method helps the model understand the desired output format and context without needing extensive training data. 

For instance, if you want the AI to generate professional email templates, you can include two to three sample emails within the prompt to set a clear tone, style, and structure. 

Few-shot prompting can save time and resources, especially in scenarios where collecting large datasets or fine-tuning the model isn’t feasible.

The structure and clarity of your prompts are essential. A well-structured prompt should clearly define the task and desired output format. 

For example, instead of asking, “Summarize this article,” a clearer prompt would be, “Summarize this article in 3-4 sentences, focusing on the main challenges discussed and possible solutions.” This ensures the model doesn’t deviate from your expectations and reduces the chances of generating vague or irrelevant outputs. 

Clarity also involves avoiding ambiguous terms or incomplete instructions that might confuse the model. A precise prompt not only improves response quality but also minimizes back-and-forth revisions, making interactions more efficient.

Allowing generative AI to assume a specific persona or role can enhance the relevance and engagement of its responses. By framing prompts that specify a character (e.g., “You are a marketing specialist” or “You are an HR professional”), you can guide the model to generate outputs that align with that persona’s knowledge and tone. 

This technique is useful for tasks that require domain-specific expertise or creative problem-solving. It also allows the model to adopt a style of communication that resonates with the intended audience.

Breaking down tasks involves dividing complex queries or requests into smaller, more manageable components. This approach helps the AI model focus on one aspect of the task at a time, leading to clearer and more accurate responses.

For example, instead of asking, “Explain the process of creating a website,” you can split it into smaller prompts such as “What are the steps for choosing a domain name?” and “How do you design a user-friendly interface?” This method allows for more detailed and focused responses. 

The integration of AI into the workplace is a game-changer because it unlocks new levels of efficiency, and business performance!

If you’re prepared to use AI in your workplace, Sprout is here to help. You can rewatch the webinar to freshen up your memory and get a better understanding on how key issues about AI adoption in the workplace were replaced. 

You can also check out our blog to learn more useful AI in HR insights, or contact us today so we can help you find the best solution to your needs!