In the rapidly evolving tech landscape, the allure of new technologies can be irresistible. Companies and individuals alike often rush to adopt the latest innovations, eager to stay ahead of the curve or simply to boast cutting-edge capabilities. However, this rush can lead to the adoption of technologies like AI not because they are the best fit, but because they are the current trend. Remember the blockchain hype, where its application was forced into scenarios that gained little from its use? In contrast, advancements like progressive web apps and AI have proven their worth, offering tangible benefits when applied thoughtfully. However, the key is not just to adopt technology for the sake of innovation; rather, it’s crucial to consider what tangible benefits these tools bring to the table.

Low Hanging Fruit

One of the most straightforward and beneficial applications of AI, specifically Large Language Models (LLMs), is in the domains of transcription, translation, and summarization. These tasks, traditionally labor-intensive and time-consuming, are prime examples of “low hanging fruit” for AI integration in business processes. By employing AI for these functions, organizations can achieve significant reductions in operational time and costs, making these essential tasks more manageable and less resource intensive.

Consider the transcription of audio and video content. Traditionally, this required hours of manual labor, listening to recordings, and typing out content verbatim. LLM AI transforms this process by automating transcription with high accuracy, freeing up human resources for more complex tasks that require nuanced understanding and decision-making.

Similarly, translation services, which are crucial in our increasingly global marketplace, can be expedited with AI, enabling businesses to communicate effectively across language barriers without the delays and expenses of traditional translation services.

Summarization, particularly of open-ended responses in surveys or robust video content, is another area where AI can provide substantial efficiency gains. This capability allows businesses to quickly grasp the essence of collected data, facilitating faster decision-making and analysis. The ability to handle larger volumes of data without a proportional increase in processing time or costs opens new avenues for gathering richer, more nuanced insights from diverse data types like long-form audio interviews or customer feedback videos.

Thus, by integrating AI in these key areas, companies not only streamline operations but also enhance their ability to handle and benefit from more complex and voluminous data sources. This strategic application of AI ensures that businesses can scale their data processing capabilities without being overwhelmed, maintaining a competitive edge in information management and insight generation.

Making Jam, Or Marmalade, Or Preserves (Pick Your Extended Fruit Metaphor)

This is where it gets truly exciting for us. As a market research services company with the soul (and chops) of a software company, we thrive on incorporating new technologies into our workflow. The thrill isn’t just in using AI; it’s in strategically deploying it to maximize its strengths while understanding and mitigating its limitations.

In the realm of data collection, AI’s applications are both innovative and transformative. For instance, AI-powered chatbots can simulate human-like interactions, making surveys more engaging and less robotic. This approach not only can improve the respondent experience but also can increase the likelihood of completing the survey, enhancing quality and quantity. AI can also dynamically tailor questions based on previous responses, making the survey feel more personalized and relevant to the respondent.

Predictive analysis and data augmentation can be added to this fruit cocktail. AI can analyze responses in real-time to predict trends and even suggest additional questions to probe deeper into specific topics revealed during the survey. We’ve even explored adding closed ended questions to surveys on the fly using AI. This adaptive questioning technique can ensure that we’re not just collecting data, but we’re collecting the right data.

From Chatbots, to AI image generation in quantitative and qualitative settings, to endless other creative integrations, I could write a book about ways that AI can be used in MR. The trend that you would see in the book (and in the paragraphs above) is the use of the word CAN. In the immortal words of Goldblum in Jurassic Park, “your scientists were so preoccupied with whether or not they could that they didn’t stop to think if they should.” Don’t get me wrong, you should. You absolutely should! But like most things, you need to find the nuanced and intelligent application for your reptile DNA splicing.

The Concern (and a Deviation From Fruit)

Too often we get asked if we can “add some AI,” like salt or pepper that can be added to any dish. And, like salt and pepper, it’s on the table and can easily be added, which can be problematic. You can add the seasoning of AI without any thought to the flavors you’re trying to create, or (a pet peeve of mine when I cooked the meal) before even tasting what is on your plate!

But I digress…

You have to ask if the AI is enhancing your tools, doing what you were already doing, or making things worse. One of the most common uses we see is adding AI to probe further on an open end. This is similar to adding a “can you please tell us more” when the word count is too low. Yeah, it’s more conversational but if you’re telling the respondent they are interacting with AI (which you should be) they likely won’t treat it any differently. What if you used the opportunity to make sure you cover the areas you are looking to know about? Like asking a hotel customer to tell you how they liked their stay and use AI to make sure they mentioned the bed, the housekeeping, and the ease of booking. Then you would get their unprompted response but also get data on the items you are most curious about.

Outside of proper implementation questions, AI can have additional drawbacks. The most significant is perhaps the risk of data bias. AI systems learn from the data they’re fed, so if the input data is biased, the output will likely be too. This can skew research and lead to misinformed decisions. Therefore, it’s crucial to constantly monitor and refine the AI models to ensure they remain unbiased and effective.

There can also be the challenge of integration. Ensuring that AI systems seamlessly integrate with existing data collection can be complex and resource intensive. It requires not just technical know-how but also a strategic vision to ensure that the integration adds value without disrupting the current process.

When preparing your research meal, AI tools can be good ingredients to have at your fingertips. Be thoughtful about how you use these tools that we have been given. Find someone that can help you in the kitchen. And, if you’re not comfortable cooking for yourself, maybe go out to eat and have someone cook for you.