The AI Research Revolution

By: Jack Wilson and Jonathan Smetherham, 2CV

Blue Planet Studio – Shutterstock

The potential for artificial intelligence to enhance the market research industry is significant both in terms of efficiency and quality of insight. 2CV is at the forefront of innovation as we test, learn, and adopt new tools and techniques. Here is our rundown of the current state of play, how AI is creating ‘data dialogue’, and why human expertise is still a vital ingredient (for now).

Market research: an industry ripe for disruption

At the heart of successful market research lies the art of curating and deciphering vast volumes of diverse data and extracting invaluable insights that empower businesses to gain a competitive edge.

However, with many traditional, manual processes still firmly entrenched in everyday practices, the industry is perpetually teetering on the brink of disruption. While some innovative strides have been made in the research technology (ResTech) domain, few have genuinely transformed the core qualitative and quantitative methods that are essential to our field.

This began to change in 2023, as artificial intelligence has moved beyond the realms of science fiction and become a palpable reality for many of us. The pace of change is rapid and the simple reality is that AI has the potential to replace (or reduce) roles that were traditionally held by researchers.

As we have tried new approaches, we’ve learned a few lessons along the way.

AI is already increasing business efficiency and part of our BAU for routine tasks

AI is already being used to automate a variety of routine research tasks, as well as to capture and analyse large amounts of data faster and more efficiently. For example:

  • During the pre-project phase – AI is being used to synthesise, edit, and summarise the myriad documents that are needed to formulate a research strategy and articulate ideas;
  • At project set-up – AI is helping to generate and refine survey questions and response codes;
  • At the analysis stage – AI aids with transcription and translation, thematic analysis and querying data.

However, across all these uses, the outputs tend to be fairly basic and more in line with what might be expected from a very junior/early-career researcher who happens to be extraordinarily fast. Their output still requires significant curating/editing.

Beware of hype and manage your expectations

AI should not be blindly heralded as the panacea for all research challenges. AI analysis can’t be trusted implicitly: be sceptical of ‘AI-washing’, whereby marketing materials tout ‘AI-powered everything’ or ‘instant answers’. Expert intervention is still required and researchers must provide editorial oversight. Checking sources and referencing source data remain essential to mitigate hallucination, bias or overclaim.

This also holds for more nuanced or high-level tasks. AI can tell you what happened, but it will generalise and overlook crucial details. Critically, since your AI model lacks a detailed understanding of your objectives or what certain stakeholders need to know, and what details might be particularly valuable, it cannot tell you what matters or why.

Data dialogue: where AI and the human touch intersect

AI truly stands out in one area: the creation of ‘data dialogue’. AI breathes life into data, enabling conversations from the very moment it’s collected (thanks to conversational AI) and throughout the analysis process.

Conversational AI attracts attention, elevates response quality, and expands qualitative techniques. At the same time, quantitative surveys transform into captivating stories that are more rewarding for respondents. Crucially, it eliminates researcher bias, providing swift and precise access to specific audience verbatims across the entire sample. At 2CV, we have developed two specific products that enable this:

  • TalkBot: 2CV’s chatbot survey tool, which sparks natural conversations, powered by AI-generated probing questions, creative exercises, and in-depth analysis. It seamlessly merges quantitative and qualitative research, transforming projects like comms testing, customer experience, innovation, and brand/product research.
  • SmartProbe: An AI-powered module integrated into 2CV’s traditional survey open-ends. It uncovers profound qualitative insights by generating follow-up questions, leading to richer, more emotional responses and actionable findings. SmartProbe seamlessly enhances quantitative research with qualitative depth and automated pre/post-tasks.

While AI-assisted analysis underpins both products, human researchers remain essential to examine the analysis, verify the evidence, and craft the narrative. Explaining research findings is an art that requires human intelligence and emotional understanding – things that AI cannot copy, at least not yet.

Conclusion

In summary, we expect AI to revolutionise research in the next 3–5 years. But for now, the human touch cannot be replaced: AI can’t improve bad data or capture emotional nuance.

Yet the benefits far outweigh the risks. To make the most of AI, exercise due diligence and rigorously test AI applications to see how they can meet your specific needs. Small-scale pilots can offer valuable insights before showing solutions to clients, ensuring a prudent and informed approach.

To understand how we can help you on your AI research journey, reach out to us today at singapore@2cv.com.

This article was first published in the Q3 2023 edition of Asia Research Media

Leave a Comment

Your email address will not be published. Required fields are marked *