Ways in Which Artificial Intelligence Can Complement Qualitative Research
Back in 1950, the famous Alan Turing question, "Should machines think?" He opened his 'Computing Machinery and Intelligence' paper. Jumping much ahead in time, the developments in artificial intelligence (AI) are manifold and cover almost all facets of our lives. AI serves millions every day, whether by your mobile, your vehicle, your bank or your home. Often you use a service such as Siri or Cortana actively, other times it's less evident and it also goes unnoticed quite much. The AI wave is having an effect on markets, including market research, across the board.
AI in Qualitative Research
AI is fundamentally oriented and powered by computational data based on mathematical models. You may believe that this somewhat conflicts with the descriptive essence of qualitative studies. Qualitative research is, after all, concerned with a selection of methodologies to explore; to dig into the depths and reasoning of human behaviour. Issues in qualitative research need to be contextual, large and non-static in order to accomplish this. Moderation is often defined as 'messy' and immeasurable as intuitive and data production, text, audio, visual and audio-visual. AI goods have progressed beyond mere numbers today, however. They are now prepared to deal with, or at least aid in the process, this 'messy' immeasurable material.
Both blogs and white papers have been written about the direct function of chatbots in qualitative moderation and the function that broader AI techniques can and will play in the collection of qual and quantum data. Analysing and interpreting.
1. Insight Community Engagement
MROCs are a perfect example of an approach to market research that offers large quantities of qualitative data production. For any community manager, the sheer volume of community member content created over weeks, months or years is a challenge. In qualitative data analysis, text and sentiment analysis are already in full swing, but there is another exciting application of these instruments in that of community participation insight.
Market research can leverage AI to predict member disengagement before it happens by applying deep learning to a combination of text and sentiment analysis, username, behaviour and profile data. This helps the (human) researcher to proactively handle member interaction, without firefighting, to minimize churn. And there is nothing to avoid the expansion of this clever AI technique into interaction optimization for deep learning to 'learn' what engages who, when and where. All of this leads to reviews of better quality for both researchers and clients.
2. Conversational Insight Activation
We have already seen numerous market research dashboard applications enter the marketplace, all designed to provide access to real-time data to internal stakeholders in an enticing way to envision, i.e. to promote insight-based action. But what if we took it a step further ... beyond the dashboard to the chatbot for insight activation.
The AI needed is still double for an entry level insight activation chatbot, i.e. one that deals with quantum data only:
1. The analytical power of a large number of different quantum data sets to identify correlations
2. The ability to express this connection in natural language, with support for visual
Working from the back of a database type dashboard, an entry level bot of this nature will be able to examine both historical and current data and summarize only important observations in bite-size qualitative chunks that can be easily digested. In order to replicate Q&A style interactions with company members, individual customer persona chatbots may also be programmed. We're using qual now to integrate insight into companies, which is pretty clever.
We could take our chatbot even further if we were to add deep learning into the mix. This will allow for constructive usage that is more innovative. You no longer have to wait to be called upon by your friendly insight activation chatbot; rather, it can push insights out to members of the organization. Feedback from recipients on relevance when they can and the bot learns what is useful for their job position, at what time, and even the language to resonate with. It's got the culture of wisdom written all over it ... with one small caveat. It would have to come with full study ties to each bit-sized perspective. A chatbot defining causation even when qual study data is applied to the quantum database? From my view, there are a few too many gaps in text and sentiment analysis for that, at least as it stands today.
3. Researcher's Assistant
The use of virtual assistants, both commercial and personal, has been brought to a whole new level by developments in both natural language and voice processing. This type of AI can be used commercially for consumer, information and entertainment services as well as buying functions and has become increasingly popular with major players such as Amazon, Facebook and Google.
But the ability to create a standardized research assistant still exists. The 'Researcher's Assistant' will provide participant information services either full-time, or out of hours, as in the commercial scenarios referenced above. The research project will essentially 'staff' it. This staffing would be limited to natural language processing and generation based on an information bank in the current AI circumstances, but the RA would nevertheless provide a degree of interactive support to improve involvement where it was previously a challenge to do so, i.e. evenings, weekends, research projects spanning various time zones, and / or free the human researcher to focus on the human researcher
Conclusion
Market research organizations and practitioners welcome AI with open arms, attempting new methods and set-ups for research. Researchers and AI are required to work alongside each other to generate (qualitative) perspectives that were previously outside the realms of possibility. AI is here to stay. Let's proactively get involved. Let's lead the way.