AI in market research: Can you use ChatGPT to chat with your consumers?
Whether you’re tired of hearing about ChatGPT or you’re actively using it in your day-to-day life, there’s no escaping the fact that this new technology is causing waves.
According to UBS, ChatGPT is the fastest-growing app of all time. It’s been estimated that 100 million people were actively using it in January, only two months after its launch. To put this into context, it took nine months for TikTok to reach the same number of users.
To get to grips with ChatGPT and what it means for you and your role, you don’t need technical know-how. Instead, it’s just worth knowing the basics.
Let’s start with the term ChatGPT. What does this actually mean?
GPT stands for generative pre-trained transformer, while the ‘chat’ element refers to the chat interactions this model has been fine-tuned for. To break this down even further:
- Generative refers to the model’s ability to recognise and generate text
- Pre-trained refers to the large data sets that have been fed into the model
- Transformer refers to the specific type of model behind ChatGPT, which has been designed to work with natural language
“ChatGPT is a new development by a company called OpenAI, and they’re known for their leading research into something called foundational models. Foundational models are machine learning models that have been trained on huge data sets, and they’re designed to be fine-tuned to a wide range of specific use cases.
ChatGPT is a fine-tuned large language model, and it’s been fine-tuned for chat interactions, as its name suggests.” said our Head of Platforms, Gavin Harcourt.
Through these chat interactions, you can ask ChatGPT a question about almost anything. You can also ask it to perform certain tasks for you, like composing an email, creating a string of code, or writing a poem in the style of Shakespeare.
But how is ChatGPT able to understand what’s being asked of it and generate a response?
“Well, the model is essentially trained on a large amount of text data, which in the field we call The Corpus. More specifically, it has been trained on books, articles, Wikipedia websites, social media platforms and more, and I think that’s why it’s so good and versatile at understanding different languages, different dialects, and different tones of voices.” explained our Product Manager of Machine Learning, Shafeeq Shahajan.
By training ChatGPT on so much natural language data, the model is now able to recognize, summarise, translate, predict and generate text. This is all achieved without any complex codes or scripts. Just the natural language we use in our day-to-day lives.
While your business might already be using this type of technology for functions like innovation and operations, have you also considered using it to get richer and more valuable insights from your consumers?
“The application for this kind of technology spans from things like customer service and self-serving chatbots to things like translation and the generation of legal documents. It’s super exciting. If you think about the consumer research pipeline specifically, there are lots of different steps it could be incorporated into.” said Shafeeq.
By condensing and analysing huge volumes of data, ChatGPT could save a significant amount of time and effort when it comes to generating a hypothesis, collecting consumer responses, analysing the data and summarising insights.
“Typically, the industry has approached data acquisition through very static, closed question surveys and Streetbees has played a big part in revolutionising the way this is done by enabling open text questions and open text data classification at scale. What ChatGPT allows us to do is take that even further by enabling freeform natural conversations with consumers to get deeper into their mind and extract richer and more accurate insights.
The second point is around the analysis of that data. If a customer responds to a closed text question, that’s pretty easy to analyse. But when it’s an open text question i.e. they respond in their own words, rather than choosing from a list of predefined answers, we’re now entering into the realm of unstructured data. At Streetbees, we have a history of effectively and efficiently processing unstructured data. Now, with ChatGPT, we can take that even further and can look at classifying open text data in real time.
The final thing is around the generation of insights. Now, I’m by no means saying that ChatGPT could completely automate this step. However, it’s a great starting point and is a very useful tool for strategic consultants to use to go even deeper with their recommendations.” shared Shafeeq.
ChatGPT has the potential to revolutionise the way we conduct market research. However, it’s important to know when to use it and when not to use it, as like any tool, this technology has its limitations.
Here’s a few you should keep in mind.
- Although the answers generated by ChatGPT might sound plausible and pretty convincing, the model has no ‘knowledge of knowledge’ and can therefore provide incorrect answers.
- The model can be sensitive to the wording we use when we ask it a question or give it a prompt, so it can provide different answers depending on the way the question or prompt is phrased.
- Although the model’s intelligence is artificial, it’s been trained on words written by people and therefore, there’s a risk it could pick up and display some of our biases in its outputs.
To summarise, ChatGPT is:
- A type of generative pre-trained transformer, which has been fine-tuned for chat interactions using natural language. No complex codes or scripts. Just the simple language we all use in our day-to-day lives.
- A technology that could have huge implications for market research and could save a significant amount of time and effort when it comes to generating a hypothesis, collecting consumer responses, analysing that data and summarising insights.
- Not a silver bullet. Although ChatGPT might seem ‘magic’ in its ability to answer questions and complete tasks, it has no ‘knowledge of knowledge’ and can therefore get things wrong. As we explore the use cases for ChatGPT in the market research space, it’s important that we approach this responsibly and with a sense of experimentation.
AI and Machine Learning is at the heart of what we do here at Streetbees and the introduction of these new generative models will allow us to get even closer to consumers and their everyday choices. We’re currently exploring new ways to incorporate the power of ChatGPT into our solution and our pilots are coming on leaps and bounds.
We’ll continue to keep you posted on these exciting new developments to bring you research that’s quicker, more accurate and more actionable. If you have any questions about ChatGPT and its implications for market research, get in touch!