Generative AI tools, like ChatGPT and Bard, are set to have as much of an impact on our daily lives as the invention of the internet and businesses have already started applying this new technology across different functions. But what do these AI advancements mean for consumer insights? Where can Generative AI technology be incorporated into the research pipeline and what kind of impact will it have?
Through our Streetbees Shares series, we're committed to sharing knowledge across the insights industry, so we’ve put together this update to bring you up to speed on what’s next for consumer research.
To help drive more innovation in the market research space, we've been doing a lot of work to understand how Generative AI can be used to make consumer insights richer, faster, cheaper and more accessible to all.
Here are some of the things we’ve learnt:
- Briefing will get 10x faster
- Pre-determined surveys will be replaced by in-depth conversations
- Insights will become instant
Let's delve into each of these.
Briefing will get 10x faster
A clear brief is essential for every research project. It outlines the research questions, the project parameters and the scope. However, the long and detailed briefs we tend to create for larger projects aren’t always necessary when it comes to more specific business questions. They can be time-consuming to create and are often changed throughout the proposal and bidding process.
With the introduction of GPT-4, you can now write a brief 10x faster!
We’ve trialled, tested and implemented this with our new AI-powered insights tool, SBX. After you answer a few simple questions from our brief assistant, the tool will create a full research brief and survey plan for you in less than 10 minutes.
Here's what we've learnt from launching this new tool and seeing clients use it:
- Setting out a clear brief will always be an important part of the research process.
- However, we will no longer need to dedicate as much time and expertise to writing these briefs, as AI-powered tools will be able to assist us.
- Instead, we should focus more time and effort on being able to truly understand and articulate the problem we’re trying to solve with the research.
- This also means that more functions across the business, who aren’t necessarily experts in creating research briefs, will be able to access and leverage rich consumer insights
Pre-determined surveys will be replaced by in-depth conversations
Many traditional surveys are designed for ‘clean’ data collection, rather than detailed and nuanced answers. Static questions like ‘how did you find the webinar?’ are often paired with closed list options like ‘I found it interesting’ and ‘I learnt something new from it’ to get an ‘easy read’ from consumers.
Although this approach means the data will fit neatly into predefined parameters, it poses more questions than answers. What did they find interesting about the webinar? If they learnt something new, what did they learn that they didn’t know already? This static style of data collection lacks the texture behind people’s perceptions and behaviours, making it very difficult to action.
The power of Generative AI and Large Language Models is changing this. Legacy, closed-text surveys with quick, predefined options that lack any richness are becoming a thing of the past. This is already the case at Streetbees, as we have always believed in letting our users respond in their own words to capture emerging behaviours and nuanced attitudes. However, before now, there was no opportunity to follow up on consumers' responses to probe further into what they had shared.
Thanks to Generative AI, tools can now capture and analyse open-text responses at scale and can optimise questions in real-time to probe for additional detail, exactly like an ethnographic researcher.
For example, if a consumer provides a high-level answer on one of our new SBX surveys, the tool will follow up with a question that specifically aims to delve deeper into what has been shared. This means that each individual user follows a unique conversation path, something that has only become possible at scale through Generative AI.
This real-time optimisation is hugely effective at capturing the ‘why’ and ‘how’ beyond the ‘what’ to ensure the insights are granular and actionable and is something we’re no doubt going to see more of in the consumer insights space.
Insights will become instant.
Would you prefer to spend less time putting your consumer insights into decks and more time putting them into action? It’s no secret that transforming raw data into strategic materials can be time-consuming and resource-intensive.
After conducting extensive tests to evaluate the capabilities of Generative AI in areas like analysis and reporting, we’ve found that the technology can in fact do a lot of this heavy lifting for you. So you can spend less time sifting through data tables and creating charts and more time on delivering and leveraging the strategic insights.
SBX, for example, uses our proprietary technology and taxonomies to automatically classify, quantify and analyse the open-text data it collects and translate this into a tailored report, designed specifically to answer your business question. This process, from brief generation to survey to report, all happens in less than 48 hours.
Here's what we've learnt from creating this new reporting functionality:
- Generative AI is very effective at processing large amounts of data at speed and can greatly increase the quality of the data.
- When it comes to reporting, it can automate the creation of charts, the cross tabulation of data and the discovery of insights.
- While it can help you get the insights you need, human review is still needed for now to translate those insights into actionable recommendations based on brand or industry knowledge.
- This technology is set to significantly speed up delivery times, however it is best suited to discrete, specific business questions, rather than big ‘open-ended’ strategic pieces.
Here are some considerations to keep in mind
Over the course of running these experiments and using our new SBX tool with clients, we’ve learnt a few things about how to get the most from Gen AI during the research process.
- The quality of the initial input still matters. Just as the initial brief is critical to the success of a larger, more strategic project, the business question is critical to the success of a smaller, ad-hoc project. In short, the more effort you put in at the beginning to work with the brief assistant to make the question targeted and specific, the better the output you can expect from the tool.
- The human touch is still key for larger strategic projects. While the technology is allowing innovation in the market research space to come on leaps and bounds, it still has limitations. Therefore, it’s best suited for certain things. It’s not yet at the stage where it can handle large, strategic projects with huge sample sizes and complex analyses. However, it’s really useful for answering specific business questions that require fast, nuanced results. Questions like “How do consumers choose between laundry detergents?” or “Why are sales of our new light beer so lukewarm?”.
- Data security is essential. You must ensure that any tools you share your data with are secure. For example, SBX leverages the power of GPT through an API but it uses only proprietary data to inform its outputs and it does not share any data with GPT or the open web. In contrast, if you’re using a tool that’s essentially a ‘wrapper’ on top of ChatGPT, you must be aware that the information you provide could be shared.
If you'd like to learn more about the SBX Insights Co-Pilot and see the tool in action, you can find more information on our website www.streetbees.com/SBX
