Published Jul 31, 2024
Renee Smith, EVP Innovation at Toluna
The insights industry is undergoing a transformation thanks to AI. At the Toluna Group, AI is at the core of our business. Our “AI Everywhere” strategy is based on our vision of the ways in which AI combined with human intelligence (HI) will transform the industry over the next 2-3 years.
Across the Toluna Group, we’re looking forward to continuing to enhance research practices as we apply AI, including Large Language Models (LLMs) and Toluna proprietary GPTs, in the areas of client insights, platforms and tools, and data quality.
There will always be a need for the types of qualitative and quantitative research we do today, especially for ground truth and validation studies. We expect adoption of new AI-based options to follow the typical technology adoption path, with early adopters leading the way. But we envision a faster rate of adoption than seen in the shift from offline to online research. Since AI-based approaches help increase speed to action, they will appeal to CMOs and insights leaders facing increased pressure for faster business growth.
Specifically, we envision new AI-based approaches across a wide range of research and insight activities occurring today.
Transformations in synthetic data – data generated artificially as a synthesis of information to mimic the statistical features of real data – are already underway. The Toluna Group offers two approaches to this:
1. Synthetic personas provide responses to questions in quantitative surveys or in qualitative discussions in real-time during fieldwork. The approach leverages Toluna’s 1st party profile data and other inputs from real humans to guide a LLM to generate a synthesis of knowledge consistent with that profile. This will launch later this year.
2. We use machine learning “synthesizer” algorithms to generate additional rows of synthetic records after data collection ends. This approach is already available.
Client benefits of using synthetic data include reduced days in-field thanks to synthetic personas that fill gaps and/or offer deeper multivariate insights from low incidence groups using either approach.
In a world of synthetic data, high-quality, up-to-date real data is key. Toluna QSphere, our proprietary quality framework, plays an important role in our strategy to produce high fidelity synthetic data. We build synthetic personas using 1st party data from our Toluna Influencer panels, which incorporate 79M+ individuals across the globe. The outputs from our post-field synthesizer algorithms learn data patterns from freshly and/or previously collected real data from projects with the Toluna Group companies or even other vendors (e.g., bring-your-own data).
One area of transformation that especially excites us is the future use of fluid, adaptive conversations (with real or synthetic panelists) for research instead of fixed, static questionnaires and discussion guides.
We are laying key building blocks for this transformation today. For instance, our upcoming QProbe launch in our Toluna Start platform uses conversational AI to gather additional information if respondent answers to open-ended questions are vague or unclear. As we move towards full-length insights conversations, Toluna Influencer panelists will benefit from a more engaging and natural experience. And the Toluna Group’s clients will gain deeper insights from the real-time, tailored capture of the “whys” behind rating scales or other structured answers – a beneficial blurring of the lines between quant and qual.
As seen in our vision, we expect widespread transformation, touching research design and end-to-end delivery practices from scoping to final recommendations. The Toluna Group is well-placed to help clients navigate these changes because transformation is in our DNA.
We started nearly 25 years ago as a panel company, moved to field and tab and then to full-service, while also developing and launching DIY QuickSurveys followed by our robust Toluna Start DIY platform with integrated sample sourcing. We set up a dedicated AI/ML Lab of software engineers and data scientists back in 2018 – earlier than most. And we were the first company in the industry to launch an insights solution that used only AI and HI – a copy testing solution called ACT Instant – and we did it in 2019.
We hope you share our vision, and we look forward to having you join us on this journey of transformation. All aboard!
Back to Blog