AI helps us to be more effective at every stage of a project’s journey but never takes the lead. We’re picky about what we use and when – quality is always paramount.
Get to the right proposal quicker
Design comprehensive surveys faster and script them more accurately
Clean data more efficiently and effectively
Analyse results faster and with a broader perspective
Create more engaging, stress-tested narratives
Communicate results more effectively and creatively

AI as a ‘critical friend’ in project design
Responding to briefs remains a fundamentally human process, but using secure GenAI to suggest alternative approaches helps us to challenge our thinking until we have landed on the best method.
AI-assisted questionnaire design
GenAI provides us a starter-for-ten on survey and question design, significantly speeding up the process and serving as a springboard for deeper human creativity.
AI survey databank
We are developing a databank of best-practice surveys for a guard-railed GenAI to draw from, improving its output and better aligning it with the style of Market Measures’ questionnaires.
Automatic scripting
We’ve developed a method of automatically scripting the bulk of surveys in Qualtrics from Word documents via a LLM – this saves a significant amount of time, allowing us to focus more on the thoughtful tasks of survey testing and optimisation.
Bot detection
Qualtrics uses AI to judge if a respondent is a bot or not. We use this as a first-pass flag for closer investigation – a human always has final say on what is real data.
Conversational open-ends
AI tools built into Qualtrics allow us to probe answers
like a qual researcher (but in an online setting) – this
gives us qual depth at quant scale.
AI-powered thematic coding
Qualtrics provides AI-driven analysis of open-ends to
give us the key themes and appropriate quotes, but
for more intensive work we can use Inca to fully code
and summarise responses using a human brief and
pre-authorised codeframe.
Stress-tested narratives
While AI is not ready to take the lead on storytelling, it
is helpful in testing narratives and providing alternative hypotheses to consider.
AI-generated voxpops
AI-powered video generation is rapidly developing,
and already allows us to create short clips of AIgenerated respondents reading out real verbatim
responses without the need to gather them in the field.
Visualising new products
Showing is always better than telling. Image-based AI visualisation allows us to demonstrate how new products and designs should look based on the data we collect in a secure way.
Synthetic data
Despite the hype, our own experimentation and reports from other agencies show synthetic data isn’t where it needs to be to replace humans, even for low-stakes projects – it simply isn’t reliable enough.
Analysis of structured data
Current AI (ChatGPT) analysis of data tables is topline at best. While individual questions can be competently summarised, unearthing wider themes, trends and cross-referencing are better left to humans.