We welcome Alex Charlton, True Stories, to this Pitch Perfect.

An LLM is a plausibility engine, not a truth engine. It hands you the most convincing-sounding answer, not necessarily the true one. Point it at your qual transcripts and it works from whatever it happens to surface – there is no stratified sampling across your dataset, no language balancing stopping the Spanish groups being drowned out by the English ones. What comes back is the middle of the distribution, fluently written, with the odd, difficult, interesting participant quietly gone.

Even qual tools that do promise stratified sampling, simply place similar quotes in piles, then label the piles. This is data summarisation, not the analysis our clients pay us for. And then there’s the question of how we combine our own (human) intelligence with the artificial intelligence we’re working with. This is a minefield in and of itself.

None of this can be fixed by better prompting. The fix has to live in the architecture around the model.

By joining this session, expect a working demonstration, some lessons learnt along the way, and a sharper set of questions to put to any vendor who tells you their tool “understands” your data.

Alex Charlton

Alex Charlton is founder and Managing Director of True Stories. He specialises in qual research and previously founded two London-based insight agencies, Essential Research and Craft. Frustrated that the AI tools arriving in research were built by engineers who had never moderated a group, he spent nearly two years building the alternative himself – writing the platform alongside AI rather than briefing a development team, so that every architectural decision was made by someone who knows what good qual research looks like. True Stories launched in January 2026 and is now used by insight teams across four continents. He works from Medellín, Colombia.

True Stories

True Stories is an AI-powered analysis platform for qualitative research. It was built by a career qual researcher, and the design assumption throughout is that the researcher stays in charge: the platform does the reading, the retrieval and the structured analysis, and you lead the analysis, assisted by AI.

Where other AI tools summarise data, True Stories analyses: drawing on 200+ marketing, consumer psychology and behavioural science frameworks to build a bespoke analysis pathway for each project. Analyses draw evenly across your data rather than sampling whatever the model finds most quotable, work in your participants’ original language without translating the nuance out, and carry confidence ratings. You can interrogate any output conversationally, push on it, then fold the conversation back into a revised analysis. Once analysis is complete, your project reports are generated automatically and include: a written mini-report, a story-driven debrief narrative, and an editable PowerPoint deck.

True Stories: Human Intelligence. Amplified.

www.truestories.earth

Register for this event here