F'innews: Beach Read
June 30th: Our new website, Digital twins POV, The coping paradox for women with ADHD (~4 minute read)
A big thank you to the New York research community for gathering once again! At our latest meetup, we took up the question of “researcher’s intuition” — whether it shapes our work, and whether handing more of our practice to AI will dull that instinct or sharpen it. What struck us most was how inventively we’re already bending these tools to our own ends: agents grinding through datasets overnight, generative databases queried for strategy and the decisions that follow. No two setups looked alike.
As more of the heavy lifting shifts to AI, qualitative research is doing what it always has — keeping our instincts sharp and us close to the people we study. The room was generous and unguarded, candid about what’s working and what isn’t. Thank you all for showing up, again and again, and making these evenings what they are.
We have a new website!
A client once described us as “the best-kept secret in the industry.” It’s both a lovely compliment but also a call to action. So we built a new site that better reflects what we actually do and where we’re headed in 2026.
F’inn helps companies understand what people want and, more importantly, why. We design custom research, work quickly, and deliver clear recommendations so our clients can make confident decisions. The new site lays out our focus areas more clearly, including our ongoing AI development and experimentation, which we’ll continue sharing here as it evolves.
Thanks to SLANT for building it with us, and to our own Avalon for bringing the brand identity to life!
Take a look for yourself: finn-group.com
Why Digital Twins Don’t Work (Yet)
At the F’inn AI Lab, we promise to share our successes as well as failures, so here’s a failure for you. We’re getting more and more pitches for “digital twins”, which are AI-generated respondents designed to stand in for real people in surveys. They promise faster, cheaper studies and access to audiences that are hard to reach. Sounds great, but can you trust them?
To find out, we put digital twins through rigorous, validated-against-human testing to see if they can be trusted to replicate real human data. They look convincing on the surface but fall apart when you dig just a bit deeper, with two (2) main fail states:
They repeat one ‘reasonable’ answer, representing consensus, not true human points of view.
They fall back to stereotypical biases so that distinctions across audiences are not accurate and marginalized voices are not heard.
Our view is simple. AI belongs in research, but not as a substitute for the humans you’re trying to understand. Its value is in analysis, like finding the psychological drivers underneath real human data and running complex advanced analytics in minutes instead of days. We believe a firm that uses AI to manufacture respondents has stopped being in the business of understanding people. We’ll keep experimenting as the technology evolves, and if it ever reliably predicts what real humans would say, we’ll be among the first to tell you. Right now (June 2026), digital twins introduce noise and invalidity into the data set and will create new risk instead of building confidence in decision making.
If you want the full argument, including what happened when we tried every workaround to make digital twins succeed, our full POV is linked here. Or shoot us a DM - we’d love to hear from you on how you’re trying to incorporate digital twins.
A few years ago I presented some of F’inn’s mental health research at the Corporate Researchers Conference, where I met Justine Bassman from Understood.org. She was working on a study focused on women with ADHD and wanted to dig deeper into our data. We shared our dataset with her team, and they recently published their findings.
The paper is called The Coping Paradox and it found that women with ADHD use more coping strategies than women without ADHD, yet their mental health outcomes are significantly worse. Only 20% of women with ADHD rate their mental health as very good or excellent, compared to 51% of women without ADHD. More effort with worse outcomes.
The standard self-care advice (eating well, exercising, meditating, maintaining sleep routines) is actually associated with worse outcomes for women with ADHD. Part of the reason is that women with ADHD already try these behaviors at higher rates, but executive function deficits make consistent follow-through hard. When the routines inevitably slip, guilt and shame move in, compounding the problem rather than easing it.
What does work is more interesting. Journaling, music, cooking, hobbies, and gardening showed significant protective effects for women with ADHD, an effect that didn’t appear for women without ADHD. Understood.org explains:
“They provide engagement through novelty. They don’t require sustained self-regulation. And they produce visible output. The data suggest they warrant consideration as targeted supports, not merely generic leisure.”
Therapy also matters more for this group than the general population, but the gap is in outcomes. Women with ADHD enter mental health treatment at nearly the same rate as women without ADHD. The implication is that care needs to be designed differently for ADHD brains, not just made more available.
It’s a well-researched paper with specific, actionable recommendations for clinicians and practitioners, and we’re proud to be a small part in it!
Worth a read: The Coping Paradox
