What’s the right mix of artificial and human intelligence when analyzing marketing research data? That’s a question that has weighed on my mind recently. My pedigree as an analyst comes from the discipline of rhetoric – the artful, humanistic pursuit of connecting language and persuasion. Because I’m wired and trained this way, any time I have an artificial intelligence opportunity before me, I’m always pulled toward adding as much of the human element as possible.
Fortunately, a recent engagement we did for a biopharma client affirmed for me the power of AI and the proper layering of human intelligence. This client is facing an increasingly common competitive environment in oncology: A decade-old entrenched treatment is ready to be unseated by a next-gen class, with members of the next-gen class (including the client’s brand) struggling both against each other and against the entrenchment. The client recently completed white-paper sales collateral research – a straightforward and tidy qualitative project that yielded actionable insights about page ordering, headline clarity, and which annotations are needed on the graphs and tables. However, when the brand leadership needed to understand deeper questions about their strategy to differentiate within the class, this final report was silent. Worse yet, the clock was ticking, as the client had to prepare materials for a major oncology conference in less than two weeks.
This client knows the value of retaining video recordings and verbatim transcripts from their qualitative research, so we had two very good places to start. We immediately brought the transcripts into S+R AQuA™, our advanced qualitative analytics application, and set our human analyst to read them and watch selected video excerpts. Simultaneously, we took those same transcripts from AQuA™ and loaded them for a granular analysis in Luminoso (www.luminoso.com), our text analytics AI platform partner.
Our talented human analyst quickly saw the parallels between this brand and another we’ve supported for many years. As he read the transcripts and watched oncologists struggle to defend their habitual prescribing behavior, he identified some of the heuristics and emotional needs at play and shared how many of these can be successfully overcome.
The AI analysis was structured to model a benefit ladder, anchored by the clinical data presented in the research stimuli. Luminoso’s instant recognition of the textual concepts related to this data allowed us to see which parts of the benefit ladder were solid and which were shaky. Even more insightful, the cluster of related concepts suggested that several benefit ladders were in play. Only one of these ladders reached a favorable outcome. Our AI-based analysis not only showed which words/phrases around the clinical data steer oncologists in the right direction, it also showed specific language to avoid.
For me, the best part of this project was watching the human and AI analysis come together to reach synergistic recommendations. The AI analysis was very textual and kept us grounded in the actual words of oncologists; the human analysis was very contextual and kept our eyes on the broad competitive environment (not just one sales aid project). We gave the client clear recommendations about how to share their clinical data: What to highlight, what to avoid, and how to get conversations back on track.
The whole client team was very pleased: Robust learnings with market-ready recommendations on a one-week timeline by leveraging research data they already had.
It was the perfect mix of artificial and human intelligence – and a really good week for our clients!
Shapiro+Raj
S+R is a research and strategy firm that uses social and behavioral sciences to solve the toughest business and marketing challenges. Our next-gen methods dig deep to unlock market-ready insights. Then our brand planners turn these into strategic marketplace actions that create brand evolution and innovation; customer experiences and loyalty; and new platforms for growth.
