Author: Cameron Maskell
Research and Insights Review November: AI Guardrails and Segmentation 2.0
Research and Insights Review November: AI Guardrails and Segmentation 2.0
December 2, 2025
Q4 pressure is peaking, 2026 plans are crystallising, and insight teams are being asked to do more – faster. From AI guardrails to attitudinal segmentation 2.0, here are four big themes shaping the insight space this November.
🤖 AI Guardrails Become a Strategic Priority
As AI tools become embedded in research workflows, brands are setting clearer boundaries, not just around data privacy, but around decision-making accountability.
Heineken introduced internal AI usage guidelines for research, setting out when AI should support, augment, or be left out entirely. One key rule: AI can summarise but never generate final recommendations without human review. This move followed a misinterpretation of sentiment in early AI-led social listening, which skewed tone-of-voice guidance.
It appears that we are now moving past the AI hype phase, and now many are looking to build smart, ethical infrastructure around it.
🛒 Segmentation Gets Real-Time and Behaviour-Led
Static attitudinal segments are being challenged by live behavioural signals…especially in e-commerce, where consumer needs shift hour by hour.
M&S moved beyond traditional personas to trial “real-time segment activation,” combining first-party data, page journeys, and session recency to adapt homepage content and email subject lines dynamically. A single user could shift between three need states in a day, and personalisation kept up. The result? A 12% uplift in click-through and stronger basket builds.
The new segmentation mindset is less “who are you?” and more “what do you need right now?”
📱 Gen Alpha Enters the Research Chat
Brands are starting to treat Gen Alpha (those born after 2010) as a research-worthy audience in their own right, not just as “kids of millennials.”
LEGO launched a co-creation community for 10–13-year-olds, with fully COPPA-compliant guardrails, moderated online spaces, and a hybrid of drawing, voice, and build-based tasks. Insights from the community fed into 2026 packaging and digital gameplay strategies, showing this age group’s strong lean toward environmental storytelling and mixed-reality play.
Gen Alpha doesn’t want to be marketed to… they want to shape what’s being built.
🧾 ‘Value’ Gets Redefined… Again!
As economic anxiety drags on, the consumer definition of valuecontinues to evolve, and brands are updating how they measure it.
ASDA conducted a mixed-methods study combining digital receipts, time-use diaries, and emotional check-ins during weekly shops. Findings showed that value isn’t just about price, it’s about reducing mental load (“can I get everything here without having to think?”). This insight has informed store layout updates, pre-bundled meal solutions, and smarter basket-building prompts.
The value equation in 2025 = price + time + headspace. Are you tracking all three?
Research and Insights Review October 2025: Rise of Re-commerce and “Second Screen” Ethnography
Research and Insights Review October 2025: Rise of Re-commerce and “Second Screen” Ethnography
November 4, 2025
By Cameron Maskell
Halloween has come and gone, campaign season is peaking, and insight teams are digging into end-of-year trends and 2026 planning. So, what’s been brewing across the research landscape this October? Here are four developments that stood out.
💬 AI Moderators Take on Qual at Scale
As online qual grows, AI is stepping in… not to replace researchers, but to support smarter, more scalable moderation and follow-up.
Unilever recently piloted an AI moderator for a week-long online community exploring teen skincare habits across five markets. The AI tool used natural language prompts to probe participants in real time, flag contradictions, and tailor follow-up questions. Human moderators oversaw and refined key threads, cutting analysis time by 40% and improving depth on niche topics.
Takeaway? The human touch still matters, but with AI co-pilots, qual is getting faster, deeper, and more flexible.
🧭 “Second Screen” Ethnography Captures Real-Life Context
Forget scheduled diaries! Second-screen ethnography is giving researchers real-world behavioural insight, in real time.
Just Eat worked with a mobile ethnography partner to capture how people really decide what to order, not just what they say they do. Participants screen-recorded their scroll behaviour on food apps, paired with voice memos explaining their thought process (e.g. “I was craving Chinese, but then the delivery time made me change my mind…”). The data led to UX tweaks and new delivery-time filters now rolling out across the app.
Takeaway? When you see the decision unfold live, the ‘why’ behind the ‘what’ becomes clearer than ever.
📦 Research Supports the Rise of Re-commerce
As resale, rental, and return models surge, insight teams are guiding how brands tap into the growing re-commerce economy.
John Lewis launched a second-hand furniture pilot and partnered with an insight consultancy to explore perceptions of pre-owned value. Through on-site pop-ups and digital ethnography, they found that product condition mattered less than storytelling. Consumers valued transparency (e.g. “tell me who used this before”) and sustainability framing more than price alone. The findings led to a test of branded storytelling tags on pre-loved listings.
Takeaway? Research and insights are showing that resale value is about reputation, relevance, and re-framing what ‘used’ really means.
📊 Insight Teams Embrace “Test-and-Learn” Budgeting
As agility becomes a C-suite mantra, insight teams are rethinking how they use and justify budget.
Sky ran a “test-and-learn” budget structure across its research road map this quarter, allocating 20% of spend for fast-turn experiments tied to live business questions. Quick audience pulses and message tests fed directly into weekly creative decisions for an entertainment campaign. The result? Shorter time-to-insight, fewer bottlenecks, and more buy-in from commercial leads.
Takeaway? Insight is becoming a testing engine for smarter business decisions.