In most insights functions, AI has been used for efficiency: tidying open-ended responses, clustering verbatims, or generating quick summaries. Helpful, but peripheral. The real test is whether AI changes how the team thinks and acts. 

In brand tracking, for instance, AI can now flag anomalies in sentiment, detect micro-segments, and generate rival hypotheses about shifting behavior. Analysts evaluate those hypotheses, bring in context, and decide how to advise leadership. 

That partnership – breadth from AI, depth from humans – is where value emerges. McKinsey’s State of AI 2025 found that firms seeing the highest returns had redesigned workflows around human-in-the-loop models, with clear governance on what’s automated, what requires judgment, and who is accountable. 

Procter & Gamble has used AI-driven simulations to evaluate product concepts before a prototype is built, cutting development time by 22%. Mars Wrigley has tested AI to generate new flavor combinations, refining them through human sensory panels. Unilever experiments with AI for advertising copy, with human teams fine-tuning tone and cultural fit. In each case, AI expands the option set, humans deliver the verdict. 

How to build confidence: 

• Choose one regular report (e.g., brand tracking or campaign monitoring) and let AI do the routine work while your team reviews the results. 

• Train analysts to interrogate AI outputs as they would a junior colleague’s work. 

• Document key changes and decisions and share the learning across the team. 

• Evaluate quality, not just speed.  

• Let each team member try AI on real projects to build confidence and skills. 

When Unilever shifted to a social-first approach, its People Data Centers stopped producing dashboards and started producing storylines: this is what people are saying, here’s why it matters, here’s the move to make. Aaron Rajan, VP for Consumer Technology, called it an “always-on approach to insights.”  

As brand discovery moves onto social platforms, the value of insight now lies in real-time storytelling, framing cultural signals fast enough that marketing and product teams can act. The most embedded insights teams run weekly synthesis sprints, bringing analysts, strategists, and business leads together to turn signals into 60-second narratives that shape decisions. 

Airbnb trains staff to convert consumer feedback into design principles that product teams can act on immediately. Nike hires strategists with editorial backgrounds to translate consumer truths into brand narratives. Both treat narrative generation as a technical skill, not an afterthought. 

How to build confidence: 

• Partner with product and marketing teams to make insights part of everyday decisions. 

• Train for facilitation and influence as core competencies. 

• Operate from one shared, live dashboard instead of static decks. 

• Track impact by where insights surface (in roadmaps, strategies, and briefs) not by volume of reports. 

Insight that moves decisions is not about data volume. It’s about narrative velocity. 

At Netflix, nothing launches without a test. In 2024 the company ran more than 70 controlled experiments each quarter, tweaking everything from thumbnail images to payment flows. Most ideas failed quietly. A few scaled. All left a data trail that sharpened the next bet. 

Adobe applies the same discipline in marketing. Its Experience Cloud teams run structured tests on campaign creative, feeding results back into product cycles. Capital One embeds insights staff and data scientists in product squads with the authority to halt a feature mid-flight if results disappoint – a practice that has saved millions in avoided spend. 

For many insights teams, the challenge isn’t understanding experimentation but building the infrastructure for real-time operations: rolling trackers, automated alerts, and governance that gives analysts the mandate to act. The World Economic Forum lists adaptability and experimentation among the fastest-rising workplace skills.  

How to build confidence: 

• Turn measurement from a one-off project into a live, ongoing system that updates as new data comes in. 

• Convert periodic data into daily data with event triggers. 

• Build a shared backlog of tests, ranked by commercial value. 

• Archive learnings in a searchable database accessible across functions. 

• Give analysts stop / go authority within predefined guardrails. 

• Capture where data prevents bad decisions, not just where it leads to good ones. 

Structured co-creation now functions less as research, more as infrastructure. PepsiCo’s Lay’s ‘Do Us A Flavor’ campaign showed how to turn mass input into marketable products. Patagonia’s expansion of its Worn Wear program demonstrated how early consumer involvement creates advocacy, not just feedback. 

Advanced teams use tiered participation models: micro-feedback loops for creative testing, advisory groups for concept development, and high-trust communities for innovation input. ESOMAR data links this structure to faster product cycles and higher satisfaction scores.  

How to build confidence: 

• Build an always on connection to consumers. 

• Close the loop: show contributors where their input landed. 

• Define what’s open to influence (packaging, messaging) and what remains internal (pricing, supply chain). 

• Track throughput: how many co-created ideas make it to market and their commercial outcomes. 

• Where possible, reward contributors meaningfully (e.g. early access, recognition, or co-branding). 

When consumers can see their fingerprints on the outcome, co-creation shifts from novelty to capability. 

In many companies, insights still live down the hall, summoned to “provide data” when a project is already underway. By the time findings reach decision-makers, context is lost and the moment to influence has passed. 

Some firms have restructured to avoid that lag. Coca-Cola, for example, launched its internal Lens platform in 2025 to give marketers, sales, and R&D staff access to live consumer signals. Insights staff were embedded alongside those functions, helping teams interpret shifts as they happened. “Democratizing insights” was the headline, but the bigger shift was cultural: the analyst had a seat in the same room as the decision-makers. 

Amazon has long required teams to start with a mock press release written from the consumer’s perspective. Embedded insights staff are critical to this process, ensuring the “working backwards” model reflects genuine consumer needs rather than internal assumptions. 

How to build confidence: 

• Place analysts within business teams and give them dotted reporting lines. 

• Define shared KPIs with other teams to ensure mutual accountability. 

• Provide unified, live dashboards accessible to all teams. 

• Train for facilitation: the analyst who guides discussion shapes decisions. 

• Audit how often insights appear in strategy and product documents. 

Embedding isn’t an org chart tweak. It’s a credibility upgrade. 

Campaigns travel fast. Missteps travel faster. 

HSBC’s “Assume Nothing” campaign from 2019 remains a cautionary tale: praised in some markets, misunderstood in others. The issue wasn’t the idea but the resonance of the language. Similar stumbles keep happening, often because teams rely on global averages without testing for local nuance. 

Global brands can’t afford that risk. Unilever’s People Data Centers scan conversations in more than 30 markets and feed cultural insights into campaign development. Netflix uses regional viewing data but pairs it with local cultural consultants before commissioning new content, ensuring that resonance is explained, not just observed. 

The insight function often serves as the first defense against cultural blind spots. That role requires more than translation. It demands fluency: the ability to interpret behavior within a cultural frame and communicate that risk to decision-makers who may not see it. 

How to build confidence: 

• Hire analysts with market and cultural experiences. 

• Maintain regional advisor networks to review unexpected findings. 

• Pair quantitative dashboards with cultural interpretation notes. 

• Test creatives in key markets before rollout. 

• Keep a library of past cultural mistakes for training and development. 

Insight teams are increasingly the first line of cultural defense, provided they have the fluency to interpret, not just translate. 

Executives no longer want “interesting.” They want impact. Bain & Company’s 2025 research showed that firms tying consumer signals directly to revenue or margin were twice as likely to outperform peers. Deloitte’s 2025 CMO Survey reached a similar conclusion: without a link to business metrics, insight rarely influences investment. 

The best teams have codified the “money slide” – a single chart that connects consumer movement to financial effect. Mondelez links brand-tracking shifts to sales lift models, guiding media reallocation. Several global banks now require an ROI estimate, even directional, in every research output. 

How to build confidence: 

• Begin each project with the business outcome and KPI it must influence. 

• Standardize the money slide format: one page, one metric, one implication. 

• Build lightweight models using internal data to prioritize efforts. 

• Track actual financial outcomes against predicted ones. 

• Evaluate teams on influence and accuracy, not output volumes. 

Insight without commercial translation is noise. The ability to connect signal to value is what earns a seat at the table. 

LEGO’s Ideas platform began as an experiment and evolved into infrastructure; a system that converts participation into innovation. The same principle applies to insights. 

By 2028, the divide won’t be between teams with more data and those with less. It will be between teams that engineered repeatable, accountable insight systems and those that stayed episodic.  

At Delineate, we help insights teams do this every day. Our always-on tracking gives you live feedback from real people, so you can see what’s changing as it happens and make better decisions faster. 

Download ‘The 7-Skill Team Readiness Checklist’