Redefining the Analyst Role: From Data Crunching to Strategic Interpretation in an AI-Enabled Insights Function

By Caplena

• article

• AI
• Artificial Intelligence
• Automated Reporting
• Competitor Analysis
• Customer Experience (CX) Feedback
• Feedback Analytics
• Data Analytics
• Data Visualisation
• Machine Learning
• NLP (Natural Language Processing)
• NPS (Net Promoter Score)
• Sentiment Analysis
• Social Media Listening/Intelligence
• Text Analytics
• Trend Analytics
• Trend Monitoring

The question facing research and insights professionals is no longer whether artificial intelligence will change their work – it already has. The more consequential question is how roles will evolve as AI automates tasks that once consumed the majority of analyst time. Will automation eliminate positions, or will it elevate them? The answer emerging from organisations implementing advanced analytics capabilities suggests a fundamental shift: from data processing to strategic interpretation, from crunching numbers to making sense of what numbers mean.

This evolution is visible in how Kia Europe’s customer experience team, a Caplena customer, operates today compared to just three years ago. The transformation offers a window into how analyst roles are being redefined across the insights industry, what new skills matter most, and why organisations that manage this transition well are discovering that AI makes human expertise more valuable rather than less.

This article covers part of the case study “Putting Feedback to Work Across Kia Europe” presented at the Insights to Action Summit in October 2025. Rewatch the webinar here:

The Traditional Division of Labour

Understanding the shift requires acknowledging how analyst time has traditionally been allocated. In most research operations, the majority of effort goes into what might be called “getting to the number” – the mechanical work of data processing, cleaning, coding, categorisation, and basic analysis. For organisations dealing with large volumes of unstructured feedback, coding open-ended responses alone can consume days or weeks per project.

This reality creates a problematic dynamic. Analysts spend most of their time on activities that, while necessary, do not leverage their highest-value capabilities. The work that should differentiate expert researchers – contextual interpretation, implication development, strategic recommendation formulation, stakeholder collaboration – gets compressed into whatever time remains after the data processing is complete. Often, that means inadequate time for the work that actually drives business impact.

The traditional model also creates capacity constraints. If coding 10,000 verbatim comments requires substantial manual effort, the organisation can only analyse so many projects simultaneously. Questions that could be answered with existing data go unaddressed simply because no one has time to do the analysis. Insights remain locked in unprocessed feedback, representing missed opportunities for improvement.

What AI Automates Well

Modern text analytics platforms powered by large language models demonstrate clear strengths in specific aspects of the analysis workflow. Topic categorisation – identifying that a comment discusses pricing, product quality, customer service, or any of dozens of other themes – can now be performed automatically with accuracy approaching human-level coding in many contexts. Sentiment analysis at the topic level (distinguishing between “I loved the customer service” and “the customer service was terrible”) similarly shows strong performance.

Translation across languages, once a significant bottleneck for multinational organisations, becomes nearly instantaneous. Preliminary summarisation of key themes across thousands of comments provides starting points for deeper investigation. Even sophisticated tasks like driver analysis – identifying which topics have the greatest statistical impact on overall satisfaction or NPS – can be largely automated.

The speed and scale advantages are dramatic. Tasks that once required days can be completed in minutes. Projects that would have been declined due to resource constraints become feasible. Questions that arose too late in decision cycles to inform outcomes can now be addressed while they still matter.

What Remains Distinctly Human

Despite these capabilities, critical aspects of insight generation remain firmly in the human domain. The most significant is judgment about what findings actually mean for the business. When analysis reveals that electric vehicle customers have significantly different handover expectations than traditional vehicle customers, AI can surface and quantify that pattern. What AI cannot do – at least not reliably with current technology – is determine whether the appropriate response is enhanced dealer training, redesigned handover processes, adjusted customer expectations through marketing, or some combination of all three.

This strategic interpretation requires contextual knowledge that extends well beyond the data being analysed. It requires understanding organisational capabilities and constraints, competitive positioning, resource availability, stakeholder priorities, and the subtle political dynamics that influence which recommendations will gain traction. It requires the ability to conduct the “sanity check” – the back-of-the-envelope assessment of whether a finding aligns with other things you know about the business or whether something might be methodologically amiss.

Human judgment also proves essential in guiding AI toward business-relevant analysis. While AI can identify patterns, humans must define what patterns to look for, how to structure topic hierarchies, what level of granularity serves the decision at hand, and how to balance exhaustive coverage against analytical parsimony. The quality of AI outputs depends heavily on the quality of human direction provided.

Finally, the work of translating insights into action remains stubbornly human. This includes crafting narratives that make findings comprehensible and compelling to stakeholders, facilitating cross-functional discussions about implications, negotiating implementation priorities among competing interests, and maintaining momentum through the difficult work of organisational change. These are capabilities rooted in communication, empathy, political acumen, and relationship management—areas where AI offers little substitute for human skill.

The Emerging Analyst Profile

The role evolving in AI-enabled insights functions looks different from traditional analyst positions. Time allocation shifts dramatically. Where analysts once spent perhaps 80% of their time on data processing and 20% on interpretation and stakeholder engagement, those ratios are inverting. In organisations like Kia Europe, analysts spend comparatively little time on mechanical coding and substantially more time on making sense of results, determining appropriate actions, and ensuring insights actually influence decisions.

This shift elevates certain skills while diminishing the importance of others. Technical proficiency in AI tools matters, but not in the sense of requiring programming or data science expertise. Rather, analysts need operational familiarity: understanding how to configure systems effectively, recognising when outputs require refinement, and knowing which AI capabilities to apply to which questions. The skill is more conductor than engineer – orchestrating AI capabilities toward useful outcomes rather than building the capabilities themselves.

Quality assurance becomes a core competency. As AI handles initial analysis, human oversight focuses on validation: reviewing categorisation samples to ensure the system understood the task correctly, checking that statistical relationships make logical sense, and confirming that edge cases receive appropriate handling. This is detective work more than production work – looking for problems rather than executing processes.

Contextual interpretation and strategic thinking move to centre stage. With AI providing quantified patterns, analyst value concentrates on the “so what” questions. Why are customers expressing these concerns? What are the feasible response options? Which interventions offer the best return on investment? How do findings connect to broader strategic initiatives? This work requires business acumen that extends well beyond research methodology.

Stakeholder engagement and communication skills become differentiators. The best insights delivered poorly have less impact than modest insights delivered compellingly. As AI democratizes access to data and basic analysis, the human analyst’s role increasingly involves curating insights, building narratives, facilitating productive discussions, and ensuring that understanding translates into action.

Implications for Career Development

For insights professionals navigating this transition, several priorities emerge. First, develop comfort with AI tools not as mysterious black boxes but as capable assistants with known strengths and limitations. The goal is not to become a technical expert but to become an effective user – understanding what these tools can and cannot do, how to direct them effectively, and when to trust versus question their outputs.

Second, invest in business acumen beyond research methodology. Understanding your organisation’s strategy, competitive dynamics, operational constraints, and decision-making processes makes the difference between analysis that is technically sound but practically irrelevant and analysis that drives genuine impact. Seek opportunities to engage with strategy, operations, and decision-making beyond the insights function.

Third, cultivate storytelling and stakeholder management capabilities. Practice translating complex findings into clear narratives. Develop the political awareness to understand what different stakeholders care about and how to frame insights in ways that resonate. Build relationships that create pathways for insights to influence decisions rather than languishing in reports.

Fourth, embrace the shift from execution to orchestration. The future insights professional is less about personally conducting every aspect of analysis and more about designing analytical approaches, directing AI capabilities, quality-checking outputs, and ensuring that the overall process yields trustworthy, actionable findings.

The Augmentation Perspective

The most useful frame for understanding this evolution may be augmentation rather than automation. AI augments analyst capabilities, handling the repetitive, time-consuming aspects of analysis and thereby creating capacity for analysts to focus on higher-value work. The technology does not eliminate the need for human expertise; it changes where and how that expertise gets applied.

Organisations that navigate this transition well, like Kia Europe, find that AI makes their insights functions more valuable rather than less. Analysts freed from data processing bottlenecks can address more questions, engage more deeply with stakeholders, and provide more strategic value. The function shifts from being primarily reactive and reporting-oriented to being proactive and advisory.

For insights professionals, this represents an opportunity more than a threat, but only for those who adapt their skills and self-conception accordingly. The analysts who thrive will be those who embrace the shift from technician to strategist, from number-cruncher to business advisor, from isolated specialist to collaborative partner. The work changes, but for those who evolve with it, the impact and organisational value increase substantially.

This article covers part of the case study “Putting Feedback to Work Across Kia Europe” presented at the Insights to Action Summit in October 2025. Rewatch the webinar here:

Author

Learn more about