Four things to demand from AI-powered consumer insights

THE 8TH GENERATION OF THE SENSIA ENGINE · FOR BRAND, MARKETING, INNOVATION AND INSIGHTS TEAMS

Why do some consumers stop choosing us? What makes competing offers more compelling? Where are we strong, and where could we establish a clearer difference?

For an established brand, these questions require a view of the market: understanding the expectations, experiences and alternatives against which consumers judge its value.

AI can produce a persuasive summary in minutes. To make it useful to your team, you need to understand what deserves attention, why it matters and what evidence supports it.

Sensia analyses consumer reviews across a category, identifies the main findings and lets teams explore them back to the source reviews. Here are four things to demand from this kind of analysis, and how the 8th generation of our engine delivers them.

Understand what makes consumers choose a brand—and what can weaken that preference

Consider a face-serum brand looking to defend its position. Its questions concern both the reasons to stay and the reasons to look elsewhere: what justifies its value to consumers, and where does that conviction weaken?

A Sensia analysis of reviews in France and the UK shows how this understanding can be developed across a category.

Example drawn from Sensia’s analysis of face serums in France and the UK. It examines drivers of preference and reasons for giving up on a product; it does not directly measure customer switching between brands.

Why select the evidence before analysing it?

A category can contain hundreds of thousands of reviews—more than a general-purpose AI model can examine in a single pass. And providing more text does not guarantee a better analysis: across very long inputs, models can struggle to use information consistently or overlook important details.

Sensia first interprets the reviews, then selects relevant evidence according to the market, products and question the brand wants to explore. Understanding disappointment, comparing competitors and investigating a positioning opportunity do not require exactly the same material.

Our specialised models provide the foundation for this selection. It prepares the evidence on which the AI develops its analysis, while preserving the link to source reviews. Teams can then explore a brand, topic or comparison further through conversation.

This preparation delivers its value through an analysis your team can build on. Four requirements help you assess it.

1. Findings you can build on

Teams need to identify what matters most: reasons to choose an offer, experiences that weaken its value and opportunities to differentiate. The order of priority matters as much as the list of topics. It helps determine where to focus the brand’s attention, resources and effort.

To test this understanding, we compared five Sensia analyses as the AI examined substantially larger volumes of reviews. On two parts of the analyses, we also used a more powerful reasoning model.

Sensia’s focused selections already supported the main strategic findings and their order of priority, confirmed by a substantially broader reading. Deeper investigation added detail and nuance while retaining the main strategic directions.

For a brand, this provides a solid foundation for supporting a direction, examining a vulnerability and investigating important questions. Sensia’s value lies in identifying priorities within a body of feedback far larger than a team can examine directly.

2. Coverage relevant to your question

To understand its position, a brand needs to see itself among the offers consumers evaluate and compare. The analysis should represent the products relevant to the subject and allow you to investigate those that matter to your project.

A launch or niche product may deserve particular attention even when it has relatively few reviews.

We checked one aspect of this coverage: did Sensia’s selections leave out products accounting for a substantial share of relevant reviews?

Across the four questions studied, selections of around 1,500 reviews included products accounting for 98.1%–99.5% of total relevant review volume. The products carrying almost all that volume were therefore represented in the samples.

This measure complements the comparison of findings: Sensia prepares a focused selection representing the products that carry almost all relevant review volume for the questions studied.

Your team can then explore a brand, product, market or topic in greater detail according to the project’s needs.

3. Evidence you can trace

Findings need to be discussed with leadership, agencies and colleagues. Being able to examine the supporting reviews and products helps people understand a finding and assess its relevance to the brand.

You should be able to move from a finding to its supporting reviews, then back to their sources. The quotations presented in Sensia come from real reviews: they are neither invented nor completed by AI.

In a Sensia analysis, teams can examine the source reviews supporting the findings.

Figures should also have a verifiable basis. Sensia calculates its indicators from reviews analysed by its specialised models. The synthesis draws on that data; a figure intended as a quantitative reference should come from the underlying calculation.

This allows teams to present a direction alongside the experiences supporting it. Consumer language can also help the brand or agency express a benefit in terms people recognise.

4. Clear directions to explore

An analysis becomes useful when it helps a brand determine what to defend, improve or explore. That may mean communicating a promise more clearly, addressing a frustrating experience, strengthening a positioning territory or investigating an opportunity competitors have left underserved.

In the serum example, a brand can examine whether its value rests on the benefits consumers expect in each market, what weakens preference and which differences it can credibly defend against alternatives.

These findings can inform positioning, communication, usage guidance, creative work or the evolution of the offer.

For creative teams, the words, situations and tensions consumers describe bring substance to a brief: what people expect, what disappoints them and what makes a difference credible.

The exploration can continue through conversation. Which brands are associated with this expectation? What are they recognised for? Where does disappointment emerge? Do the same perceptions appear in another market.

Explore more questions as your projects evolve

Keep a view of the market as your projects evolve

Major studies and tracking programmes provide the foundations of market understanding. Between those milestones, new questions arise: an alternative becoming more attractive, a promise losing its appeal or a brand strength that could be used more effectively.

Sensia lets teams return to consumer voices to examine these questions, with a category-wide view, accessible source reviews and the ability to investigate further through conversation.

What would your team like to understand about its place in the market?

About the benchmark: five production analyses, comprising twenty sections, across three reference datasets in skincare and pet food. Analyses were regenerated using six review volumes. Two sections were also compared with a high-end reasoning configuration. Selection reach was measured for four questions. Main findings and their order of priority were assessed by comparative reading. The conditions, measurements and full results are available in the benchmark report. Tests conducted in October 2026.