Six readings of the same batch, one picture
Every batch goes through all six. Each analyzer asks one question of every record — not a sample — and returns a finding with the counts and the customer's own words underneath.
Sentiment
What do customers feel — and about what, exactly?
In a batch of 169 reviews, 155 came back negative, with an average score of −0.72. Billing was the loudest topic — not an impression, but counted, with the reviews that say so underneath.
Churn risk
Who is about to leave, and what is driving them out?
The same batch put 22 of 169 customers in the risk bands. Service, not price, was the driver — a distinction that changes what the retention call offers before it is made.
Customer journey
Where do journeys stall, and at which stage?
Across 50 claims calls, 34 ended without an outcome. The failures concentrated at claim assessment — one named stage to fix, not a vague sense that claims are slow.
Root cause
What is actually causing the complaints?
One cause recurred across the batch: high claim excess. The report carried the fixes callers suggested in their own words, so the remedy arrived with the diagnosis.
Archetypes
Who are these customers, behaviourally?
Cooperative claimants were the largest segment in the batch, and none of them sat in a churn band. That tells a retention team where effort is wasted as clearly as where it is needed.
Strategic intelligence
Given all of it, what should we do first?
The batch closed with 4 at-risk customers to contact before renewal, ranked by the strength of the evidence. Not a chart to interpret — a list to work through.
One page per analyzer, read in a morning
Each analyzer returns its finding at the top, the counts behind it, and the customer's own words underneath. The sentiment page goes to the exec meeting, the journey page to operations, and the strategic page becomes the week's call list. Nobody logs into a dashboard — the batch report is the meeting document.