Growth · Part of CSAT
Verbatim sentiment analysis
Verbatim sentiment analysis reads the open-text answers on your CSAT (Customer Satisfaction: a post-interaction score for how satisfied a customer was.) surveys and tags each one positive, neutral, or negative, alongside the top keywords customers use. A sensitive permission gate keeps this free-text layer separate from ordinary score reads, so only authorized roles see what customers actually wrote. The default analyzer is a heuristic English and French wordlist, and any tenant can opt in to a production model (AWS Comprehend) for higher-accuracy scoring.
For the operator
You turn it on per tenant, and once it is live every open-text response carries a sentiment tag and a short keyword list. A QA (Quality Assurance: the program that scores and reviews agent interactions.) lead with the sensitive verbatim permission opens a survey response and sees the positive, neutral, or negative label plus the words driving it, while a supervisor with only score access stays limited to the numeric CSAT read. You start on the built-in English and French wordlist analyzer, then switch to AWS Comprehend from your tenant settings when you want production-grade scoring. Keyword surfacing points you straight at the phrases customers repeat, so you skim a week of verbatims in minutes instead of reading each one. The sentiment tag and keywords sit next to the raw comment, so nothing is hidden behind a separate export.
Business impact
Because sentiment and keywords attach to each survey, you can report negative-comment trends per client and per LOB (Line of Business: a distinct client program or queue within an operation.) instead of one blended number, which is the view clients ask for in QBRs. When a client challenges a CSAT dip, you show the actual verbatims and the sentiment split behind it rather than defending an average. Coaches use the keyword patterns to target the exact behaviours customers complain about, so coaching time lands where it moves the score. The sensitive permission gate keeps customer free-text away from roles that only need the number, which protects the trust clients place in you with their end-customer data. And it runs inside FrontLine, so you get sentiment analysis without buying and integrating a separate text-analytics tool.