Using text analytics technology to analyze call-center transcripts and customer surveys, emails and tweets can improve customer service, in part by avoiding pitches that merely annoy customers, according to Ernst and Young.Every major financial institution records its inbound calls, but few examine the data for insights into their customers, the company says in a report by consultants Heidi Boyle, Bernhard Klein Wassink and Avril Castagnetta.Text analytics software searches textual communications for insights that could help businesses reduce customer attrition and target cross-selling efforts for greater success, the Ernst and Young report says. Without text analytics, executives miss the huge pool of comments and complaints describing customer attitudes about the company, its products and employees.For example, if customers are calling frequently with billing questions, it may be a sign that a confusing billing statement needs to be redesigned.Going further, the combination of text analytics and voice-recognition software can determine the emotional content of customer communications. If this is done while the customer is on the phone, the call-center script can be "adjusted to the customer's emotional state," the report notes."A customer calling about a replacement debit card is not ready to hear a pitch for a new mortgage product. Likewise, a caller who is upset about service and billing problems is probably not willing to be up-sold whole life insurance," the report says.Text analytics software can help customer-service reps, the report says, by "suggesting relationship-building actions, instead of standard cross-selling messages that stand little chance of success."Follow everything from CIO.com on Twitter @CIOonline, Facebook, Google + and LinkedIn.