An insurance broker who stammers has serious concerns about voice recognition analysis systems and the affect that they will have on us when calling a call centre to report a claim.
I have worked in the insurance industry for twenty five years and have always been aware that fraud and deception costs this industry millions of pounds a year. This cost is passed on to all customers in increased premiums. The main area of concern has always been motor and household claims where a policyholder submits either a fraudulent or exaggerated claim.
For many years insurance companies have tried to identify these claims so they can turn down the claim and even report the deception to the police. Until a few years ago, a claim that looked potentially flawed would be passed from a claims handler to a supervisor who would decide whether to ignore the issues and pay the claim or pass it on for further investigation. Commonly, this would result in a pre-arranged visit to the claimant by the insurance company's representative. For a small claim this would be a claims person, for a larger claim this would involve a member of staff who was maybe an ex-policeman. The visit was not only a face to face discussion with the claimant, but a first hand observation of the claimant's lifestyle and circumstances that would also add to the investigation.
The whole service industry is now moving from direct personal contact to call centres, help lines and other faceless methods.
The latest development is voice recognition analysis (VRA). This is a computer based system that records a phone call from the telephone to a computer and uses the caller's voice pattern to determine if they are telling the truth. Simply, it is no different to a lie detector. Some insurance companies are already using this and others are piloting it, but it is rapidly becoming an industry standard. Some will tell the caller that the call is being recorded, others may even say that the call is going through a VRA, but some will just say nothing. The VRA has no human involvement, the computer software decides if you are telling lies. It works during a say 15 minute call to report or discuss a claim and in the early part of the call the computer uses your voice when saying "Hello" and exchanging the usual introductions to set a standard. The operator will then ask you a lot of questions about the claim and within those there may be 10 or 15 key questions that are scored based not on the answer, but how you answer. Typically, the VRA will negative score a question if you hesitate in replying, do a lot of umm-ing and ahh-ing or just beat around the bush.
Going back to the days of the visit, they are more interested in how you answer and not the answer to the question. Usually, a negative score of two will put your claim into a high risk category of being fraudulent. From then on you will be mistrusted. The insurance company will investigate you in the best ways they can, credit rating, CCJ's mortgage arrears, tenant or homeowner, criminal record. Also, you will be asked to meet a representative who will already have in their mind that you are potentially a thief; yes, even a small exaggeration of a claim to obtain a few extra pounds is theft and carries the same charge as stealing goods. A fluent claimant will just go straight through the system and insurers will settle the claim.
Many people who stammer are comfortable in their own zone and when making a call to a call centre will have planned the call, what they are going to say in the introduction, and possibly use avoidance techniques and carefully chosen words and be happy to give yes and no answers. But knowing that if they block or stammer in more than two key questions could make them out to be a potential criminal in the mind of the operator does make life hard for those of us who stammer and could have serious repercussions.
Is this discrimination? or is it yet another pitfall for those of us who stammer?
From the Summer 2006 edition of Speaking Out, page 16