To recap, there are several factors to consider when trying to predict changes in the ways in which customers are identified:
• businesses want to reduce the cost of fraud
• customers want convenience but also their personal information and assets protected • businesses need to comply with existing and new laws and regulations
• contact centers spend excessive amounts of money on identifying and verifying customer identities • existing methods of identity verification (e.g. PIN, password, etc) are not secure and are user-unfriendly.
The emergence of biometric technologies
Biometric technology uses physiological or behavioral characteristics to verify a person's claimed identity. Physiological biometrics includes fingerprints, iris, or retina recognition, and voice verification. Behavioral biometrics includes signature verification, gait and keystroke dynamics.
Of these, voice is the only biometric that can be used over the phone. In fact, a voice verification system’s strength lies in its ability to work over the phone or web making it a viable identity verification solution for contact centers. Voice verification systems use spoken words to generate a voiceprint. A voiceprint can be compared with a previously enrolled voiceprint to verify a caller’s identity. The most sophisticated systems generate a voiceprint by using spoken words to calculate vocal measurements of a caller’s vocal tract thereby creating a unique digital representation of an individual’s voice. These systems are not affected by factors such as the caller having a cold or using different types of phones, or aging. Voice verification systems are now delivering levels of accuracy and security that have proven robust enough for use by banks and insurers.
Data collected from Nuance deployments indicate that voice biometrics based ID&V provides higher levels of accuracy than PIN or agent handled verification. Specifically, voice biometrics based ID&V reduces the number of fraudsters that can reach the call center while increasing access levels for legitimate callers. As such, even though voice biometrics does not provide 100% accuracy, it provides the call center with measurable improvements in security and customer satisfaction over alternative ID&V methods.
The customer’s experience
Since speaking is natural and intuitive, a well-planned implementation can result in a better customer experience that eliminates the need for PINs or passwords. For example:
• ‘Account Number’ based voice verification – the caller is asked to speak their account number. The account number identifies the caller, and the spoken words are used to generate a voiceprint that verifies the caller is the account holder
• ‘Challenge Response’. Typically the customer is asked to repeat a series of numbers , e.g. “Please say ‘one seven three four’”. The spoken words are used to generate a voiceprint. The numbers spoken are usually different each time the caller phones.
In cases where a two-factor authentication process is required, voice verification can be combined with a ‘something you know’ – such as an answer to a memorable question.
The business benefits
Businesses benefit from two types of savings. These can be illustrated in the following example:-
A contact center receives 10 million inbound calls per annum with the existing identity verification procedure taking on average 26 seconds and being performed by an agent:
• Eliminating the time taken by an agent to verify a caller’s identity can save 60.9c per call ($6.09m per annum)
• Secure automated identity verification enables a broader range of fully automated services to be offered, reducing agent cost.
The potential benefits for the business are huge, and the customer also gains through a better experience, longer opening hours and greater identity protection.
Assessments performed by Nuance customers demonstrate that
implementing voice biometric ID&V in the call center results in a compelling ROI, with hard cost savings realized through reduced agent handled ID&V. These savings typically are sufficient to justify a voice biometric deployment, however, additional financial benefits can be realized from increased self- service usage, reduced fraud and decreased agent turnover. The most impactful benefit to the bottom line however lies in the ability to retain and gain customers by delivering a superior customer experience.
Voice verification can also be used to protect the enterprise against repudiation (where the customer says at a later date that they did not do it) as it can verify the physical presence on an individual at the other end of a phone line. Interestingly, this capability is already used by various US law enforcement agencies to check that released offenders are where they should be.
For procedures such as internet password resetting, the higher level of security achieved with voice verification can enable businesses to offer real-time password resets or reminders. This benefits both customer and business and can reduce up to 70% of helpdesk calls.
Voice verification has the advantages of near-ubiquity (the vast majority of people would be able to use it) as well as improving levels of security and reducing costs. The increasing demand of the public for identity protection, coupled with businesses’ permanent desire to increase profits mean that voice verification is an option that any company concerned about identity authentication should now seriously consider.
Feedback collected from Nuance deployments globally indicate that a third key benefit is realized when call centers implement a voice biometrics solution for automated ID&V – improved customer satisfaction. Although many deployments are initiated with the objective of reducing costs & improving security, improved customer satisfaction tends to drive expansions of voice biometric deployments as companies realize the competitive advantage derived from providing a positive ID&V experience at the onset of each call.
CALL ROUTING AND QUEUE MANAGEMENT
Collecting information about customers before an agent has spoken with them is a contact center technique which has been around for decades, under the wider auspices of CTI (computer telephony integration). CTI
infrastructures route calls and automate information retrieval to help agents deal with issues quickly and accurately, without transferring callers or leaving them on hold.
CTI and SIP
SIP ("Session Initiation Protocol" - see the section on IP) will enable companies to use CTI functionality throughout IT infrastructures instead using a proprietary hardware/software layer. Applications will use SIP commands to carry out call-related activities and also non-call functions such as presence management.
A contact center's CTI servers use caller ID information from PBXs to retrieve customer information from various databases. This information is then passed to the agent along with the voice call as a screen-pop, cutting down the time spent at the beginning of a conversation. If calls are transferred, the information follows the voice call so the customer does not have to repeat information they have already given. CTI is extremely effective, however, it's also expensive, proprietary and complex. The use of SIP will enable CTI functionality to be used far more widely and effectively.
Enabling a voice device to communicate with a data network has required the CTI middleware layer to translate TDM (Time Division Multiplexed) voice traffic into data. If a contact center uses SIP and a pure IP infrastructure, it removes the need for CTI as a separate layer in the IT infrastructure because SIP enables mobile phones, laptops, smartphones, etc to communicate directly with IT resources. In pure IP networks, calls will enter the infrastructure as VoIP traffic and travel to a SIP proxy server which initiates sessions with the necessary applications to perform call routing and customer information searches that the CTI server was formerly responsible for.
Standards-based SIP proxy servers are much cheaper than CTI servers and can be implemented on standard hardware. Integration is easier and quicker as all the input and output in the network is one standard protocol, which opens this up to smaller operations too, however all voice traffic must be through VoIP.
Those contact centers which use touchtone IVR or speech recognition considerably more than average have traditionally been found in the telecoms, utilities and finance sectors: often high-volume environments where a few seconds shaved from a call or a reduction in misrouting can save considerable amounts of money. Most financial services companies have many products which require specific skills and product knowledge. As such, routing based upon selection criteria such as customer account numbers, sales/service and specific product choices can take place, supported by an IVR front-end, functionality which is often known as 'auto-attendant'.
It is worth setting some definitions here. While some IVRs are ALSO used to front a contact center, they are typically designed to filter out and resolve the majority of calls, rather than route them. The real benefit of an IVR is that it has the ability to provide ALL the assistance that a customer needs, such as company or event information or when integrated with customer recognition technology, account balances and other information sourced from internal databases. A solution used solely to route callers would simply be the (cheaper and less sophisticated) auto-attendant functionality within the CTI solution.
This particular section of the report investigates the business purpose of the IVR / CTI application as it is used for routing calls, rather than providing a full-service solution (which is looked at within the 'Self-Service' section). As such, 'IVR' will be used here to represent this particular call routing functionality, although the actual technology may differ between contact centers.
In the past, less-automated or volume-based contact centers, such as public sector, and sales-focused operations, such as retail, have shown less of a demand for IVR call routing solutions. However, over the past few years, the majority of respondents from every sector have reported that they use touchtone IVR for call routing purposes. A further 1 in 8 use speech recognition for the same purpose, with retailers standing out this year as being ahead of the others.
Figure 77: Use of touchtone IVR and speech recognition to route calls, by vertical market