Directorates
·	Director General's Office
·	Data Development
·	Strategic Evaluations
·	Partnership and Continuous Evaluations
Publications
·	Evaluation Reports
·	Evaluation Briefs
·	Labour Market Development Agreement Series
·	Lessons Learned Series
·	Evaluation Tool Kit Series
·	Evaluation Forum Newsletter
·	From Theory To Practice

• This section highlights the U.S. experience to date with examples of new technologies and automation in the areas of:
• information/inquiry systems;
• claims taking activities;
• continued claims;
• automated payment delivery systems;
• expert systems;
• investigation and control;
• performance standards; and,
• general administration.

• For each of these areas, the discussion presents some of the models being adopted or tested in the U.S. -- and available information on the assessment of these models and lessons learned.
• Where possible, the impacts of these technologies are assessed according to their ability to provide more accurate, efficient, cost-effective and timely service for UI customers. These assessment criteria are customer-focused and results driven. As such, they are consistent with the principles guiding the new service delivery vision in Canada.
• It should be noted, however, that our discussions with U.S. contacts and key personnel at the ITSC indicate that hard data regarding these assessment criteria is still quite limited. Therefore, the following discussion draws together the information that is available, and also notes areas where further studies are underway.

• Three main types of information/inquiry systems are being used or tested in the U.S.:
• telephone operators located at call centres;
• automated telephone inquiry systems; and,
• automated self-service kiosks.

• Colorado has telephone operators located at a centralized claims centre to respond to inquiries and take initial claims. The call system was introduced statewide in April, 1991. By 1992, however, concerns about the length of waiting times led to the setting up of two separate telephone lines. Since then, one telephone line has been used for claims taking and a separate line has been used for inquiries.
• Separating the claim and inquiry lines has aided the management of incoming calls, reduced claimant waiting time and improved the level of service.
• In managing the call centre, priority is given to the line for claims taking. Therefore, the average waiting time is allowed to be longer on the inquiry line. The average length of an inquiry call is three to four minutes.
• Many calls on the inquiry line are asking whether a cheque has been mailed.
• The Colorado call system is discussed in detail in the section on initial claims.

• Colorado is expanding its telephone services to include a VRS system for inquiries and continued claims. If the state was starting today to introduce new technologies, it would probably implement a VRS for inquiries and continued claims before introducing a telephone system for initial claims.

• In states using a VRS, or considering the use of a VRS, the system usually includes the capacity to answer:
• basic questions about the UI program (e.g., the location and hours of the nearest local office, what to bring to the local office); and,
• questions about claim status (e.g. whether a claimant's cheque has been mailed and the amount of the cheque).

• There is particular emphasis on using VRS technology to deal with inquiries about claim status, because of the frequency of these types of calls. Providing this service usually involves linking the VRS to a data base of claimant records, and enabling claimants to access key parts of their own records through a touch-tone telephone.

• States using a VRS for information/inquiries generally report that their systems work well and improve customer service by increasing access to information.
• The automated inquiry system in Wisconsin was implemented between December, 1993 and April, 1994. The average length of an inquiry call ranges from 2.25 to 2.57 minutes. Since routine questions are answered by the VRS, local office staff are more available to deal with complex questions.
• About 80 percent of the calls taken by the automated inquiry system in Maryland are asking whether a cheque has been mailed. The VRS handles this type of question in about a minute. Under the old system of calling the local office, callers often found they were getting a busy signal for days.

• One of the important aspects of a VRS for inquiries is overall call management.
• In Washington state, the volume of calls to local offices increased after the VRS was introduced because callers were being referred to the local office whenever the VRS could not handle a question. Local offices found that their old telephone technology could not handle the additional calls.
• In Maryland, calls to the local offices decreased with the introduction of the VRS because a few telephone operators are located at the VRS call site to handle questions that the VRS menu cannot answer.
• In Wisconsin, the number of inquiry calls received by local offices has not declined, but the type of inquiry has changed. The VRS handles the routine questions, while the local office handles the more complex questions.

• To help identify and resolve start-up problems, states generally recommend thorough testing of new information/inquiry systems. They also recommend a staggered implementation schedule that applies a system in only a few new areas at a time.

• Several states (e.g., Pennsylvania, Minnesota, Ohio and Alaska) began the use of a VRS by first applying these systems to inquiries.
• Minnesota concluded that the main payoff for a VRS in their state was in continued claims, rather than inquiries. Therefore, they are expanding their system to include the taking of continued claims.
• In the case of Pennsylvania, Ohio and Alaska, the approach is to phase-in the use of a VRS for continued claims and inquiries by beginning with inquiries. Pennsylvania has completed the full implementation process and found that the phase-in approach aided the process of identifying and resolving start-up difficulties (including technical and programming problems).

• As indicated in Table 1, many states are experimenting with kiosks for Employment Services. The kiosks have key boards or touch screens that provide access to interactive services.
• Washington state is obtaining positive and highly visible results in using kiosks to improve access to information on job openings.
• If more of these experiments are successful, there is potential to expand the use of kiosks to other programs, including the provision of UI-related information.

• Pennsylvania is one state that is testing the use of multi-use kiosks. These kiosks serve a wide range of functions. For example, they provide information on job openings, UI and tax rules. They also enable users to renew their driver's license and apply for a fishing licence.
• Pennsylvania has found that kiosks located in public areas (e.g., malls) need to be more secure than kiosks located in local offices. This increases the cost of applying this technology because the price of a secure kiosk is about U.S. $20,000, compared to U.S. $5,000 for a kiosk located in a local office.

• Some states report problems with their kiosks.
• The kiosks in Iowa, for example, are registering a lot of down time even after 6 months of operation. It is still uncertain whether the problem is with the hardware or the software.
• Nebraska has encountered major disputes with the contractor.
• In Minnesota, there are disagreements over how kiosks should be used.

1. An automated telephone inquiry system is able to improve customer service and reduce calls to local UI offices by handling routine questions, particularly questions regarding claim status. However, call management is very important to the success of these systems.
• The inquiry line should be kept separate from call lines used for claims.
• Although most routine inquiries can be handled by an automated response, a few telephone operators should be linked to the VRS to deal with questions that are not part of the automatic menu.
2. Automated self-service kiosks appear to have considerable potential, although success depends on choosing hardware and software that match the use of the equipment.
3. Successful implementation of a new information/inquiry system requires thorough planning and testing.
• Thorough testing and a staggered implementation schedule are important ways to identify and address start-up problems.
• If the plan is to implement a full telephone system for inquiries and claims processing, it is best to begin with the inquiry system. If the inquiry system is not functioning well, the telephone lines for claims processing are likely to be filled by callers with routine questions.

• Three main types of claims taking systems are being used or tested in the U.S.:
• in-person filing at a local office;
• telephone processing; and,
• automated self-service kiosks.

• The traditional way of taking an initial UI claim involves collecting information from new claimants at the local office.
• While many states continue to use this method for some or all initial claims, new technologies and automation are modifying the traditional process.
• For example, on-line systems for initial claims and claim histories enable local office staff to key computer information while speaking with the claimant.
• In Indiana, claimants enter own information at PCs located in the local offices.

• Existing systems can impose limits on the potential for using some types of new technology. For example when the ITSC made initial assessments of client/server implementation, they found that existing system infrastructures may not support client/server architecture. Also, the cost of updating system infrastructure can be prohibitive given current budget constraints and high conversion costs.

• Several states are developing or using methods to process initial claims by telephone. The main motivation for telephone processing is increased customer service, higher customer satisfaction and cost reduction. The four main models in this area are:
• North Dakota's call and visit system;
• Colorado's centralized call system;
• the decentralized call system being pilot tested in California; and,
• Wisconsin's system of combining the use of VRS and telephone claimstakers.

• In North Dakota, the initial claims process begins with a telephone call to an 800 number. At the call centre, an operator takes basic information (e.g., name, social security number and reason for separation). Following the call, the local office is notified and schedules a visit by the caller to collect the rest of the information.
• North Dakota has been using this approach since the 1970s, and finds that the system is a convenient and relatively inexpensive way of servicing their rural population. The call and visit system enables the local office to manage its workload, and is particularly useful in the case of large layoffs.

• Colorado was the first state to replace in-person filing at the local office with telephone processing. The calls are taken by a central claims unit located in Denver. With the introduction of centralized telephone processing in April 1991, local UI operations were closed. Although local Job Service offices remain open to provide employment services, these offices are no longer involved in processing UI claims.
• The call system enables claimants to initiate all transactions associated with the establishment of new claims over the telephone with the aid of claimstaking staff. For confirmation, a prepared form is mailed to each claimant. These forms must be reviewed and signed by the claimant, and returned to the Denver central office before payments can be processed.
• Over 96 percent of initial claims are taken by telephone. The rest are taken by mail through bulk claims for major layoffs by large employers.
• Although claimants can initiate claims over the telephone, persons who are not exempt from work search must visit a Job Service office to register for work and to review their job-finding strategy.

• The U.S. Department of Labour has conducted an evaluation of the transition to centralized telephone processing of UI claims in Colorado. The evaluation compares performance indicators before and after the transition.
• Before the centralized call system was introduced, UI claims could only be initiated by visiting one of the 26 full-service Job Service Centres or fourteen satellite offices. Continued claims were centrally handled, with claimants mailing claim forms to the Denver central office on a bi-weekly basis.
• The evaluation study found that the level of accuracy for initial claims was not adversely affected by telephone processing. However, some aspects of performance did decline. In particular, there were longer first payment time lapses for claims within the state.
• The call system encountered major problems when the Emergency Unemployment Compensation Program (EUC) went into effect late in 1991. Under the EUC, persons who had exhausted their benefits could come back into active status. The telephone service had been designed to handle up to 40,000 calls per week — but weekly calls suddenly rose to 300,000 by January 1992. Average waiting time rose from 3 minutes to 75 minutes. One part of the solution was to establish separate lines for information and claims taking, because about one third of the calls were general inquiries.

• The Department of Labour evaluation also conducted a telephone survey of about 900 claimants. One sample had initiated claims under the old system and the second sample had initiated claims under the new system.
• The survey found that filing a claim by the call system required about half as much time as the old system.
• The survey also found that about 7 percent of the claimants who used the telephone service indicated that they would not have filed for benefits under the old system of going to a local office. More recent information indicates that UI take-up may be as much as 12 percent higher in the case of telephone processing. The number of denied claims is also higher. This suggests that some claimants respond to the convenience of a call system by filing marginal and short-term claims.

• Recent discussions with state officials indicates that Colorado's call system has produced a number of benefits.
• Reduced staff requirements and the closing of local UI offices has achieved significant savings. The call system is estimated to cost 20 percent less than the old system.
• Centralization has improved the ability to train and manage claimstaking staff. The use of a single site also makes it easier to improve procedures and telephone scripts. Through training, Colorado has compressed calls to file initial claims into about seven and a half minutes. The claimstaker is polite but moves the caller quickly through the required information.

• Colorado's experience highlights a number of lessons.
• The initial claims process can be nearly paperless.
• Active management of the call site through the use of an automated call distributor and separate lines for information and claims reduce the waiting time and improve customer service.
• Training and continuous improvements are necessary ingredients to success.
• Over 90 percent of claimants are satisfied with the call system.
• Introducing a call system at the same time as other major changes in UI can lead to major problems.

• California is pilot testing the taking of UI claims by telephone in 5 northern locations. The pilot dates are between January 1994 and June 1995. A new pilot is scheduled for the spring, summer and winter of 1995—to test telephone filing in a metropolitan area (i.e., in San Diego and Ventura locations).
• Potential claimants call a local telephone number. An automated call distributor allows the calls to be routed to the first available claimstaker. The claimstaker gathers all claim information from the caller, and keys the data into the computer during the telephone interview. Forms that verify eligibility are mailed to the caller, along with a handbook containing benefit information.

• The issue of fraud has been a major consideration in the California pilot projects. Fraud prevention activities include:
• asking basic questions for identification (e.g., questions regarding name, date of birth, prior addresses, social security number, base period employers); and
• examining the potential use of Personal Identification Numbers (PINs), voice recognition equipment, and blocking telephone numbers from other states.

• Preliminary results for the 5 northern locations indicate that, on average,:
• waiting time (being on hold) is less than 1 minute;
• filing a regular claim takes 6 minutes; and,
• filing an interstate, military or federal claim takes 11 minutes.

• The fourth major model is the Telephone Initial Claims Project underway in Wisconsin. This system has been piloted since February 1995, with plans for statewide implementation in the fall.
• The system is to have two UI call centres. One located in Madison and the other in Milwaukee. These call centres are expected to eliminate the need for UI local offices by fiscal year 1996.
• For callers with touch-tone telephones, the basic information is entered by VRS. Once this is done, the call and the information given to the VRS are simultaneously transferred to a UI specialist for the completion of more complex information and benefit rights information.
• Callers using a rotary telephone are immediately transferred to a UI specialist.
• The UI specialist can access wage records and drivers licence information to help confirm the identity of the caller.

• The Wisconsin model essentially takes the existing claims procedure and automates it. Wisconsin officials predict that their model will achieve cost savings similar to the Colorado model (i.e., savings of 20 percent). However, it is too soon to say whether the reduced UI staff time will pay for the increased level of technology (including the costs or hardware, software and maintenance).
• The ISTC is undertaking a post implementation analysis of the telephone processing systems in Wisconsin and California. The study will review how well the systems are saving costs, increasing customer satisfaction and meeting other goals. Their report is planned for the end of September.
• The ISTC is also developing a claims taking simulator to assess the costs and benefits of taking claims by telephone as compared to current procedures. The simulator will also provide a way to test various claims taking approaches and telephone scripts for length of time, ease of use and claimant reaction. The simulator will be ready by the end of September.
• Early simulation results suggest that a voice response unit (VRU) takes about twice as long as a telephone claimstaker to ask the caller for certain information. In the case of a VRU, some of the additional time is to explain how the caller should enter the information by touch-tone telephone.

• Oregon is now using about 120 kiosks for employment service information while looking at a possibility of submitting UI initial and continued claims by kiosk.

1. There is considerable interest in the U.S. in processing initial UI claims by telephone. The reasons include:
• convenience for UI claimants (e.g., no travel time, no transportation costs, no parking fees and no child care expenses); and,
• processing of initial claims by telephone is seen as complementing telephone systems for inquiries and continued claims.

2. The choice of model is still a matter of considerable debate, however. Currently, the two main models are the Colorado system of telephone claimstakers and the Wisconsin system of combining VRS automation with telephone claimstakers.

• The Colorado system is estimated to cost 20 percent less than the old system of in-person filing at a local office.
• The ISTC post implementation analysis of the telephone processing system in Wisconsin (due to be available in September) will add important information to this debate.

3. The use of a call system for initial claims can increase take-up and the filing of marginal claims. In Colorado, there has been a 7 to 12 percent increase in take-up plus an increase in the number of denied claims.
4. A call system raises the issue of what type of work search system to put in place to meet UI program requirements. Consideration of this issue becomes particularly important if a call system is to be accompanied by a major reduction in the number of local UI offices.
5. The problems encountered during the implementation of Colorado's call system highlights the importance of:
   • planning for increased caseloads during a recession;
   • anticipating potential problems if the call system is to be introduced at the same time as other major changes in the UI system; and,
   • using pilot tests and staged implementation to help minimize start-up and transition problems.

• Two main types of continued claim systems are being used in the U.S.:
• mail-in claims; and,
• automated telephone processing.

• The traditional way of submitting a continued claim involves mailing in weekly or bi-weekly status reports. Some new technologies are helping to streamline this method.
• A number of states use bar code scanning (e.g., Alaska, Minnesota and Pennsylvania).
• In Minnesota, bar codes appear on cheques and forms used to continue a claim. States report that this is an inexpensive way to track forms and aid data entry.

• Other states use full form scanners (e.g., Alabama, New York, Louisiana, Maryland, North Carolina, North Dakota and Oregon).
• In Maryland, for example, the scanner reads the social security number, key dates and answers to key questions. The scanning process generates a computer record and flags anything that requires the attention of clerical staff. Only about 10 percent of the scanned documents require manual interventions.
• As a lesson learned, states suggests that forms to be scanned may need to be re-designed to meet the needs of scanning equipment. To minimize errors, the forms should be changed before the scanning system goes into operation.

• Several states are going to an imaging system for documents (e.g., Maryland, North Carolina, Ohio and Oregon).
• An imaging system is considered to be more accurate than scanning. It eliminates the need to retain a hard copy or a micro fiche version of the original because it takes and stores an electronic picture of each document. Also, it facilitates creation and managing of a UI system based on electronic records and files.
• Maryland is implementing an imaging system for quarterly reports, wage records and continuing claims. The system will combine imaging with some OCR scanning to lift certain information from the documents to the mainframe. The new system is expected to increase efficiency, particularly where imaging and scanning replace manual processing. However, estimates of potential cost savings are not available.
• The cost of imaging is considerably more than scanning. Although other states are interested in imaging to improve records management, some find that such a system would only be cost-effective over the long term.

• A number of states report considerable satisfaction with high volume, cut sheet lazer printers (e.g., Indiana, Mississippi, Nebraska and Maryland).
• With these printers, states have been able to streamline the printing of forms. They have also been able to streamline mailing and reduce costs by combining their mailing of cheques and continued claim forms.

• As shown in Table 1, a large number of states have recently started to use a VRS or are installing these systems for continued claims. Most of these states seem to be offering claimants a choice between a VRS and mail-in claims.
• Maryland plans to operate a dual system which allows claimants to file continued claims by either mail or VRS. Once operational, the state expects 75 percent of clients to use the mail system, and 25 percent to use the VRS.
• In Wisconsin, 85 percent of continued claims are filed by telephone.

• The general view is that claimants like having access to a VRS and that these systems are well suited to continued claims because most of the information is numeric or simply "yes" or "no".
• In Wisconsin, feedback from claimants indicate that 90 percent rate the new system as good or excellent.

• Iowa provides an example of a decentralized VRS, with 15 VRUs located in field offices and 5 VRUs at head office. Their experience points out several lessons.
• It is important to have an effective method of managing calls when the host computer goes down. In Iowa, the VRUs take off-line calls and then transfer the information to the mainframe. Their system also has a "roll-over" capacity which transfers calls to the head office when a local unit goes down.
• Introducing a VRS requires considerable expertise for both hardware and software, because VRUs need good programming.
• Other states have also noted the importance of hardware and software expertise and compatibility. In Tennessee, for example, the VRS had to go through two mainframe computers.

• South Carolina has a VRU in each local office so that callers can use a local number. The information is then transferred to a central location for processing.
• South Carolina implemented its system in 5 or 6 waves, rather than in all local offices at one time. This approach allowed time to work out the problems encountered at each site.
• By contrast, New York started its VRS in 45 locations to provide state-wide coverage. This approach was considered to be a maintenance nightmare. Also, the on-site people were often not well equipped to identify and resolve problems. As a result, there was a lot of down time.
• The VRS in Virginia operates in a central site. Their experience highlights the importance of understanding the limitations of hardware and software. They found that the first VRS they tried was too small and could not be extended.

1. Although there is wide spread interest in taking continued claims by VRS in the U.S., there is not much hard evidence to compare the cost-effectiveness of this approach with an automated mail-in system.
2. Experience with automated telephone processing highlights the importance of technical support and managing calls. In particular, it is important to:
• ensure compatibility among various parts of the system;
• have good hardware and software expertise; and,
• ensure no information loss in the event of hardware or software problems.

• Only a few states offer the option of a direct deposit of benefit payments (e.g., Iowa, Minnesota and Oregon).
• Iowa offers a direct deposit system for VRS users. North Dakota is planning to survey claimants using its VRS to determine interest in a direct deposit system.

• The first payment made by direct deposit is usually slower than a mailed cheque, because of the time taken to set up a direct deposit with the banks. Oregon estimates that the delivery of the first cheque is two days behind the mail service. In Minnesota, the delay can be longer.
• After the first cheque, however, direct deposits can occur the same day as the continued claim is received and processed.

• The states offering direct deposits find there is not much take-up of this option. This result is probably due to a number of reasons:
• the delay in the first cheque;
• a lack of promotion;
• not everyone has a chequing account (e.g., chequing services are often used in the inner cities); and,
• unemployed people seem to have a preference for receiving benefits in the form of a cheque.

• At the state level, there is also some reluctance to introduce direct deposits because:
• some banks require that claimants receive a notice in the mail, and this eliminates the potential for saving on postage; and,
• there may be some legal issues (e.g., the courts have not yet tested whether making a withdrawal from a bank account is the same as cashing a cheque).

The early experience with direct deposit systems for UI in the U.S. indicates that some practical difficulties are impeding the application of this technology. For example:

• interest among UI claimants seems to be quite limited; and,
• it is not clear how much cost-savings can be achieved, particularly if banks require that recipients of direct deposits also receive a notice in the mail.

• Expert system technology refers to software designed to provide a form of artificial intelligence. The software gathers key facts by prompting the user through a series of questions, and renders certain decisions for the user on the basis of these facts and a set of programmed experiences or rules.
• Since 1987, The U.S. Department of Labour has funded five expert system demonstration projects. These projects were undertaken to determine the feasibility of using expert system technologies to aid the day-to-day operation of UI.
• The projects were located in Kansas, Maine, Missouri and Texas (which had two projects). Oregon has also experimented in this area.
• Kansas, Maine and Oregon tested systems to aid non-monetary determination.
• Oregon also tested a system to determine disaster unemployment assistance.
• Texas tested two systems. One designed to assist claims examiners and one to determine employment.

• The U.S. Department of Labour has conducted a formal evaluation of the Kansas expert system which was designed to assist in applying UI rules regarding voluntary quits. The evaluation found that the expert system:
• was able to gather facts and render decisions with reasonable accuracy;
• helped to structure interviews so that information necessary for the determination of eligibility was collected, extraneous information;
• ensured that decisions reached were consistent with state law;
• could be operated by relatively inexperienced personnel; and,
• could be used as a training aid.

• Regarding acceptance of the expert system, the evaluation noted that:
• the initial attitude of claimstakers ranged from indifference to open hostility—although their comfort level increased as they realized that the system was not designed to replace staff or reduce the "people factor"; and,
• deputy examiners were slow to realize the benefits of the expert system.

• At the present time, expert systems for UI are being used in only a few states.
• Rhode Island uses an expert system for disputed claims and reports that the system has achieved consistency in applying rules across the state.
• Alabama has recently implemented a claims taking system in eight offices. Their users have given the system a good report. In particular, they found it helpful in focusing the claimstaking interview.
• Florida has an expert system for adjudication. However, they have encountered a number of problems getting started and will probably require additional resources to bring this application into the mainstream of their delivery system.

• The U.S. experience with the use of expert systems in UI applications offers a number of lessons learned.
• The general view is that expert system technology can work well for relatively simple questions that are factual or involve mathematical calculations. However, expert systems that involve many variables, judgements or questions of intent become too complex to work well in day-to-day operations.
• Setting up a system requires plenty of help from the vendor, programmers and consulting services.
• Systems that fail to gain user support are much more difficult to integrate into the program delivery process.

• To assist the development and use of expert systems for UI in the U.S., the ITSC has undertaken to identify expert system applications which are candidates for sharing among states.

• Experience in other program areas suggests that expert systems may be easier to develop and apply to federal programs than state programs. In the case of a federal program, an expert system is more readily transferrable from one region to another.
• Texas has achieved considerable success and good publicity with an expert system for a federal program to provide disaster assistance.
• The Texas system has also been used by Florida, Louisiana and California for disaster assistance.
• The U.S. Department of Labour considers expert systems to be a potentially useful approach to providing better and more consistent service, particularly for federal programs.

1. It may be easier to develop and use expert systems for UI in Canada than in the U.S. because Canada's UI is a federal program.
2. The U.S. experience suggests that expert systems can be potentially useful, particularly for federal programs, because this type of technology can:
• render accurate decisions;
• help to structure client interviews;
• ensure consistency in the application of program rules; and,
• simplify certain parts of program delivery -- particularly areas involving simple questions that are factual or involve mathematical calculations.

3. The U.S. experience also highlights the importance of gaining user support for expert systems.

• Three ways in which states are using new technologies and automation to aid investigation and control are:
• automating aspects of the appeals process;
• sharing data across programs; and,
• experimenting with VRUs.

• About 75 percent of the states have automated all or a portion of their appeals process. These systems seem to be working well.
• In some states (e.g., Illinois and Kentucky), automation is used throughout the appeals process.
• Others (e.g., Montana, Georgia, Pennsylvania and Vermont) use an automated system to tracks appeals.
• Some (e.g., Nebraska and Texas) use automation to help schedule appeals.

• The automated appeals system in Texas was developed around three co-ordinated programs:
• a program for notices facilitates the entry of a received appeal into the host data base;
• a scheduling program schedules proceedings before the Appeals Tribunal; and,
• a host communication program brings forward data from the host mainframe for the purpose of manipulation on a PC.

• Maryland is developing a local area network (LAN) based system. They are also using a bar code system to help manage files and track each case. Their system is reported to have resulted in reduced clerical support, added to quality control and standardized all legal documents.

• In 1993, Wisconsin undertook an issue resolution project and an issue resolution appeals project. In response, the state is:
• establishing a VRS for employer reports and basic appeal questions;
• continuing the process of eliminating paper files by capturing information electronically; and,
• consolidating 25 adjudication sites into 4 sites.

• Another way in which states are using new technologies and automation to aid investigation and control involves data sharing across programs. As indicated in Table 1, there is data sharing in most states.
• In Wisconsin, claimstakers will be able to access drivers licence information to help confirm the identity of callers filing initial claims by telephone.
• Some states have online access to data for certain agencies such as the welfare department and the Internal Revenue Service.
• Also, most states provide information from employers' wage records to other agencies on request.

• Sharing data across programs is an area where states are planning to make further progress over the next few year.

• A primary cause of overpayments is claimants returning to work but failing to report their earnings.

• Several states use a mail-in post card system to obtain hire reports from employers. In 1992, Maryland reported that they were receiving about 25,000 record of hire forms per year (usually about 80 per day).

• Some of these states are considering the use of VRUs to take hire reports.
• Maryland is examining the potential for replacing its post card system with a VRU system.
• Georgia is already using VRUs. Employers call on the same line as UI claimants, but they press a certain code and then enter their employer account number, the hire date, the social security number of the person hired and date of birth.

• The ITSC is examining the potential for using VRS technology to replace manual post card systems for hire purchase reports. The results of this project are expected in September.

1. The U.S. experience indicates that there are considerable opportunities for using automation to streamline the management and processing of appeals. However, hard evidence on the cost-effectiveness of using automation and new technologies in this area is still quite limited.
• Wisconsin provided one estimate when it predicted that its new system for filing and processing appeals will save about U.S. $0.5 million per year, with a payback period of 2.5 years.
• Since states are generally not taking a re-engineering approach, additional cost savings may be possible if automation is part of a full-system re-engineering.

2. In the U.S., data sharing across programs is becoming an integral part of the investigation and control process.
3. Automation is also being used to help identify cases where UI recipients fail to report a return to work. In the U.S., the current emphasis is on experimenting with the use of VRUs to take hire reports from employers.
• In Canada, it may be useful to consider a wider range of models, including a mail-in system linked to the collection of other information regarding new employees.

• Currently, the performance testing of telephone systems for information and claim processing tends to emphasize waiting time (time on hold) and processing time. In this area, performance testing is aided by the fact the automated call distributors used to route callers to the first available representative can also keep statistical data on how many callers are waiting in the queue and how long they have been waiting.
• The recent Government Act regarding performance and results is generating considerable interest in performance and customer perception of performance.
• In this area, the Department of Labour is working with a group of state and regional representatives to develop a way of measuring/assessing performance. Their work will include a national survey of UI clients to determine customer satisfaction among people in the program. The survey is planned for next year.

• A number of states are also reported to be undertaking projects to develop ways to measure and monitor customer satisfaction.
• Wisconsin is planning a study of voice/data performance monitoring. The purpose of the study is to develop an approach to measure and track the systems performance of remote claim telephone sites.
• California plans to establish a co-ordinated ES/UI measurement system to measure customer satisfaction.
• Florida is developing an assessment process and strategies for using customer satisfaction data as the basis for continuous improvement. The process includes obtaining information from UI claimants. The information is to be used to develop customer service standards.

Given the current interest in performance standards in the U.S., it would be particularly useful to monitor the developments in this area.

• Colorado provides the most dramatic example of managing a major change in program delivery. In that state, all local UI offices were closed with the 1991 introduction of a centralized call site for the processing of initial UI claims.
• Prior to centralization, one-third of operations was located in central Denver, another third located in the Denver area, and the remaining third in other parts of the state. Majority of the central/suburban Denver workforce was moved to the central call site. Remote offices moved less staff to the processing centre.
• The change was disruptive to office operations and field staff. However, the state also reports that centralized staff have advantages for training and management (e.g., improvements in procedures can be made more quickly and telephone representatives can be moved from the inquiry line to the claims line to meet a rise in calls on the claims line).
• Colorado has a deliberate policy of having more experienced and better trained claims-takers. This has led to better quality of claims being taken and greater claimant satisfaction.

• Wisconsin is considered to be the most forward looking in terms of managing the transition of human resources. To minimize disruption and avoid layoffs, Wisconsin developed a three year plan during the early stages of its streamlining initiatives.
• The three year plan included information sessions and focus groups for staff, as well a redeployment opportunities.

• Other states are keeping local UI offices but co-locating UI and Employment Services in the same office. In some states (e.g., Rhode Island, Vermont, Ohio, Florida and Illinois), this includes cross training and cross utilization of the front-line workforce. As discussed in the next section, states are now planning to move more in the direction of one-stop centres.

The experience with major changes in Colorado and Wisconsin highlights the importance of human resource planning as a way of reducing disruption and avoiding layoffs.