Tag Archives: Technology Articles

PCI Scope Reduction Can Save Tens of Thousands of Dollars Per Year



By Art Coombs

High-profile stories of compromised credit cards and data breaches and their sobering aftereffects have dominated the headlines in recent years. As such, increasing security and reducing fraud is on the minds of many business leaders. This is particularly true of call centers, where credit card transactions are at the heart of their operations. These companies are challenged to provide a secure environment to accept credit cards while keeping the associated costs down.

The leading credit card companies set up the Payment Card Industry Data Security Standard (PCI DSS) to help businesses that take card payments reduce fraud. Built on solid security principles that apply to all sorts of data, it covers areas such as retention policies, encryption, physical security, authentication, and access control. According to the Verizon 2017 Payment Security Report, almost half of companies that accept credit cards fail to protect their payment card data on an ongoing basis.

The explanations vary widely as to why this is the case, but one of the primary reasons is the expense associated with maintaining full PCI compliance. In many cases, it’s prohibitively expensive. Fines levied by banks and credit card institutions for not being PCI compliant in the event of a breach can range from five thousand dollars to five hundred thousand dollars, highlighting the need for compliance despite the cost.Two approaches call centers can employ to reduce or even eliminate PCI scope is to use DTMF (dual-tone multi-frequency) suppression and SMS text messaging. Click To Tweet

Companies Face Mounting Costs

PCI-compliance costs add up quickly. Companies can expect to pay handsomely for items such as vulnerability scans, penetration testing, training, and policy development. Overall, there are twelve standards and more than four hundred controls outlined in the PCI DSS.

Often the largest direct expense (aside from remediation requirements resulting from a breach) is usually the PCI assessor and assessment fees, which, depending on the complexity of an organization, cost tens and even hundreds of thousands of dollars each year. These annual and biannual assessments are conducted by Qualified Security Assessor (QSA) companies, independent security organizations that have been qualified by the PCI Security Standards Council to validate a company’s adherence to PCI DSS.

The PCI Security Standards Council maintains an in-depth program for security companies seeking certification as Qualified Security Assessors and recertification each year. The requirements are stringent and comprehensive. Because of the time and energy individuals and companies invest in certification, they are justified in charging a premium for the assessments they conduct.

Reduce PCI Scope and Save Money

The litany of requirements is as costly as it is formidable, but call centers, as well as any company accepting credit cards, need to be aware that there are distinct ways to reduce the burden of applicable PCI controls. This means they can easily reduce the number of areas in the scope of PCI compliance that the company is responsible for. Reducing or eliminating areas of PCI scope can greatly reduce costs now and in the future and still provide a secure system.

Two approaches call centers can employ to reduce or even eliminate PCI scope is to use DTMF (dual-tone multi-frequency) suppression and SMS text messaging. These bypass the agents and contact center infrastructure, going instead directly to a tokenization service provided by the company’s payment processor and acquiring bank.

DTMF represents the tones the numbers on a phone make when pressed. DTMF suppression is a method that enables customers to enter their credit card information using the keypad on their phone. The agent stays on the line and never sees the numbers or hears the tones.

The second approach is to leverage SMS, or texting, so customers don’t have to give their credit card information verbally over the phone to the agent. SMS and an accompanying payment portal are a secure and smart solution for accepting payment for several reasons. Most consumers are already familiar with their mobile devices and SMS. This saves agents from having to explain a complicated web portal and payment screen. The consumer doesn’t need to download an app or go through a credit card terminal to make payments via SMS. SMS payments can be accepted around the world without any agents seeing or hearing the information.

The systems the company uses (CRMs, CMS, and payment systems) receive a confirmation or token validating that the transaction went through, but the credit card data never touches the company’s infrastructure. This greatly reduces risk: the company doesn’t have the credit card data, and it isn’t present, stored (recorded for quality assurance), or transmitted within the company’s systems. This reduces or eliminates PCI scope.

It’s important to note that regarding fraud prevention, even the most robust, 100 percent PCI-compliant environment could still be at risk when human agents, including employees, decide to commit fraud or theft. If they verbally receive numbers over the phone, they can memorize the critical information and then write it down once they leave the office or record the numbers and use them for their own nefarious purposes. In any card-not-present environment, there is risk. These approaches take that risk out of the picture.

Reduce Scope to Qualify for Self-Assessment

By using technologies that employ DTMF suppression and SMS, companies can reduce the scope of what’s required under an assessment so much that they’re no longer required to hire a consultant to conduct an assessment. Instead they can conduct a self-assessment, write a report, and submit it to the PCI council themselves, instantly saving tens of thousands of dollars or more while dramatically improving security.

Art Coombs is a published author on leadership and methodologies for BPOs, contact centers, and technical support. Art has more than twenty-five years of experience with several global firms and their call and BPO centers worldwide. He is president and CEO of KomBea, a fifteen-year-old software company that develops solutions for contact-center environments to help deal with the myriad of regulations and standards they face, including PCI compliance and HIPAA. For more information visit www.kombea.com.

Best Practices for Surviving a Ransomware Attack


Startel, Professional Teledata, Alston Tascom


By Jim Graham

In 1989 the first known ransomware attack occurred when twenty thousand floppy disks containing malware were distributed to researchers across more than ninety countries. In 2017 Symantec recorded an average of 1,242 ransomware complaints per day, not including the infamous WannaCry and NotPetya attacks. According to a survey conducted by Malwarebytes, one in six organizations impacted by a ransomware attack were down for twenty-five hours or more.

A recent attack on one of our clients was a painful reminder that ransomware continues to be a genuine threat to individuals and businesses worldwide. Our client received the virus upon clicking on a bad link in a “spear phishing” email. Their business was down for twenty-four hours before they were able to process calls.

The longer a business is down, the harder—and costlier—it is to recover. The financial impact can be just as staggering, with one hour of inactivity costing small businesses as much as $8,500. That doesn’t include lost business opportunities or the personnel cost associated with downtime.

Common Best Practices

There are many best practices, tips, and recommendations to mitigate a ransomware attack. The options can be overwhelming. However, you can lessen the likelihood you’ll become another statistic and decrease the impact of an attack by implementing these best practices.

1. Be Educated: Staff training is the first and best line of defense against ransomware. In most cases, systems are infected by user-initiated behavior such as clicking a malicious link in an email, opening an executable email attachment, or unknowingly giving a password to a potential hacker.

Educate staff about recognizing suspicious links and attachments. Phishing expeditions have become more sophisticated and targeted. These “spear phishing” attempts typically include client-specific information you’d assume no one else knows, making them much more believable. Never click on email links unless you’re absolutely certain of the identity of the sender.The longer a business is down, the harder—and costlier—it is to recover. Click To Tweet

2. Be Prepared: No matter how well-trained your staff is, be prepared for the possibility of a ransomware infection. This is where robust system and data backup strategies become essential. It’s critical to backup your data, software, and configuration settings frequently. Without a backup, you could permanently lose data. Create three copies, on two different media, and keep one copy stored securely off-site. Then test all backups to ensure you can successfully recover data.

A detailed incident response plan can make these instances a little less daunting. Take the time to put together an incident response plan, and test it each year. Also, consider investing in a business continuity and disaster recovery solution. These solutions minimize downtime and help ensure customer data remains secure and accessible 24/7.

Finally, in the unfortunate event you’re impacted by ransomware, consider enlisting the assistance of qualified IT professionals skilled at recovering from an attack. They’ll be able to get your company up and running and help minimize the impact on operations.

3. Stay Proactive: Once staff is well-trained and you have a strategy in place, continually monitor other areas of your business that may be vulnerable to ransomware. Implement these approaches to stay proactive:

  • Update operating system patches and antivirus software. On average, Microsoft releases several “critical or security”-related updates each month.
  • Limit administrative rights to only those that need to have them.
  • Deploy strong spam filters that block executable files.
  • Consider using a secure email gateway (SEG) in addition to your email client filter.
  • Set firewalls to block known malicious IP addresses.
  • Lock down your firewall from inside out to prevent data from being extracted.

HIPAA and Other Compliance Implications

A breach caused by a ransomware infection can have significant HIPAA and other compliance-related implications. Whether or not data has been taken, a successful attack is still considered a breach by HIPAA standards. Be sure you’re maintaining backups and log files for all systems that touch electronic protected health information (ePHI), because your company security policies will be subject to review by auditors. Proper HIPAA training is also essential in protecting ePHI.

Disclaimer

No matter how well prepared your business is, you can still be a victim of ransomware. However, following these recommendations will lessen the likelihood and impact of an attack.

StartelJim Graham co-founded Professional Teledata (PTD) in 1993 and served as vice president until the merger with Startel in September 2015. As the CTO of PTD, Jim draws upon his thirty years of computer and software development experience and twenty-three years of call center experience. Startel, Professional Teledata, and Alston Tascom provide unified communications, business process automation, and performance management solutions and services. They leverage their solutions and industry knowledge to empower organizations to improve agent productivity, reduce operating costs, and increase revenues. For more information, call 949-863-8776 or visit www.startel.com.

The Bots Are Coming!



Automated and AI-Driven Programs for Business

By Elena Langdon

Automation and artificial intelligence (AI) are all the rage these days—for good reason. The technology behind once too-good-to-be-true tools like facial recognition and 3-D printing has advanced rapidly. Many of us own or pine for smart devices and use dozens of apps a day for personal purposes. So what about business? How much can automation and AI boost productivity and profit at work? And what are the no-go zones for this exciting area of development?

First, Some Terms

“Automation” and “AI” are often used interchangeably, but there are important differences. Automation refers to processes that can be undertaken through a chain of events that trigger each other without human interference. We’ve seen it in manufacturing for decades. Simple contemporary business examples are Hootsuite or Buffer, programs that help automate a business’s social media participation.

AI refers to machines undertaking processes and making choices on their own, based on their programming and what they learn from it. There are different levels of AI, and the most powerful two—levels at which a machine can understand human thoughts, and be self-aware, respectively—have not been reached. So what can be accomplished now?

The Digital-Assistant Revolution

While C-3PO from Star Wars or Ava from Ex Machina are not in our immediate reality, AI is a driving force behind many business applications.

Personal digital assistants such as Siri and Cortana are good examples of AI-driven programs that can boost productivity, save time, and facilitate our lives. With one of these programs, you can delegate scheduling, play music, and check the stock market, all without typing, thanks to voice recognition capabilities. Pen, paper, and typing can be eliminated from the entire process.

Google Duplex is a more recent digital assistant that takes automation to a new level. It makes calls to humans to schedule appointments, request information, and order food. Instead of speaking with a typical robotic tone, Google Duplex mimics real speech patterns and uses fillers such as “um” and “hmm.” Plus, this bot interacts with human responses and can carry on a conversation. For this reason, its reception so far has included a mixture of awe and trepidation.The more complex the task, and the more it involves human reasoning,the less likely it will work for business. Click To Tweet

Proceed with Care

Caution might be needed for that type of digital assistant, especially from ethical and privacy standpoints. Should a human receptionist know that he’s talking to a machine? Is he being recorded so Google can learn from the exchange? Nevertheless, most of the tasks accomplished by Google Duplex involve little personal risk. If your haircut gets scheduled at the wrong time, it would be a nuisance but not a big loss.

However, you should approach some types of AI-driven programs with caution when it comes to business because of the risks involved. For example, in language translation, the technology can’t yet match the human capacity for communication. Automatic translation engines are great for getting the gist of a letter or website, but using them for business can result in embarrassment, misinformation, and even financial loss.

Most companies put time and money into writing compelling and clear texts; foreign-language copy requires the same attention. Despite recent advances in deep learning, machine translation is not like Google Duplex—it does not sound human, and it’s much less eloquent. More importantly, accuracy is seriously compromised with automatic translation—just think of all the menus with indecipherable items such as “The water fries the potato” and signs saying, “Beware of safety.”

Apply the same caution for verbal translation or interpreting, which has made headlines with programs that combine machine translation with voice recognition. Holding a conversation with someone in a language you don’t know by using “translator earbuds” might work for casual exchanges with inconsequential outcomes. However, if you need to speak to an employee about her performance or to an international branch manager about next quarter’s sales goals, you cannot rely on AI to accurately transmit your message. Between speech recognition flaws, cultural differences, and the incredible creativity behind any human being’s speech, it’s best to stick to a professional interpreter for bilingual business communication.

Lawyer Up or Bot Up?

If creative speech is one reason not to trust the machines, what about legal discourse? Does it make sense for a business to rely on automated contract-writing programs or document-reviewing apps? As with many machine-based applications, such programs can work, albeit in a limited context for limited purposes.

AI-driven programs will review legal documents at a fraction of the cost of a lawyer. This review process takes humans significant time, and lawyers take years to master it, yet computers have apparently learned the skill. That said, even apps’ websites make it clear that the apps will not provide legal advice and should be used only for the specific purpose of reviewing documents.

The formulaic language and boilerplate nature of legal documents lends itself well to AI and frees up time and money for actual legal strategy. In some ways, it’s like translation—you can get some entry-level tasks done, just not anything that requires tactics or nuanced meaning. And of course, nothing involving any risk to your business.

Look Both Ways Before You Leap

So the next time you see an ad for a new app that looks like a miracle cure for what’s ailing your business, by all means, don’t ignore it. There are many good applications for automated and AI-driven programs. Just be sure to research the program and consider its uses. The more complex the task, and the more it involves human reasoning, the less likely it will work for business—at least in an all-encompassing manner. Work patterns and skills are certainly changing, but the bots aren’t taking over just yet.

 

Elena Langdon is a certified Portuguese-to-English translator and interpreter and an active member of the American Translators Association (ATA). The American Translators Association represents over 10,000 translators and interpreters across 103 countries. For more information on ATA and to hire a translation or interpreting professional, please visit www.atanet.org.

Embrace the Intelligent Virtual Agent


Onvisource


The Promises, the Challenges, and the Pitfalls

By Ray Naeini

Adoption of intelligent virtual agent (IVA) or chatbots is a popular topic in today’s industry, as it can offer a broad range of benefits to both enterprises and their customers. A survey published by DMG Consulting in January 2018 showed that “increasing use of self-service” is one of the top three “enterprise servicing goals for 2018.” IVA uses artificial intelligence to automate customer service for chat or audio interactions with customers. It has the potential to operate as, or improve the performance of, live agents. Today’s customers mostly prefer self-service, especially through digital channels. IVA is a promising solution for improving customer satisfaction.

Automation Is Inevitable and Evolutionary

Automation has been a progressive, irreversible, and unstoppable trend. Automation has made drastic changes to our way of living and doing business. A few decades ago, customer service started with live switchboard agents manually connecting customer calls to the right customer service agents. In the 1980s live switchboard agents were replaced by interactive voice response (IVR) that could automatically prompt questions and route calls. IVA is the next evolutionary step in automation, going beyond the IVR functionality. It penetrates deeper into the enterprise organization and further automates various functions of customer service.Today’s customers mostly prefer self-service, especially through digital channels. Click To Tweet

Benefits of IVA

IVA offers a broad range of compelling benefits. It can assist live agents with real-time access to knowledge management systems, improving the quality and the speed of service. In certain cases, it processes customer service requests directly without the need for a live agent. In general, IVA can significantly improve the quality and the speed of the service while reducing live agents’ workload or payroll costs. It also reduces enterprise challenges related to live agent staffing, training, and retention.

IVA is available 24/7 from anywhere and can offer consistent customer service with an unlimited, real-time access to information during customer engagements. It automates repetitive tasks and can assist with or take over sophisticated customer service transactions. It supports multichannel via text-based chat or audio-based interactions. The use of IVA can go beyond customer service to benefit other departments such as sales and marketing for customer surveys or lead generation and qualifications.

Artificial Intelligence Is the Brain Behind IVA

What makes IVA smart enough to intelligently automate customer service is its use of artificial intelligence (AI) technologies. AI is a broad concept that started in the mid-1950s. It promised delivering intelligence similar to the human brain through progressive technological milestones. Advancements in mathematical modeling and natural language understanding, combined with faster and more cost-effective computers, make each technological milestone more capable of offering solutions to real-world problems.

The first two AI technological milestones that provide real solutions are called machine learning (ML) and deep machine learning (DML). The concept is to create mathematical models capable of continuously receiving, parsing, and categorizing a vast amount of relevant or training data to progressively increase the capabilities of the computers in natural language understanding, image processing or recognition, medical diagnostics, and so forth. This is similar to the basic functions of the human brain, as we were born with an inherited ability to continuously receive enormous amounts of data through our senses and then parse and categorize the information.

The use of ML and DML in IVA mainly focuses on natural language understanding (NLU) to converse with customers. In an IVA driven by ML, the data is analyzed and categorized by trying to understand the intent of the data (conversation) and extract the information associated with the intent (called entities) to prepare a response. The more intents and entities are analyzed and categorized, the more intelligent the IVA becomes. The ML approach, however, has certain limitations due to its single-layered analysis. In a DML-driven IVA, the data is analyzed by multiple layers or stages (using technologies such as neural network), and then at the end a collective scoring of the results from all layers is used for categorization.

IVA Challenges and Pitfalls

While IVA can deliver many benefits, it also creates challenges. Experiences related to the deployment of disruptive technologies tell us to avoid the hype of IVA and focus on applying it to each specific application systematically and progressively. IVA requires continuous training using a significant amount of valid and relevant data to improve its accuracy and performance.

We should also remember the pitfalls of early deployment of IVR that created significant customer dissatisfaction and try to avoid those mistakes. IVA deployed in a contact center environment should be capable of seamlessly integrating with the contact center’s overall workforce optimization (WFO).

OnviSourceRay Naeini is the chairman and CEO of OnviSource.

AnswerNet Acquires Synergy Solutions

AnswerNet acquired Synergy Solutions in an asset transaction. Synergy Solutions specializes in high-touch consultative customer service and sales support programs for many of the nation’s leading brands.

Headquartered in Phoenix, Arizona, Synergy Solutions was established in 1999 and focuses on innovative customer interaction solutions. This allows Synergy Solutions to provide superior results for its clients in fast-growth retail and e-commerce, as well as traditional verticals such as healthcare, insurance, and financial services.

Gary Pudles, president and CEO of AnswerNet stated, “Synergy Solutions furthers AnswerNet’s continued growth in high-touch customer engagement space for well-known companies and brands. AnswerNet’s customer care business has been growing exponentially over the last three years, and adding the incredible Synergy team further deepens our strength in providing solutions that help our clients continually stand out in providing support to their customers.”

Synergy’s president and co-founder Lori Fentem is staying with AnswerNet and will work closely with Pudles on building and executing AnswerNet’s growth strategy. Fentem is a well-regarded leader in the contact center industry. “We are thrilled by the opportunity to become part of the AnswerNet family,” says Fentem. “This acquisition allows Synergy the ability to continue to offer high-touch customer experience solutions. I believe that integrating with AnswerNet will present the ideal environment for Synergy employees and clients.”

Automation Success Requires Human Involvement



By Dan Somers

Automation and artificial intelligence (AI) can help save contact center costs, but primarily it increases customer satisfaction by speeding up responses and reducing customer efforts. Contact center automation falls broadly into three categories:

  1. Speeding up or automating the helpdesk agent (staff who capture and triage queries)
  2. Speeding up or automating the case handler (staff who resolve queries)
  3. Increasing self-service automation (chatbots, searchable FAQs, and self-help tools)

AI Challenges

Certain limitations of AI cannot guarantee the accuracy expected by customers, however. Some of these limitations are temporary, such as the comprehension capabilities of speech recognition, which will continue to improve. But other limitations relate to how machine-learning robots work.

All machine learning relies on studying real-life training data to predict or classify current data. The training data needs to be “labeled”—that is, it must have an outcome or class (tag) assigned to it, as judged by a person. For example, if a query comes in that says, “My server has crashed and is showing a blank screen,” then the chatbot will assign the best label it has in its training set, which might be “server crashed.”

However, in this example, a label of “faulty screen” might be assigned instead. The customer would be annoyed if the bot attempted to address a faulty screen issue instead of a server crash. This is an example of potential ambiguity. Furthermore, new issues will appear from new product launches, changes in quality, and evolution in the market. Lastly, the way people describe or view the same problem is more variable for certain issues than others.

Human-in-the-Loop

The only safe way of deploying bots within a contact center is to have a human-in-the-loop. This person will validate what the bots are doing, preferably with minimal impact to the customer.

So, who and where is the human-in-the-loop? It turns out that there are four general ways for humans to validate some or all of the process:

  1. A helpdesk agent can validate suggested responses before sending.
  2. The customer can validate that the response—or the question they asked—was comprehended.
  3. A third-party solution provider can check the performance of the bots and curate the process; this might be an internal or external data science team.
  4. The knowledge base manager can check the bots for satisfactory performance.Automation of contact centers yields promise, although not without humans-in-the-loop to maintain its performance. Click To Tweet

Considerations of Humans-in-the-Loop

There are pros and cons of different human-in-the-loop approaches. Some of these points are technical in nature but have substantial implications.

Agent: Some solutions on the market have AI recommend the next “best response” for the agent. The agents validate the response, not the categorization. For example, if two queries—“The strawberries I bought were tasteless” and “The strawberries I bought made me sick”—both lead to the same recommended response, “We’re sorry; please accept our voucher,” then the categorization models will degrade as they are not being updated with the accurate root cause.

Also, the insight generated by the models won’t allow executives to monitor product quality, design, and usability to then generate the self-service tools that can reduce contact center traffic. With this solution, other humans-in-the-loop will still be required elsewhere.

Customer Validation: If customers provide the required validation, it is scalable, but customers may not like having to correct their original query or the responses. If the query produces a new category, then there must be a process to deal with it. Fundamentally, the system cannot be relied upon with just these humans-in-the-loop.

Solution Provider: This is the status quo for most machine-learning deployments in real-world environments: a data science team, either internally or a third-party, sets up, curates, and retrains the models on a regular basis to maintain their performance. The pros are that these are the only humans-in-the-loop required. The cons are that these professionals are in short supply.

Knowledge Base Manager: This role has the most hidden potential benefit for having a human-in-the-loop. In a nontechnical environment, they will provide business rules on how to handle queries, as well as the training, trouble-shooting guides, and fault tree analysis to resolve issues.

In terms of their day-to-day role, they will be aware of product launches and modifications, but they also can use the rich insight of the labels coming from the contact center (both triage and resolution) to make improvements to both the knowledge base and the process. This includes updating the FAQs so customers can better use self-service. Also, this insight can inform other functions, such as product quality, product design, and customer experience, to help guide improvements.

Optimized Learning

A new approach that only requires a few humans-in-the-loop can exist because of a new technology called optimized learning. This is a form of machine learning that builds models but invites training from a human in such a way to minimize human input and still provide maximum performance. It is ideal for spotting new signals and improving existing ones.

Optimized learning doesn’t need to be in-line and suffers from none of the downsides of other approaches. Instead, it requires a fraction of the labeling otherwise required, even in a changing environment. The implications of this are profound. It means that a call center would only need to retain a few agents after the automation implementation, and they would handle the training that the optimized learning invited them to do in an offline capacity. This would maintain the models for labeling queries to generate both automation and insight, thus speeding up and reducing issues.

The rest of the automation would come from the rules originating from the knowledge base manager, as informed by the bots. This paves the way for improving chatbots and self-serve, searchable FAQs to free up contact center staff.

Conclusion

Automation of contact centers yields promise, although not without humans-in-the-loop to maintain its performance. There are many different flavors for human-in-the-loop AI automation. With new technology appearing, an optimized system is possible with a minimum number of humans who don’t need any data science skills. There is now no reason why the contact center of the future needs to look like those of the present. The same applies for the customer experience too.

Dan Somers is the CEO of Warwick Analytics, which provides call center automation solutions to address voice of customer (VoC) data, chatbots, service desks, and complaint handling.

Artificial Intelligence and the Call Center



Predict Who Is Calling and Why

By Nancy Lee

Artificial intelligence (AI) is intelligence demonstrated by machines, in contrast to the natural intelligence displayed by humans. In computer science, AI research is defined as the study of intelligent agents: any device that perceives its environment and takes actions that maximize its chance of successfully achieving goals. The term artificial intelligence is applied when a machine mimics cognitive functions generally associated with the human mind, such as learning and problem solving. Artificial intelligence algorithms improve their performance automatically through experience by finding patterns in a stream of input, such as telephone automatic number identification (ANI) information.

AI in the Call Center

Artificial intelligence is nothing new to Amtelco, which started developing a unique dataset for predictive call handling in 2005—the Intelligent Series (IS) Clairvoyant Agent feature. The latest iteration of this development is the innovative IS predictive intelligence feature. Predictive intelligence uses the ANI for a call to recognize the caller, eliminating the need for operators to reenter information. Predictive intelligence determines who is calling and the most likely reason why.

Predictive intelligence enables the call center system to learn with each call it receives and progressively improves the performance of subsequent tasks.

Time-Saving Features

Predictive intelligence saves a tremendous amount of time for both operators and callers. Operators simply select the reason for the call from the list, helping ensure error-free information.

For multiple callers from the same phone number, predictive intelligence sorts the list of callers by the number of times each person has called and displays the list in a pop-up box. The most frequent caller from a number appears at the top of the list. The operator can quickly choose the current caller from the list, and predictive intelligence retrieves that caller’s information from previous calls. For each caller, the list of reasons for previous calls displays with the most frequent reason for calling.

The textbook for teaching the call center system about predictive intelligence is comprised of a script template, an SQL server database separate from the IS system database, and an administration script for managing the database. The script template is the starting point for creating new predictive intelligence functions within existing scripts and using it to direct the disposition of the call. The separate database makes it possible to track callers and their reasons for calling and provides information about past calls to the script. The administration script is a user interface that enables supervisors to easily create a customized list of reasons for calling without having to access the database.Artificial intelligence gives the adage 'live and learn' new meaning for call center systems and their human operators. Click To Tweet

Looking to the Future

Although advanced AI could replace or reduce the need for human interaction, reducing the workload on call centers by answering simple or commonly asked questions, it could also enhance human interaction to improve and streamline the customer experience. This would retain the personalized aspect of calls that human interaction provides while augmenting it with AI technology to provide higher-quality service.

Voice recognition technology can be coupled with artificial intelligence capabilities to offer adaptive response suggestions to operators based on key words and phrases. This would reduce wait times and enable agents to instantly respond to a wide variety of caller needs with appropriate solutions.

Tone and voice analytics can gauge customer satisfaction and agent friendliness and enthusiasm to help provide individualized performance assessments. This form of AI also could monitor agent and customer call temperaments in real time, alerting managers when a situation escalates and allowing them to listen or step in if necessary.

Artificial intelligence gives the adage “live and learn” new meaning for call center systems and their human operators. And there seems to be no limit to what call center systems can learn to make life easier for call center agents.

AmtelcoNancy Lee is in marketing and advertising at Amtelco, a developer and supplier of call center and communications solutions located in McFarland, Wisconsin. Contact her at nlee@amtelco.com

How to Ensure That Your VoIP-Based Call Center Is Always Online



By Steve Walker

The industry trend towards Voice over IP (VoIP)-based PBXs is causing a shift in the technology underpinning the call center business. VoIP PBXs bring with them tremendous features and flexibility, but they also create some unique technical challenges. Since a VoIP PBX is essentially software running on a computer, how will you keep your agents active and phone lines up when your VoIP telephony environment goes down?

Computer problems are not a question of if, but when. And VoIP PBXs are no exception. VoIP PBXs may encounter problems on their own (for example, a hard disk failure), network problems may block them, or they may go idle in the event of a local VoIP carrier problem. (VoIP call centers on the East Coast of the United States will remember an extended outage of a particular carrier in November 2017). Any one of these events (and more) can bring your VoIP telephony environment to a halt and idle your agents.

If you’re planning to deploy a VoIP-based PBX, you need to ensure that you implement high availability (HA). In simplest terms, HA means that if one PBX fails for any reason, another will rapidly take its place and restore telephony services. This is normally achieved through “clustering,” which means having a standby PBX ready to take over for the primary PBX if things go wrong.

If you ask your IT person about HA or clustering, you might get an answer well-suited to an office computer but not appropriate to a telephony environment. To design a HA solution suitable to a mission-critical telephony environment, you need to consider the following six criteria:

1. Autonomy

This criteria is the most important requirement when designing a HA telephony environment. It means that damage or failure of one PBX in the cluster cannot negatively affect the others; they must be autonomous (share nothing). Simple or cheap solutions share hardware, software, and disk drives between primary and standby PBXs. But enterprise-caliber solutions, including those serving public service answer points (PSAPs), must have fully autonomous cluster members. Make sure your clustered PBXs are fully autonomous.

2. Synchronization

The information held in the PBX must be kept consistent between the primary and standby PBXs in the cluster, so that either can take over for the other on a moment’s notice. Solutions that share data break the first rule of autonomy, but solutions which synchronize data are ideal. Look for a solution that synchronizes data, not one that shares a data storage device. Just as important, ensure that the PBXs will automatically turn off synchronization if one of them is in poor health. Sharing data that may be corrupted by a failing PBX can destroy the other one, resulting in the call center going off-line.

3. Failure Detection

Simplistic HA solutions define failure as a black-or-white scenario (for example, a power outage affecting the building shuts down everything). But VoIP PBXs fail in their own unique ways. A software bug might prevent the PBX from connecting calls, or a memory error may prevent calls from reaching agents. Enterprise-caliber solutions require sophisticated health sensing and failure detection. This ensures that the PBX is running and telephony services are fully functional. Avoid solutions with simplistic failure detection.

4. PBX Separation

While putting the primary and secondary PBXs side by side is convenient, it minimizes the magnitude of failures the cluster can withstand. Instead you will want to place one PBX in your primary call center and the other far away, perhaps in a different state. That way, if you suffer a local or regional power or carrier outage, the backup PBX running far away can take over. Then agents can connect with mobile phones or work from home. Note as well that simplistic synchronization solutions break down whether the two PBXs are placed far away or one is placed in the cloud. Therefore, make sure your synchronization solution can handle any degree of physical separation of the two PBXs.

5. Rapid Detection and Failover

Your call center will suffer immensely if it takes fifteen minutes for your PBX to detect that something went wrong, and it will suffer again if it takes twenty minutes longer to switch to the backup. And a lengthy outage may put your call center SLAs (service level agreements) or contracts at risk. Ensure that your HA solution can rapidly failover from one PBX to the other and that failure detection (health monitoring) can trigger a failover in under one second if things go wrong.

6. Encryption

If your call center handles personal health information (i.e., for a medical facility), then information contained in the PBX (such as voicemails) may be protected health information (PHI). Voicemails synchronized between the two PBXs may be deemed “ePHI in transit,” which could violate rules pertaining to the protection of this information. Regulations like HIPAA in the USA, PHIPA in Canada, PDPA in Singapore, and so forth may impact your HA solution. You must ensure that communications between the two PBXs are encrypted to secure that information; this will also help protect the PBXs from internet hackers.

Conclusion

These six criteria define a minimum set of capabilities your HA environment must meet to ensure you maximize PBX uptime and maintain the productivity of your call center. Since VoIP PBXs are fundamentally software running on a computer, you will find a range of HA solutions from free and open-source (generic computer HA) to commercial products specifically for PBXs.

As you select your HA solution, evaluate your options using this criteria to find the solution that’s right for you. Don’t wait until your first VoIP PBX outage to start implementing a high-availability solution.

Steve Walker is the CTO at Telium, a manufacturer of telephony and telematics solutions specializing in VoIP.

Machine Learning Puts the “Intelligence” in Contact Center AI



By Bob Kasten

With the advent of computers, thoughts quickly turned to speculation that a computer could someday match human intelligence. In 1950 Alan Turing devised the Turing Test that became a threshold for when a machine is said to become intelligent. The test uses a human evaluator that watches a conversation between two parties. The evaluator knows that one of the parties is a machine, and if the evaluator cannot distinguish between the human and the machine, the machine is said to be intelligent.

While the notion of artificial intelligence (AI) can bring thoughts of computers someday becoming self-aware, we do not have to worry about this just yet. In our era, AI has become an important tool that can be used in contact centers to become a performance differentiator.

Machine learning is a branch of AI. It uses data to feed algorithms that automatically learn and improve. Machine learning falls into two broad categories: supervised and unsupervised. In supervised learning, the output datasets are provided and used to train the machine and get the desired outputs for future datasets. Unsupervised learning does not use output data, but instead the data is clustered into different classes and then analyzed.Having a conversation in a text format gives the ability to mine agent and customer interaction. Click To Tweet

Many industry verticals have become commoditized to the point where offerings are similar or the same across the competitive market. This applies to BPOs (business process outsourcers) as well as the companies that use their contact center services. In order to gain and keep market share, it is essential to give a best-in-class customer experience. Machine learning can help with this process.

Contact centers generate a multitude of data. Data sources include systems for CRM (customer relationship management), billing, collection, agent QA (quality assurance), call recording, chat, email, CSAT (customer satisfaction), social media, and so on. All this data tells a story about the customer and the contact center’s interaction with them.

The recent progress made in AI parallels similar progress with voice recognition and natural language-processing technology. It is now possible to convert real-time conversations or voice recordings to text with a high degree of accuracy. Having a conversation in a text format gives the ability to mine agent and customer interaction.

The easiest way to understand how data and machine learning can work together in concert is to describe a common scenario. The process uses an archetype that can be applied generally to supervised machine learning.

Here is a list of generic process steps, followed by a specific example that uses a voice recording as the data source. Note that the data could come from any of the above sources.

  • Identify data source: Locate a voice recording.
  • Generate learning data: Convert the voice into text.
  • Machine learning analysis: Process text data.
  • Machine learning correlation: Connect success and failure outcomes with patterns in the agent and customer conversation. Find out how the best agents generate success. Identify the criteria of the call that causes the interaction to be successful. The output of this step will produce actionable insights.
  • Make improvement suggestions: Use the insights from the previous step to make enhancement recommendations. Improvements can come in many forms, including agent training as well as agent and customer matching. This provides real-time customer data the agent can use during future sales opportunities or script changes.
  • Implement recommendations: Take action by improving the script or training the agent.
  • Predict success: The insights gained from machine learning can be used to predict the likelihood of success. A benefit of predictability is that an action can be taken based on the predicted behavior.
  • Feedback: Verify that agents are using the training recommendations by running the process iteratively to confirm that the feedback is contributing to attaining key success metrics. Feedback is also used in the success prediction.

Determine what successful agents say is the basis for building on that foundation and give it to agents who are not as successful. Once you know the why of success, you have taken the first step in answering how to improve it. Positive customer experience plays directly into understanding the success equation. A customer who makes a purchase or gives a positive customer satisfaction rating for the service they received will be identified by using the machine-learning process.

Other data sources can be added into the model to improve the predictability quotient. CSAT could be used to help further identify success patterns. Another possibility is using customer attributes that can be obtained commercially. Commercial data includes purchase habits, household income, age, gender, home market value, and occupation.

These data marts are a good way to give agents more direct information about customers. In addition, they are used to improve the accuracy of the predictive modeling. Customer attributes are used to enhance the customer experience and increase the likelihood of success. Many attributes about the customer that are found using their phone number or email address can give further insights.

Using a computer-generated call that speaks as if it were a human being brings this model full circle. This application has been deployed in the past couple of years. It still has a long way to go, because customers can tell the caller is a computer rather than a person. In 1950 Alan Turing probably never envisioned that his Turing Test would be used with automated dialing systems.

Using machine learning in conjunction with learning from the mined voice text, with the advances in natural language constructs, brings new opportunities, as seen with assembly-line automation in the early twentieth century. Calling and speaking using a computer reduces staff and leads to labor cost savings. It should be noted there are TCPA requirements, as well as state and federal laws, that should be thoroughly understood and followed before using this type of technology.

When we get to the point that a customer cannot tell that a machine called and is speaking to them, the Turing Test will pass, and the machine will be considered intelligent. Once that happens, we may be closer than we would like to the day when computational awareness becomes a reality.

 Bob Kasten is the founder of contactcentertools.com. His tools provide a holistic suite of agent performance management modules that includes KPI performance tracking dashboards, voice analytics, QA scoring, knowledge testing, goals, coaching, and secure clean desk agent communication. He spends his time consulting on contact center information technology projects and enhancing his agent performance management tools. He can be reached at bob.kasten@contactcentertools.com or www.linkedin.com/in/bob-kasten/

Add Resilience to Your Call Center

By Wayne Scaggs

Do I get up after a knockout event? My first, second, and third answer is “yes.” Why? It’s because it is in me to find a way over, around, under, or through the issue.

For the purpose of this article, issue refers to your call center system, the beating heart of your business. The processors in your system clock millions of cycles per second, and the software sends the data you need through wires to the screen or printer to provide the information needed.

I do not know who discovered the following proven method of troubleshooting, but I will say that I have used it and it often works. One of the first places to start to look for issues in your system is to verify what is happening between the keyboard and the monitor. This should be approached delicately, because you do not want to disrupt the balance that exists. It can be a tremendous time-saver when this intricate part of your system is at its peak form.

Here are some of the mitigating approaches to data loss you may want to consider to protect and preserve your company’s most valued resource: the data on your servers. This information is based on our experience and research in creating both the hosted and cloud systems that protect and preserve our clients’ data. We are required to protect and preserve that data. Click To Tweet

Procedures you may want to have in place and verify their readiness include Windows updates, snapshots, planned failover, checkpoints, server replication, virtual servers, SQL transaction log shipping, backups, and spares. All procedures may not be applicable to all systems; use the procedures that work best for your business.

  • Windows updates can be an issue. Sometimes they won’t let you shut off your computer or allow you go to work until the update is completely installed. Updates break things, and updates rearrange your comfortable setup that you worked so long to get the way you wanted. Something to remember about Windows updates is that they help keep your computers secure, and the longer the time between updates, the longer it takes to update your computer. More importantly, you cannot operate your business without computers and computers must be updated; it’s mandatory.
  • Snapshots of your server are taken at an instance in time, which you can go back to in order to restore your computer to the time the snapshot was taken. You should use snapshots before a major update or component change-out. Snapshots are performed automatically, and there is a limit to the number of snapshots before they are overwritten.
  • Planned failover is used when you want to use a different server for operations and replicate back to a server. This method keeps the operating server and the replicating server in sync.
  • Checkpoint is a manually created mirror of your server used to retreat to a known working point of your server.
  • Server replication is a scheduled event that replicates your server—both the operating system and all the programs.
  • Virtual servers can operate multiple servers in one physical computer. The physical computer has enough resources (disk drive space, processing power, and memory) to accommodate servers that require fewer resources. A virtual server can be used to house your replicated server. If a failure occurs on your operating server and you have that operating server replicating to a virtual server, you can now bring your virtual server online and continue to operate your business while the failed server is being attended to.
  • Transaction log shipping is a SQL server tool to reestablish your SQL database back to a point, determined by the frequency of the transaction log shipment. To make the transaction log work, you must have a separate and complete SQL server where you can ship the transaction logs.
  • Backup options include online storage, flash drives, physical hard drives, and cloud storage. Performing multiple routine backups and maintaining off-site storage is mandatory.
  • Spares could include virtual servers, disk drive raid arrays, network switches, backup Internet access, and backup online servers.

We live in a software-driven world, with software controlling the data of our businesses; we are required to protect and preserve that data. This article provides a view of a variety of systems from a higher level. Know what you have in place, and then choose to incorporate other options that add resilience to your call center.

Wayne Scaggs is the president of Alston Tascom, provider of call center database information and network telephony systems.

Save

Save

Save