You will notice, in the introduction, that I used the term business drivers and not technology drivers. This nomenclature is an important distinction. Individuals outside of technology don't care all that much about how things work. They don't care that new technology is driven by serverless technology or uses object storage. Those items might be great from a technology perspective, but it isn't going to light a fire under someone like cutting call duration by a minute per call would. Those minutes of employee time add up quickly and end up being significant savings.

Think for a minute about a situation where you could save a minute per call that ended up reducing just one employee on the call center staff. That would mean that there is one less of the following to worry about:

• A salary
• A set of benefits
• A desk that needs to be occupied
• A person to train and onboard
• An annual review that needs to be completed
• A computer, phone, and headset

The list doesn't stop there; I'm sure that I forgot some essential expenses. You should be able to see that there is a significant impact possible with Connect, even with something as seemingly insignificant as a minute of call savings. Now that we have covered why the why is such an essential part of your project plan, let's discuss the benefits of Amazon Connect.

I believe that one of the most significant benefits of using Connect is to improve your user experience. Let's face it, who likes calling into a call center? If you have arrived at the point of calling into a company, you have already exhausted everything you can do on your end to resolve your issue. You have already tried to place an order online, return a product, or fix your problem, but nothing has worked. Now you have to call in to talk to a person to get the assistance that you need. Needing to talk to someone isn't the issue. The issue is the endless question-and-answer loops, the poor menu choices, and the failed call transfers. If you are like me, you might be wondering if they are just trying to get rid of you as a customer sometimes.

You don't want to run your call center like this. You want to provide an excellent experience where your customers on the other end of the line are wowed. Today there are a million companies that do a horrible job with customer experience. It wouldn't take much to wow a customer and ensure they become a repeat customer. Here is an excellent example.

My wife and I shop mostly on Amazon. Sometimes issues arise with what we purchase, and we must send it back. Amazon has built its entire business on being customer-centric. When we need to return something, it takes a few clicks and dropping the item off at the UPS store down the street. We couldn't ask for a better process. However, my wife and kids buy things online from clothing retailers. Many of these companies have horrible return procedures that require you to call in and plead to send back something that didn't fit. With some of these companies, especially during COVID-19, the customer service was so bad that they refused to shop there again. Does it cost Amazon more with their process? Maybe, maybe not. What I can tell you is that it cost the other companies a small fortune in lost sales from my wife and kids.

Connect makes it easy for your company to provide the kinds of experiences that make your customers happy. We will cover improved interaction, self-service, AI services, single-channel experiences, surveys, and sentiment analysis in the following sections to demonstrate these benefits.

Improved interaction

Practically everyone has been on a phone call where you have a horrible interaction experience with the call center's contact flow. A contact flow is a path that is programmed into the system to lead callers down an intended path. The problem with conventional systems is that they are hard to program, and the engineers that run them would rather be performing more interesting work. This situation is where you, as the caller, get frustrated because the menu has twenty-seven options that you have to listen through. Another common problem is when there isn't an option that fits your needs, and there is no way to get to a person to explain it to. Ultimately, you end up choosing any of the options and hoping that the person that answers your call can transfer you to the right individual.

The beautiful thing about Connect is that it's easy to set up your contact flows using a graphical interface. Everything is done via point-and-click and drag-and-drop motions. This is not to say that everything is rosy. You can get yourself into trouble if you don't know how to set up contact flows properly and allow them to get too large and unruly. We will cover this later in Part Two of the book to save you from that frustration. However, as the implementation engineer, the system allows you a more accessible and efficient way to configure and make changes to the system. Instead of a static system that gets changed rarely, you can change and adapt to the changing market and customer needs efficiently. Customers' needs change, and your interactions need to change as well.

Interfacing with enterprise applications

As we have covered before, Amazon Connect allows interfacing with other AWS services to expand its capabilities. One of the capabilities that you can add is interfacing with external enterprise applications. By itself, Connect is like any other telecom system. It doesn't contain any information about your customers, products, or operations. To provide capabilities such as AI bots, you need to interface with the systems and databases that contain all of this information. This interface is much more difficult in a conventional system and wouldn't have as rich a feature set as AWS.

I guess that you would say that interfacing with AWS services is the biggest background benefit to your company. However, these AWS services by themselves aren't very useful. They are the building blocks for you to construct your individualized services. The second biggest benefit of Amazon Connect is being able to use these services to create interfaces to your applications and databases. These application interfaces are what will make your AI bots, surveys, and sentiment analysis more impactful for your business.

The most effective way to interface between Connect and your applications is via an API. APIs allow you to query information for clients as well as update information easily using standard web calls. Most modern applications have some form of API interface for integrations. If your application does not have an API, all is not lost. Some instances of applications that might not have an API are older applications or potentially internally developed applications. You can still create interfaces for these applications, but you will have to have access to read the database directly. This method is a bit more complicated. You will need to know the data layout to query the data properly. Also, this method would best be left to read-only operations. Writing and mirroring the operations of the application can be quite complicated. You could corrupt data accidentally if there isn't a strong understanding of the application's operations.

Self-service

I'm willing to bet that you have run into a situation where you were stuck on a call waiting for someone to answer a straightforward question. Wouldn't it be better if you could get that information for that simple question yourself? The fact of the matter is that most call centers don't have self-service capabilities. Sure, the big banks do. They let you get your account balance and make transfers and such. But many companies do not have this capability. This inconvenience exists primarily because the big banks and retailers are the only ones that could afford to have this kind of interfacing set up for them. An advanced self-service functionality was outside the reach of the everyday company before Amazon Connect.

Since Connect allows you to utilize AWS services as part of the call center operations, you have at your disposal all the pieces required to provide self-service capabilities. To provide self-service, you need to tie your call center to your enterprise applications that hold vital data, such as your Enterprise Resource Planning (ERP) and Customer Relationship Management (CRM) systems. The interfaces for conventional on-premises call centers are very costly. The problem is further exacerbated by the lack of availability of these interfaces for your application. With the plethora of applications available on the market, the probability of an interface for yours declines. With Connect, as long as you have a way to communicate with your application via an API, you will be able to interface the two systems together. This ability to connect to an API gives you nearly unlimited capabilities to create a self-service experience for your customers.

AI services

Just like the self-service capabilities we just discussed, Connect also allows you to incorporate AI services. AWS has several off-the-shelf AI services that don't require any machine learning capabilities or massive training data. They are ready to use with a simple configuration and can then be interfaced with Connect. Typically, on-premises AI capabilities are out of reach for most companies. Even large corporations such as banks and large-scale retailers don't have elaborate AI capabilities. You might recall a situation where you called into one of these large organizations. You were prompted with a question where you could say "yes" or "no" to the prompt. The natural language processing power of these systems is relatively limited to understanding only a small response set.

There are several AI services that Connect can interface with that fit very well. The two major ones that most call centers will use would be Amazon Polly and Amazon Lex. Polly is a text-to-speech service that allows your call center to communicate with your customers verbally. Lex is a chatbot that uses natural language understanding and automatic speech recognition to provide a lifelike communication experience with your clients. These services work to reduce the burden on your call center staff by removing mundane workloads such as identity verification from them. Also, these AI services serve as the foundation of your self-service capabilities that improve customer satisfaction. We will cover the implementation of Polly and Lex later in Part Two of this book.

Multi-channel experiences

The internet and mobile applications have changed the way that we as consumers communicate with companies. Back in the olden days, when I was young, and rocks were soft, we only had three ways to communicate with a company. I could go to their place of business and speak in person, I could call them on the phone, or write them a letter. Later, you could send a fax, but it would be some time before people had fax machines in their homes. Today, we have dozens of pathways. You can still call them on the phone, but you can also use mobile applications and websites. When utilizing these different paths, you want to accomplish two things. The first is that you want your customers to have a unified experience no matter what method they use. Secondly, you want to minimize your staff requirements and utilize the staff you have to the fullest effect.

Before late 2019, if you had chat functionality in your application or website, it would have to be serviced independently from your call center staff. At least, it should have been. If you were utilizing your team, they would be operating in two different systems and would have to move in or out of the call center system while servicing individuals in the chat system. This set-up is suboptimal as it distorts your reporting and analytics, as it becomes hard to determine how efficiently your staff is operating. Seeing this deficiency, Amazon created a chat integration in Connect and released it in late 2019. Having chat integrated into your call center solves a lot of problems for your company. Issues such as the productivity issue discussed earlier, staff reduction, and providing customers a consistent experience become moot.

Surveys

One of the best ways to get feedback from your clients is to perform a survey about their experience. Using conventional on-premises systems conducting a survey wasn't the easiest to accomplish. You obviously wouldn't want to have the operator taking the call also perform a survey at the end. This conflict would mean that you would have to transfer the call to another operator to do a survey. This method is costly and uses staff for a function that might not utilize their expertise. After all, you hired them to take orders or answer support questions, not to take surveys. Often, these surveys would be handled by an outside firm. This mode saves your employees time but at a considerable operating expense to the company.

Since Connect allows the utilization of AWS services, you can create an entirely automated survey interface, utilizing the AI capabilities of Polly and Lex. This method will allow you to gather your customers' insights without impacting your staff or bottom line. Best of all, you can utilize this feedback to improve your overall customer experience. You can modify your contact flows to achieve a better experience and get nearly immediate feedback through the survey mechanism. You are essentially creating a continuous feedback loop on how your call center is operating.

Sentiment analysis

The last benefit of Connect that I wanted to cover is the ability to do sentiment analysis. Sentiment analysis is when you take a look at the conversations happening in your call center and determine the sentiment of the client's interaction by the words they use – to put it bluntly, to see if they are happy or not. Utilizing this information, you can tell if they are unhappy with a particular product or feature, if the call center agent needs more training, and when they like their experience. Before Connect, the implementation of this capability would have been too costly. You would have needed specialized processing power to handle this type of AI computation.

AWS makes sentiment analysis possible through the implementation of Contact Lens. Contact Lens uses several underlying AWS AI services to construct the sentiment analysis. You might be thinking it is complicated to set up the ability to recognize the speech in a phone call and then analyze it for sentiment. However, it is relatively straightforward, and although not easy, it's not complicated. Similar to surveys, you can use sentiment analysis to create feedback loops that continually improve your products and your customer experience. We will cover how to implement sentiment analysis using Contact Lens in Part Two of this book.

Now that we have covered the customer experience benefits, you will hopefully see that Connect can provide a far superior experience; that is, if you implement the correct technologies and set up the proper processes and feedback loops to improve continually. Using Connect, you should think of your call center as a living, evolving organism. You will need to break free of the contemporary patterns of set it and forget it used with conventional systems. You should find at least one, if not more, compelling why business stories for Connect's implementation within these benefits. We can now move on to the reduction of hard costs, which will help make a compelling financial story.

I've been in IT for over 25 years, and it seems that no matter how hard I try to drive projects with business value, it always comes back to costs. It probably originates from the fact that at most companies, IT is a cost center and doesn't drive any revenue directly. That is, unless you are a Software-as-a-Service (SaaS) company. As you are reading this, you might be thinking the same thing: I'm going to have to show how this implementation project will save us some money, or at least not cost us any more than we are paying today.

I'm reasonably confident that Connect will save you money over operating on-premises with a conventional system. However, there are two types of costs associated with a call center. They are the soft costs and hard costs. The soft costs are things like time savings, and hard costs are the real dollars you pay for equipment and services. Hard costs are easier to substantiate, and typically where I focus the majority of my efforts, so we will cover those first. Although they are genuine, soft costs are often qualitative versus quantitative and don't sway people into decisions easily. We will cover soft costs in the next section.

No circuit costs

One of the more apparent savings with Connect is that you no longer have to pay for any circuits for telecommunications for your call center. These would typically be in the form of T1 or T3 circuits that come into your building from the telephone company. Conventional phone carriers charge you for both the circuit and the phone calls going out over that circuit. Since Connect is an over-the-internet product, you won't need these phone circuits. AWS will handle all of your circuit needs for you. Like the phone company, AWS will charge you for the minutes you use only.

The Connect model is attractive to businesses because there are no sunk monthly costs with the pay-as-you-go model. For instance, let's say that your company sells Christmas ornaments, and your busy sales season is from June to November. However, in December, since all the store displays are already set up, and everyone has their inventory, your call volume drops to near zero. On premises with a conventional system, you would still have to pay for those circuits even though they aren't being used. If using Connect, your costs would drop to near zero along with your call volume.

It should be reasonably easy for you to identify the costs for your circuits that you currently use. If you are setting up a new call center, it might make sense to call a phone company and find out how much a circuit would cost, so that you can include an on-premises comparison. This comparison will help paint a picture as to how Connect is more cost-effective than other solutions.

No long-term contracts

If you have ever worked with a phone company with any circuit larger than a standard phone line, you will know how they lock you into a long-term contract. Typically, a T1 circuit could have a contract as long as three years. While this might not sound like an expense above and beyond the circuit costs we just talked about, it can be. The contracts that you enter with a phone company lock you into a specific usage pattern at a particular location. If any of those two things change, you might be on the hook for expenses to terminate the contract early. If you were to move to a different state where the provider didn't have a presence in order to move your circuit, you would probably owe them termination fees. These fees immediately recoup the cost of the installation from you.

Phone companies lock you into a contract because it costs them money to bring the circuits to your building. They spread this cost out over time to recoup them. It's similar to how you used to get a cell phone for nearly nothing but had to sign up for a two-year contract with the cell carrier. The cell carrier was just recouping the cost of the phone from you over a longer period.

Since Amazon Connect uses a pay-as-you-go model, these costs will never exist in the first place. You can create and remove phone numbers, call flows, or even whole Connect instances with no penalty. Although you might not be in a situation where this kind of hard cost would affect you, it does make a compelling argument against a conventional system.

No long delays for implementation and upgrades

With conventional telephone circuits, there is a lengthy delay with the installation. At best, you can get a circuit installed within 30 days. However, it wouldn't be uncommon to see an install timeline in the 90-day range. Unfortunately, there isn't much that can be done to accelerate the phone companies and get the line installed faster. While you're waiting for the circuits to be installed, your business isn't stopping. This delay is where the costs come in. You can't expand to meet increasing customer demand or launch that new product. While at a standstill, your company is burning real dollars in lost revenue and decreasing your customers' satisfaction.

Again, since Connect is a pay-as-you-go internet-based product, you don't have to wait for anything to be installed. You get to start using it the second that you want to. This lack of delay means that you can immediately address increased customer demand or launch that new product line. This agility and speed mean that your company can start to reap the benefits as quickly as you implement your Connect instance. Once Connect is up and online, it will scale to meet your needs, big or small. I might sound like a broken record right now because I keep stating all of the pay-as-you-go benefits, but they are numerous and essential.

No demarcation extension costs

Unless you are fortunate in terms of the layout of your building and where the phone companies' circuits come in, you will probably run into some costs associated with extending the circuit to your suite or telecom room. The phone company will only bring your circuits into the building. They aren't responsible for the rest of the way. The responsibility to extend that circuit is up to you. Typically, these costs are not astronomical, maybe a couple of thousand dollars. However, it is an inconvenience as another piece in an already complex puzzle.

Since we already discussed that with Connect, you wouldn't have any circuits in the first place, you wouldn't then need to extend the circuit. However, this is another instance where it's an important cost to incorporate when making a financial comparison against an existing or potential on-premises deployment. Many of these small costs get lost when performing an analysis since they don't occur often. Employee churn is very common in IT, and it wouldn't be unheard of for someone who installed a circuit to leave before the next one is provisioned. When the employee left, they took the tribal knowledge required for the circuit extension. These types of oversights lead to cost overruns or incomplete comparisons.

Extra equipment

I have saved the best hard cost for last: extra equipment. Since Connect is based on a software platform, there isn't any need to buy phones, cards for a telephone system, or special power over Ethernet switches. Suppose you take a second to think about all of the world's phones out there, sitting on desks unused. I'm sure the costs of those phones are in the hundreds of millions. The COVID-19 pandemic made this problem even worse. Phones are a depreciating asset that you don't have to purchase when using Amazon Connect.

With Connect, your users will use the computers and headsets that they already have. You won't need to purchase any extra equipment that depreciates or sits unused. Determining how much phones would cost is an important variable when comparing Connect to typical call center systems. The cost of phones alone adds up to be a significant expense. To make matters worse, the phones themselves depreciate very quickly. At one company I worked for, we purchased our phones second-hand and saved significantly. Others' misfortunes were our gain.

We have now covered the hard cost savings that can be attributed to utilizing Connect. If you can allocate costs appropriately for an existing system, you should be able to compare to your current system and show how migrating to Connect will save you a small fortune. Likewise, suppose you are going to be creating a new call center. In that case, you can properly compare and contrast costs and demonstrate how a software- and internet-based solution is far superior to on-premises systems.

The interfacing of other AWS services gives Connect the ability to reduce soft costs in your organization. In the User experience section, we touched on several capabilities that also decrease your soft costs. Before we start reviewing these savings opportunities, let's review a comment made earlier that soft costs are more qualitative and harder to use for financial decisions. Since soft costs are items such as reducing employee efforts and reducing staff, they become harder to quantify. Employees are different. They work at different speeds, have different career objectives, different skill sets, and so on. Therefore, whenever you estimate these soft costs, there is always some amount of doubt. This doubt is what makes crafting a financial decision based on them much more difficult.

To demonstrate this, let's take a look at this scenario. You have designed your Connect implementation and have created an AI bot that allows your customers to use self-service. Your customers can now call in and check their last order's status without speaking to an agent. You have chosen this self-service option because you spoke to the call center staff and asked them the most frequent question. You then asked the staff how long they would spend on the phone with a client looking up their order and letting them know the status. After you gathered all of the information, you took the average hourly cost of a call center agent and calculated the savings based on the estimated number of calls. In total, you have figured that the new self-service AI bot will save the company $124,546 in employee time.

Based on this scenario, you should be able to see how Connect can save you a significant amount of money. Unfortunately, soft cost savings are more of a bonus when seeking financial and project approval. If you were to base the entire project on the merits of the $124,546 alone, you would run into a lot of questions, such as the following:

• Do you have exact numbers on how many times this question was asked of staff?
• Are the time savings are based on an average of employees, efficient employees, or poor performers?
• Does the volume of this question ebb and flow based on the time of year?

There's almost an infinite number of permutations of the questions you will be asked. You won't have answers for all of them. That is why I suggest building your solid foundation for approval on hard costs first, and make the soft cost savings the icing on the cake.

We have already covered how an AI bot may potentially save you money by reducing staff commitments. This reduction is probably one of the biggest contributors to saving you cash with Connect. The possibilities for self-service using bots are nearly endless. Some possible options to think about are as follows:

• Order status
• Shipping status
• Last paycheck deposit amount
• Current vacation balance
• Repeating a customer's last order
• Pausing a standing order
• Canceling an order
• Getting test results

These are some examples of what can be automated with self-service. Of course, this is highly dependent on the industry that you are engaged in. The only way that you will be able to extract what these automations might look like would be to communicate with the call center team. They will be able to provide the required insight as to what would be most effective to implement. If you are implementing a new call center, you might extract information from the business unit requesting the call center. However, I would suggest a follow-up post-implementation after a few months to revisit and see what improvements can be made. All of these self-service automations are reliant on a very important concept. This concept is the interfacing of Connect with enterprise applications. We will cover these interfaces in depth in the next chapter.

Now that we have covered the benefits of Connect, you have hopefully created some strong stories as to how your company can use these benefits to its advantage. The customer experience benefits should give you some great "why" stories to communicate with other business units. The hard and soft cost savings will give you a great financial story to show. Finally, all of these things are possible by utilizing AWS services to create application interfaces. You may have some additional effort to complete the costing analysis and estimates. However, you should be well on your way to crafting the initial parts of your implementation plan.

The next part of the planning process involves gathering the business requirements from your stakeholders. You will not be able to judge the project scope, or how much the finished implementation will cost, unless you know what the business is trying to achieve. In the next chapter, we are going to cover how to extract this information, and how views will need to change to adapt to Connect's capabilities.