June 1998

Designing Customer Interaction Networks
How Service Providers Can Put Business Rules
In Front Of Technology

BY BRIAN GALVIN

Companies that operate call centers are demanding more of CTI. Once, these companies emphasized the need to reduce costs. Now, they've added another requirement: the ability to deliver truly customized interactions to each customer. Customized interactions, however, demand more intimate pairings of business rules and the technologies that execute business rules. In other words, companies that want customized interactions with customers will likely need customized customer interaction solutions, or uniquely configured Customer Interaction Networks. It is this need for customization, for uniqueness, that presents opportunities to service providers -- systems integrators, value-added developers, and resellers.

Customer Interaction Networks, like relational databases, enterprise resource planning, and the Web, must be highly customized to deliver real business value. Businesses have differences that must be reflected in managing their customer interactions. These differences are important because real-time interactions with customers constitute a continuous series of "moments of truth," instances in which a business redefines itself with respect to its customers, one at a time. These interactions are mission-critical for every company, as is the ability to design their Customer Interaction Network "their way." This constitutes one of the chief competitive differentiators (good or bad, and whether they like it or not!).

Implementations of enterprise-scale frameworks for customer interaction networks therefore represent a historic opportunity for service providers -- just as database integration and enterprise resource planning have in the recent past. Systems integrators, value-added developers, and resellers will be essential to each company's Customer Interaction Network implementation.

Perhaps even more enticing to the leading service vendors (many of which are already building dedicated practices for the implementation of Customer Interaction Networks), these services will deliver very high margins for years to come. This is because the skill level required to successfully design and implement Customer Interaction Networks is daunting -- practitioners must combine advanced knowledge in business process design, telecommunications, call center operations, database design, application integration, and workflow management. The supply of skilled professionals is currently very tight, and will become much tighter before any easing occurs. Meanwhile, the demand for their services is booming because companies are aware of the opportunity -- and the commensurate threat to laggards -- posed by the Customer Interaction Network.

DESIGN CONSIDERATIONS
The key areas of opportunity for service providers (integrators, resellers, and developers) in building Customer Interaction Networks map very cleanly onto an orderly design methodology. This methodology is built around four "views," or models, of the call center, which offer complementary insights into the current state of the call center, and the means by which it can be converted into a Customer Interaction Network.

The models include the value model, the call flow model, the workflow model, and the data model. Each model suggests certain questions, and each tends to focus on a different aspect of the call center system. Perhaps most illuminating, each model corresponds to one of the key constituencies within the customer organization to which service providers sell their services and are held accountable (Table 1).

The Value Model
The value model looks at the call center (and the Customer Interaction Network) from the point of view of the line of business owner. This model requires describing how the call center fits into the value chain of the business unit (including purchasing, manufacturing, inventory, sales, fulfillment, shipping, etc.). This chain represents a series of steps in which economic value is added for the end user or customer. Typically, call centers have narrowly defined roles in the value model. For example, many call centers exist strictly for demand generation (telemarketing centers) or debt collection (collections centers). Others exist for post-sale customer service (technical support centers and the like).

Looking at the call center from the value model point-of-view helps focus the design team (which should include members from the platform provider, as well as the service provider and the customer) on questions such as "what are the goals of the call center?" and "how can we measure progress?" In addition, the design team should look at the value model for opportunities to expand the role of the call center in the value chain.

For example, a "simple" collections call center often receives calls from customers who rank in the top quintile of a company's customers, but who happened to have missed a payment. Currently, they are treated like all other collection calls -- push for the promise to pay and then get off the phone! However, a look at the bigger picture suggests that this may be one of few times per year that the company speaks with this high-value customer. Wouldn't it be best to work on establishing a relationship with this customer, building loyalty and brand equity, and possibly selling up, rather than just getting a promise to pay? In the most fundamental sense, the call from such a customer is an asset. Put another way, it is a moment of truth during which the company is redefined for this particular customer. It is an opportunity that shouldn't be squandered.

The Call Flow Model
The call flow model focuses on the actual progress of the voice connection over time during the customer interaction. Examining the call flow model tends to force one to focus on telecommunications devices and functionality -- such as the PBX/ACD, phone set, IVR, how the call is routed, and what treatments are played during queue times.

In addition, the call flow model represents a view from the customer's point of view, since the voice channel is typically the customer's only interface to the customer (recall that only a tiny fraction of today's call center traffic involves the customer using any media other than voice). Questions which immediately arise in examining an existing call model are "how is a call routed today?," "how long do customers wait in queues?," and "how do we prioritize calls, if at all?"

The call flow model also leads the design team to look at the call center from the telecommunications manager's perspective, since the telecom manager owns all the systems involved. It is in the call flow model that one comprehends issues such as load balancing among sites, least-cost routing, IVR strategies, and so forth.

The Workflow Model
The workflow model looks at call center operations from the agent's point of view. The agent doesn't deal with a customer interaction until the call arrives at the agent's desk, so the agent's view of interactions is quite different. It is application-centric, and has more to do with managing the real estate of the agent's desktop, the time activities conducted by the agent over time, and the flow of data and voice to and from the agent.

Accordingly, the workflow model is more closely aligned with application development and management teams. It is important to realize that the call flow and workflow models work hand in hand. For example, if most calls are handled completely by IVR scripts (a call flow model consideration), then there is more time available in the workflow model for the agent to add value for the customer and the firm, perhaps by cross-selling or relationship building.

The Data Model
This model is familiar to IT types around the world -- the conceptual design and representation of how data is stored and managed in an enterprise. Thus, the data model represents the call center world from the point of view of the IT Department, perhaps even more so than the workflow model.

Relationships between data entities and the flow of data are routinely analyzed and diagrammed already. What is somewhat new here is the notion that the data model supports the call flow and workflow models by providing persistence and context management capabilities.

Persistence is simply the maintenance of detailed information about a customer interaction after the interaction is complete. Much of this is already done today, but usually in many disparate systems that do not communicate. For example, call detail records from carriers and ACDs are not readily correlated, and neither of them is easily associated with transaction records maintained in business application databases.

DESIGN METHODOLOGY
With these four models in mind, we can now examine the methodology for building Customer Interaction Networks, focusing as we do so on the opportunities at each stage for high value, and therefore high margin, service provision by integrators, resellers, and developers.

Discovering Opportunities
The first stage, as one might expect, is the discovery phase. This phase is best conducted by using a cross-functional team of people from the customer, the service provider, and the platform provider. The goal of this phase is to describe the current state of the call center, using each of the four models to generate a complete view.

Typically, the best way to conduct discovery is to focus first on the value model, carefully describing the call center's role in the customer's value chain. Then, as a group, identify opportunities to enhance the call center's role in the value chain. Usually, this preliminary analysis generates many "low hanging fruit" ideas about how to move forward.

Discovery also includes a complete description of the existing call flow, agent workflow, and data models. Again, many obvious areas for improvement usually suggest themselves (such as noting that one center uses a proprietary hardware-based dialer that is three years old and unsupported, and which does not interact with the business application).

Defining Goals
The team, having completed the discovery stage, then compiles a list of needs and opportunities that are uncovered during discovery. This usually occurs during a brainstorming session, and it is generally quite easy to create a list of very exciting opportunities and compelling needs.

The service provider generally acts as a facilitator during the discovery and the needs and opportunities phases, using the phases to educate the customer on the potential of Customer Interaction Networks in a non-threatening, collaborative way. Often, customers who start out asking for nothing but screen pops soon recognize a much broader range of possibilities. At this point, they are tempted to eschew screen pops. The careful service provider, however, will help customers see that screen pops, while pedestrian, are still highly valuable and offer an excellent opportunity for a confidence-building first phase.

Redesigning The Call Flow And Workflow Models
At this point, the telecom team is asked to study the call flow model, and the IT team is asked to study the workflow model. (Also, the finance and line of business team is asked to consider the return on investment.) The call flow and workflow teams must work closely together, as changes in one area always affect the other.

Several key opportunities present themselves to service providers at this stage:

• Development of business rules-driven routing strategies. Moving beyond skills-based routing, this entails designing a system that asks, each time a customer initiates an interaction, the following questions:

-- Who are you? This is usually done via an IVR prompt for account number, followed by a database lookup. The data dip gives demographic and value information about the customer.

-- What do you want? Again, IVR is the usual medium, via menus.

-- What else do you want? By comparing the explicit request with the customer's demographics and buying history, business rules can identify cross-sell, up-sell, or relationship-building opportunities tailored to this customer.

-- What else is happening? Since the goal is to route the interaction based on the total business context, it is helpful to know what else is happening in the enterprise's business environment. For example, is there a power outage (utility example), or is the Dow dropping like a rock (finance example)? These factors can be as essential to how you route calls as who the customer is!

-- Who is available? Given the answers to the above, one can now look across the entire enterprise for a resource (an agent, an IVR port, a voice mail message, etc.) capable of resolving the customer's explicit and implicit needs under the control of business rules.

• Integration of CTI into business applications. This is being done today by most major applications vendors in the call center space. They are choosing to use the most powerful CTI engines in order to deliver maximum functionality to the agent desktop, including screen pops, intelligent transfer, and full telephony support in the native application interface. There are many opportunities beyond the obvious, however, which offer significant customization and new application development opportunities:

-- Data prefetch utilities. Since the Customer Interaction Network fully integrates telecommunications and data systems, it is possible to take full advantage of every second of the interaction. One example is prefetching data and attaching it to the call while the call is in queue (something available only with third-party call control, available from leading CTI vendors). It is possible, for example, to maintain a call object throughout the life of a call (regardless of the number of dissimilar switches the call may be carried on, and including call wrap-up time after the call is released). This means data can be attached to the call object by an easily developed utility (using either ActiveX or Java), regardless of where the call is. Over the course of the interaction's life, this makes it possible to accrete contextual information directly to the call. This information can then be used to dynamically control the workflow by varying the screen presentation based on the call's context.

-- Customized alarms and thresholds. By essentially transforming switches and phone networks into programmable devices accessible to Visual Basic, C++, or Java developers (and there are millions of these!), it becomes easy for a service provider to provide comprehensive alarms and thresholds. Moreover, most major CTI vendors provide native SNMP 2 support, so the use of network management systems such as HP's OpenView to manage the Customer Interaction Network becomes possible. To do this also requires customer integration -- customized management solutions are a major service opportunity.

-- Enhanced supervisor functionality. Again, with full access to switch- and carrier-based events, it is easy to build powerful new applications to allow Customer Interaction Network supervisors to dynamically adjust business rules, agent prompts, and routing strategies. Conceiving and designing the solutions to do this is another major service opportunity.

• Integration of diverse media. While it is the role of the Customer Interaction Network framework to provide access to various media via a media abstraction layer, it is the role of the service provider to help the management of a Customer Interaction Network design the systems that will actually present these media to agents or other interaction destinations. For example, e-mail, by its asynchronous nature, is very different from voice communications, which are inherently synchronous. Blending e-mail and voice at one agent station requires careful management of the workflow and call flow models by the service provider during the design phase.

Redesigning The Data Flow Model
Once the call flow and workflow models are redesigned (to meet the goals identified during the discovery and goals phases), the team examines the existing data model to identify data elements and relationships that must be added or changed. These changes will be needed to support the context passing required by the call flow and workflow models.

For example, business rules-driven routing uses real-time statistics (maintained by the Framework of the Customer Interaction Network) and database accesses to derive the context by which routing decisions are made. The tables of data needed for this may not exist in the customer's pre-existing data model. Or, the data may be contained in a fragmented fashion in a difficult-to-access database.

Depending on the time sensitivity of the data, some of it may profitably be pre-processed off-line and then cached in RAM for rapid access (which we call meta-cached data, and for which a patent is pending). For example, exact customer account balance is rarely necessary for routing; classification of customer wealth by tiers is usually sufficient. Since most people's wealth does not change on a call-by-call basis (unless the call is from Publisher's Clearinghouse...), a one-digit wealth index is suitable for routing decision-making. This digit can be reviewed offline once per week, or month, as desired, without compromising the speed of the routing decision. These meta-cache designs are clearly the territory of specialized service providers.

Reporting
Finally, after the required data elements are in place, the service provider leads the Customer Interaction Network design team in the report design phase. After a review of goals, call flow, and workflow, the design team decides what business reports will be needed to measure the effectiveness of change.

The service provider encourages the team to get beyond the "measure what you can, then manage by it" outlook. Such an outlook is common within call centers, since many of them have had to rely on reporting systems from hardware vendors, and these systems are often incompatible and usually provide only limited information.

A broader approach to report design is a major benefit of the Customer Interaction Network. Report design and development is both a business process reengineering and a technical issue, and is best done by a service provider.

SUMMARY
Customer Interaction Networks, which offer customer-by-customer customization of business services provided by phone or other medium, offer many exciting opportunities for service providers. It is likely that this trend will accelerate in coming years, as we are just at the beginning of the adoption curve for the technologies involved. Much like what has happened in the database and enterprise resource planning markets, we expect to see the emergence of many niche players focusing on the service and value-added development opportunities, as well as the emergence of dedicated Customer Interaction Networks practices among the large systems integrators.

Brian Galvin is director of channel development for Genesys Telecommunications Laboratories, Inc. Genesys combines computer and telephone systems to improve interactions between companies and their customers, increase productivity, lower costs, and achieve greater customer satisfaction and loyalty. The company's enterprise-wide platform and applications software constitutes an innovative suite of inbound, outbound, blended, network-based, and Internet-based call center products. For more information, visit www.genesyslab.com or call 888-436-3797.

What Is A Customer Interaction Network?

A Customer Interaction Network is a network of elements that work together intelligently, which means the elements act on contextual cues under the control of business rules. The Customer Interaction Network personalizes and optimizes the management of each (usually remote) interaction between an enterprise and its customers, based on the context of the interaction.

In this sidebar, we will examine each of the components of this definition:

• Elements. The elements of a Customer Interaction Network include the telecommunications systems used by an enterprise (the carriers, switches, IVRs, and phone sets, as well as the e-mail and other new communications systems), the data repositories (relational databases, data warehouses, etc.), the applications (desktop agent and enterprise resource planning applications), the framework (which connects all of the pieces, manages the routing of interactions, and tracks the state of interactions) and the people who interact with customers.

• Contextual Cues/Business Rules. In a Customer Interaction Network, elements work together by passing contextual cues from one system or device to another, and applying business rules based on this context. For example, how you pop a screen on an agent desktop should depend not only on who is calling, but also on what else is going on at the time within the enterprise (this is also part of the context of the interaction). The routing of calls -- in any medium -- is similarly subject to business rules in a Customer Interaction Network.

• Personalized Interactions. By using business rules and contextual information, the Customer Interaction Network can react to each customer in a personalized, highly customized fashion. For example, a brokerage house may want calls from high-value customers to automatically be routed to the customers' personal brokers, rather than queued for the next available broker. Also, cross-sell and up-sell prompts for agents could be tailored on a call-by-call basis to match the buying habits and buying power of the customer who is calling at the moment.

• Optimized Interactions. By putting business rules in charge of handling customer interactions, it is possible to deliver superior, customized service to valuable customers while meeting at least the minimum needs of all customers. This is because companies do not need to apply the same level of relationship building to all customers, and can handle some customers in a more automated, less costly fashion while still driving more revenue. In this way, firms seeking to optimize their customer interactions can do so where it counts (highly profitable long-term customers), while delivering different, more cost-effective service to those customers which actually cost the firm money. This is similar in concept to the idea that businesses should feel pressure to fire some of their customers.

• Remote Interactions. While dealing mostly with remote interactions (remote here refers to the physical relationship of the person requesting service and the person delivering it), Customer Interaction Networks could be equally applicable to in-person interactions. For example, a business rules-driven Customer Interaction Network could moderate the natural (and annoying) inclination of service employees to give precedence to telephone calls over the person standing in front of them (by not allowing any but the most important telephone calls to arrive at an occupied clerk's station). However, the usual use of Customer Interaction Networks is to manage the telephone, e-mail, and other electronic interactions between companies and their customers.