Setting up the right foundation for customer-centric, AI-driven CX requires a complete understanding of current performance without advanced capabilities.


EDITED 01.12.21

Customer-centricity and reducing customer effort, strategic to realizing competitive business advantage.

Competing on CX is the new battleground. Today, CX is strategically focused on customer-centricity and empowering customers to reduce their level of effort to complete transactions, using omnichannel contact centers, along with web, mobile, social media and apps, said Karl Walder, Managing Director, Oigo CX.

"If you want loyalty from customers, don't make them work hard or struggle for assistance. Every interaction...should be easy and convenient," said Emolytics.

Know who, what, how and which support experiences customers consider acceptable.

“Effort is the strongest driver to customer loyalty,” according to Gartner.

Being customer-centric means knowing who your customers are, what they like, how they prefer service and purchasing. That knowledge is used to craft curated cx and fuels, conversion, personalization, customer intimacy, longevity. And, ultimately, lifetime value. Not only within an organization, but across companies and the digital spectrum.

When you know how your customers engage over myriad interaction points to accomplish transactions with less effort, you understand what they consider an acceptable support experience.

Customer-centricity is realized in the convergence between pricing, product, customer service and CX.

A company’s competitive advantage no longer depends on differentiated pricing or product strategies, only. It’s now best realized in the convergence of pricing and product strategies, customer service and CX, across all customer interaction points, said cmswire.

During the last decade of the 1990’s, CX was strategically focused on customer relationship management (CRM) to identify the best customers across all touch-points over disparate platforms. Many CRM efforts became synonymous with sales force automation, lead and account management. Marketing adopted emerging internet and social technologies to drive brand awareness.

The first decade of the new millennium ushered in more focus on customer-centricity, delivering value to customers. Key performance indicators (KPIs) began measuring customer-centricity and customer effort scores (CES) to know how hard customers had to work to solve an issue or complete a transaction.

Today customers expect intelligent experiences, which are driven by data and machine learning. Omnichannel platforms generate a large portion of the human-to-human transactional data part of the equation. The full picture requires the aggregation of other platforms from marketing, web sites, mobile, sales and logistics to understand the KPIs for full customer-centricity.

What KPIs drive machine learning target objectives?

Before introducing advanced machine learning, predictive analytics, and AI into CX, it’s necessary for business users to have a comprehensive understanding of the KPIs that drive current performance without these advanced capabilities.

Target KPIs: These are some of the most important KPI, they represent the end-result stakeholders are looking for, examples include:

  • Business Process Results: customer lifetime value (CLV); churn rate; retention rate; conversion rate; problem resolution time (PRT); first response time (FRT); cost per interaction; customer satisfaction (CSAT); up-sell and cross-sell; average order value (AOV) and customer journey.
  • Customer Experience and Journey: CES, NPS, FCR, CSAT, channel switching and migration.

In-Process KPIs: These are the operational statistics for transactional measurement used to derive the output KPIs from particular business areas. From an omnichanel contact center perspective, these KPIs include:

  • Contact Traffic Analysis: Total inbound contacts, outbound contacts, average handle time, abandon rates, average speed of answer, campaign list penetration;
  • Agent Utilization Metrics: Wait-time, talk-time, break-time, and utilization/occupancy;
  • Agent Contact Results: The end result of the contact transaction. What action was taken or sales made.

Derivation of target KPIs requires a single source of truth.

In addition to omnichannel transaction data, it's necessary to have a combination of data from other platforms like marketing, sales, logistics and other enterprise areas.

Traditionally, getting a complete understanding of all this data meant gathering vast amounts of disparate data sets, hiring administrators, analysts and programmers to generate spreadsheets or custom applications. Organizations lucky enough to have large budgets employed teams to aggregate disparate enterprise data resources into data warehouse solutions.

However, these traditional approaches introduce time lags that fall outside the near-real-time interval required to manage contact centers effectively.

If an organization is unable to assess performance in near-real-time, or at most 15-minute increments, it’s an indicator it doesn’t have the foundation required for AI. Therefore, a descriptive analytics foundation is required to enable AI-driven customer-centric experiences.

Omnichannel platform example.

Transaction data generated by omnichannel contact center platforms are one input toward a complete customer-centric, single source of thruth data store.

Omnichannel platforms are characterized by the following:

  • Vast quantities of time-stamped contact detail record (DCR) transactions (phone, SMS, social media, web, chat, and emails) handled by self-service and agents;
  • They add data dimensions for time-stamped workflow data consisting of agent utilization statistics, self service transactions from IVR and chat bots and direct customer-to-agent results, along with audio, screen, and chat data used for quality assurance.

Can the organization's business users do self-service analysis?

If a company cannot easily provide both, single source of truth updates in near-real-time and self-service analytics and reporting information to business stakeholders, it needs to create a descriptive analytics foundation.

Enabling the self-service user experience is dependent on a descriptive analytics database foundation that facilitates high speed search over large quantities of data. For example, my company, a CLEC telco carrier and contact center provider, has captured over 3 billion contact-detail transactions over a 4-year period.

A descriptive analytics foundation provides the following:

  • Captures time-series data as raw transactions from operational systems;
  • Organizes data into high speed search hierarchies to explore dimensions in data;
  • Sets up a foundation for advanced machine learning, predictive analytics, and AI.

What is a self-service analytics and reporting experience?

A descriptive analytics foundation must allow business users to do the following:

  • Enable business users — not programmers — to explore data, capture reports, visualizations and dashboards and easily share with stakeholders;
  • Define data governance rules that enable the creation of roles for different types of business stakeholders;
  • Roles are used to push informational dashboards according to operational responsibility and access level, such as read-only, read-write and publish;
  • Explore performance using meaningful drill paths that show data dimensions and specified metrics and KPIs;
  • Select any set of metrics and organizational views;
  • Create reports, visualizations and dashboards for different business users;
  • Publish the information to all applicable stakeholders;
  • Deliver this experience over any desktop, mobile device, or app.

Requirements for omnichannel data dimensions.

In the context of omnichannel operations, teams must understand and manage to both input and output KPIs, and be able to assess the results across a defined set of business hierarchies and data dimensions to understand performance and profitability.

From the contact center operations perspective, these data dimensions include:

Target KPI categories:

  • Agent occupancy and utilization statistics;
  • Contact traffic inbound-work-queue analysis;
  • Outbound campaign results, and resultant return calls, chats, social posts, etc.

In-Process KPI categories:

  • Workflow results statistics and profitability;
  • Performance of organization stockholders, including executive, operations, account management, contact center, marketing, sales, and customer service.

The legitimate challenge for many operations teams is getting a complete picture and full understanding of how the volume and multiple dimensions of data affect overall business performance.

Immediate benefits to business stakeholders

Moving to advanced AI is time consuming and expensive. Therefore, migration has to take an incremental approach in order to generate immediate paybacks at each step of the evolutionary process.

Starting with a descriptive analytics foundation results in an immediate payback and real benefit to stakeholders. Armed with a single source of truth, they have the ability to do the following:

  • Understand what is driving performance in their area;
  • Deliver information immediately when making decisions;
  • Implement, test, and measure their operational changes;
  • Measure, monitor, and manage to organizational KPIs in near-real-time.

This information and analysis agility increases operational insight for management, monitoring, and performance. Arming stakeholders with a complete picture of their operational areas drives improvement of Target and In-Process KPIs.

Understand KPIs for each business stakeholder.

Stakeholders need both, similar and separate sets of data, to realize profitable operations. Once descriptive analytics views are defined, teams will know what KPIs to drive, and which machine learning objectives to prioritize.

In conclusion, a descriptive analytics foundation is necessary to initiate advanced machine learning and predictive analytics for customer-centric, AI-driven CX.

Elaine Sarduy is a freelance writer and content developer @Listing Debuts