The Key Missing Ingredient In GenAI Transformation Of Customer Engagement

Introduction

According to McKinsey, generative AI (GenAI) can deliver up to a 45% productivity improvement in customer engagement. However, many GenAI projects have hit roadblocks that have prevented enabling it at scale.

This article examines the foundational role of knowledge management in the GenAI transformation of customer engagement.

GenAI Can Transform Customer Engagement, With Some Help

A recent Forbes article featured a discussion with Morgan Stanley’s CTO that explained how generative technologies “are likely to rejuvenate knowledge management” and that Morgan Stanley has prioritized developing GenAI tools to assist financial advisors. Gartner also calls out knowledge management as a critical capability (subscription required) to harness GenAI for customer service and support.

Simply put, GenAI generates conversational responses based on natural language context, prompt and training. One way to introduce GenAI in an operational process is to treat it like a “new college hire” in your team. The recruit is bright and quick but unaware of your business-specific requirements, process and know-how.

To onboard this “new hire,” you assign it appropriate tasks while supporting it with relevant know-how and quality assurance. This is exactly what a knowledge platform provides to GenAI. All of these capabilities outlined below must be orchestrated to enable easy, effective use of GenAI for customer engagement:

  1. Best-Practice Prompts: Optimized prompts are at the core of getting good output from GenAI. A business-friendly prompt management service can offer a library of configurable, best-practice prompts for your use cases. Think of prompt management service as a supervisor who would personalize instructions for the “new hire” to effectively execute tasks.
  2. Relevant Content: Without trusted content input, GenAI output cannot be trusted. Content creation and curation at scale are critical to feeding a GenAI tool. Further, this content must be attributable (who is responsible for what content), permissioned (who is allowed to use what content) and personalized (who should see what content elements). The “new hire” must be fed the right material to be useful.
  3. Business Controls: GenAI must be used when appropriate and with suitable controls. For instance, compliance-heavy content may not be interpreted, summarized or restated by GenAI. Configuring and controlling knowledge content for GenAI access, therefore, is crucial. For example, the “new hire” should not be allowed to restate a legal disclaimer while responding to a customer.
  4. Quality Assurance: GenAI can hallucinate, so its output must be checked in real time for accuracy. Modern knowledge platforms can assess GenAI output against input context and content used to generate the prompt, ensuring relevance and accuracy. The knowledge platform must seamlessly invoke other response methods in the event of GenAI hallucination so customer experience is not compromised. When your “new hire” delivers subpar work product, the manager should be able to use alternatives.
  5. Closed-Loop Analytics: GenAI in customer engagement must be continuously measured and managed. Each iteration of prompt improvement or output feedback should be A/B tested to drive continuous improvement. GenAI operation should be closely monitored to optimize impact. Measuring your “new hire” is key to onboarding a competent resource.
  6. Composable Architecture: As GenAI technology develops, there are likely to be leapfrogging advances along the way. Situating GenAI within a composable knowledge platform minimizes switching costs and vendor lock-in for GenAI services. Occasionally, your “new hire” may not work out as expected, so you should be able to easily replace it with someone better.

Orchestrated together in a knowledge platform, these capabilities can de-risk the effective use of GenAI for customer engagement automation.

Challenges And Best Practices In GenAI

Innovative organizations are recognizing this key missing ingredient in GenAI transformation and investing accordingly. That said, companies may face challenges with adoption. Here are a few steps to consider for success:

  1. Connect process and content silos to empower AI: Using GenAI as a stand-alone tool with limited access to enterprise content is a common challenge among businesses. Companies trying to build a knowledge base using GenAI must recognize that the know-how is best assembled by extracting relevant information—such as top customer questions from contact recordings—and marrying that information, for example with answers generated from correct, long-form documents sitting in silos. With these connections in place, companies will be better able to accelerate the process of using GenAI.
  2. Orchestrate AI and humans in the flow of work: Often, AI can perform a task within a process, but it must work in concert with human experts and editors who can review AI output to be scalable and deliver trusted automation. Streamlined orchestration of AI and humans is vital to maximizing impact while preserving accountability and buy-in.
  3. Centralize and leverage knowledge: Some businesses build GenAI capability focused on digital channels, targeting customer self-service capabilities. However, they ignore the value of knowledge assets they already have in contact centers. Building digital capabilities from scratch is a recipe for “building fast and scaling slow.” Consider leveraging existing agent knowledge. By using GenAI’s capability to repurpose agent knowledge into digital channels, you are likely to see a bigger and more sustainable business impact.
  4. Crawl, walk and run quickly with a narrow focus: Given GenAI excitement, there is a tendency to try to bite more than you can chew. Some businesses pursue the big prize out of the gate: using GenAI to automate customer self-service. The capability can work but can also have public failures that could become PR flash points. Instead, consider starting with supervised GenAI use cases and quickly building knowledge bases. Then you can move to agent guidance use cases and, eventually, to customer-facing capabilities. This can increase the risk/reward aperture as you develop GenAI competence.

Conclusion

Knowledge management is one of the key ingredients for using GenAI transformation of customer engagement. By leveraging a knowledge platform and understanding the challenges and best practices, organizations will give themselves the best chance to harness the potential of GenAI in customer engagement safely and easily.

Previously published on Forbes.com
Contact us

Skip to content